maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity
llvm-project/lldb/source/Expression/ClangExpressionDeclMap.cpp

2354 lines
83 KiB
C++
Raw Normal View History

Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
//===-- ClangExpressionDeclMap.cpp -----------------------------*- C++ -*-===//
//
// The LLVM Compiler Infrastructure
//
// This file is distributed under the University of Illinois Open Source
// License. See LICENSE.TXT for details.
//
//===----------------------------------------------------------------------===//
#include "lldb/Expression/ClangExpressionDeclMap.h"
// C Includes
// C++ Includes
// Other libraries and framework includes
// Project includes
Updated to latest LLVM. Major LLVM changes: - Sema is now exported (and there was much rejoicing.) - Storage classes are now centrally defined. Also fixed some bugs that the new LLVM picked up. llvm-svn: 114622
2010-09-23 11:01:22 +08:00
#include "clang/AST/DeclarationName.h"
Improved logging and cleaned up a redundant return statement. Now when ClangExpressionDeclMap returns a variable for a name, it pretty-prints that variable to the log instead of printing a (fairly useless) NamedDecl pointer. llvm-svn: 117972
2010-11-02 07:22:47 +08:00
#include "clang/AST/Decl.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
#include "lldb/lldb-private.h"
#include "lldb/Core/Address.h"
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
#include "lldb/Core/Error.h"
Updated the expression parser to use proper logging when looking for external variables. Also cleaned up the log messages coming from the DWARF interpreter. llvm-svn: 106613
2010-06-23 08:47:48 +08:00
#include "lldb/Core/Log.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
#include "lldb/Core/Module.h"
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
#include "lldb/Core/RegisterValue.h"
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
#include "lldb/Core/ValueObjectConstResult.h"
Added a new class, ASTDumper, that provides verbose diagnostics of Clang AST classes for the purpose of debugging the types LLDB produces for DWARF objects. The ASTDumper is currently only used in log output if you enable verbose mode in the expression log: log enable -v lldb expr Its output then appears in the log for external variables used by the expr command. llvm-svn: 124703
2011-02-02 07:43:26 +08:00
#include "lldb/Expression/ASTDumper.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
#include "lldb/Expression/ClangASTSource.h"
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
#include "lldb/Expression/ClangPersistentVariables.h"
Endian patch from Kirk Beitz that allows better cross platform building. llvm-svn: 124643
2011-02-01 09:31:41 +08:00
#include "lldb/Host/Endian.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
#include "lldb/Symbol/ClangASTContext.h"
Modified the lldb_private::Type clang type resolving code to handle three cases when getting the clang type: - need only a forward declaration - need a clang type that can be used for layout (members and args/return types) - need a full clang type This allows us to partially parse the clang types and be as lazy as possible. The first case is when we just need to declare a type and we will complete it later. The forward declaration happens only for class/union/structs and enums. The layout type allows us to resolve the full clang type _except_ if we have any modifiers on a pointer or reference (both R and L value). In this case when we are adding members or function args or return types, we only need to know how the type will be laid out and we can defer completing the pointee type until we later need it. The last type means we need a full definition for the clang type. Did some renaming of some enumerations to get rid of the old "DC" prefix (which stands for DebugCore which is no longer around). Modified the clang namespace support to be almost ready to be fed to the expression parser. I made a new ClangNamespaceDecl class that can carry around the AST and the namespace decl so we can copy it into the expression AST. I modified the symbol vendor and symbol file plug-ins to use this new class. llvm-svn: 118976
2010-11-13 11:52:47 +08:00
#include "lldb/Symbol/ClangNamespaceDecl.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
#include "lldb/Symbol/CompileUnit.h"
#include "lldb/Symbol/Function.h"
#include "lldb/Symbol/ObjectFile.h"
#include "lldb/Symbol/SymbolContext.h"
#include "lldb/Symbol/Type.h"
#include "lldb/Symbol/TypeList.h"
#include "lldb/Symbol/Variable.h"
#include "lldb/Symbol/VariableList.h"
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
#include "lldb/Target/ExecutionContext.h"
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
#include "lldb/Target/Process.h"
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
#include "lldb/Target/RegisterContext.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
#include "lldb/Target/StackFrame.h"
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
#include "lldb/Target/Target.h"
Handle the case where you make a ClangExpressionDeclMap without a selected frame. llvm-svn: 121099
2010-12-07 09:56:02 +08:00
#include "lldb/Target/Thread.h"
Improved logging and cleaned up a redundant return statement. Now when ClangExpressionDeclMap returns a variable for a name, it pretty-prints that variable to the log instead of printing a (fairly useless) NamedDecl pointer. llvm-svn: 117972
2010-11-02 07:22:47 +08:00
#include "llvm/Support/raw_ostream.h"
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
using namespace lldb;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
using namespace lldb_private;
using namespace clang;
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
ClangExpressionDeclMap::ClangExpressionDeclMap (bool keep_result_in_memory) :
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
m_found_entities (),
m_struct_members (),
Order of initialization lists. This patch fixes all of the warnings due to unordered initialization lists. Patch by Marco Minutoli. llvm-svn: 129290
2011-04-12 03:41:40 +08:00
m_keep_result_in_memory (keep_result_in_memory),
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_parser_vars (),
Order of initialization lists. This patch fixes all of the warnings due to unordered initialization lists. Patch by Marco Minutoli. llvm-svn: 129290
2011-04-12 03:41:40 +08:00
m_struct_vars ()
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
EnableStructVars();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
ClangExpressionDeclMap::~ClangExpressionDeclMap()
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
{
DidDematerialize();
DisableStructVars();
}
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
void
ClangExpressionDeclMap::WillParse(ExecutionContext &exe_ctx)
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
EnableParserVars();
m_parser_vars->m_exe_ctx = &exe_ctx;
Fixed a bug where the parser-specific members of persistent variables were staying around too long. This caused the following problem: - A persistent result variable is created for the result of an expression. The pointer to the corresponding Decl is stored in the variable. - The persistent variable is looked up during struct generation (correctly) using its Decl. - Another expression defines a new result variable which happens to have a Decl in the same place as the original result variable. - The persistent variable is looked up during struct generation using its Decl, but the old result variable appears first in the list and has the same Decl pointer. The fix is to destroy parser-specific data when it is no longer valid. Also improved some logging as I diagnosed the bug. llvm-svn: 112540
2010-08-31 06:17:16 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (exe_ctx.frame)
m_parser_vars->m_sym_ctx = exe_ctx.frame->GetSymbolContext(lldb::eSymbolContextEverything);
Handle the case where you make a ClangExpressionDeclMap without a selected frame. llvm-svn: 121099
2010-12-07 09:56:02 +08:00
else if (exe_ctx.thread)
m_parser_vars->m_sym_ctx = exe_ctx.thread->GetStackFrameAtIndex(0)->GetSymbolContext(lldb::eSymbolContextEverything);
Adding support for fetching the Dynamic Value for ObjC Objects. llvm-svn: 130701
2011-05-03 02:13:59 +08:00
else if (exe_ctx.process)
m_parser_vars->m_sym_ctx = SymbolContext(exe_ctx.target->GetSP(), ModuleSP());
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (exe_ctx.target)
m_parser_vars->m_persistent_vars = &exe_ctx.target->GetPersistentVariables();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
}
Fixed a problem where function calls on i386 weren't being generated correctly. Also added a messy way to single-step through expressions that I will improve soon. llvm-svn: 117342
2010-10-26 08:31:56 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
void
ClangExpressionDeclMap::DidParse()
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
{
if (m_parser_vars.get())
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
for (size_t entity_index = 0, num_entities = m_found_entities.GetSize();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
entity_index < num_entities;
++entity_index)
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP var_sp(m_found_entities.GetVariableAtIndex(entity_index));
if (var_sp &&
var_sp->m_parser_vars.get() &&
var_sp->m_parser_vars->m_lldb_value)
delete var_sp->m_parser_vars->m_lldb_value;
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
var_sp->DisableParserVars();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
}
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
for (size_t pvar_index = 0, num_pvars = m_parser_vars->m_persistent_vars->GetSize();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
pvar_index < num_pvars;
++pvar_index)
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP pvar_sp(m_parser_vars->m_persistent_vars->GetVariableAtIndex(pvar_index));
if (pvar_sp)
pvar_sp->DisableParserVars();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
}
DisableParserVars();
Added extra logging, and made sure that the argument struct for expressions is deallocated when the ClangExpressionDeclMap is taken down. llvm-svn: 116028
2010-10-08 09:58:41 +08:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
// Interface for IRForTarget
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
const ConstString &
ClangExpressionDeclMap::GetPersistentResultName ()
Added automatically generated result variables for each expression. It is now possible to do things like this: (lldb) expr int $i = 5; $i + 1 $0 = (int) 6 (lldb) expr $i + 3 $1 = (int) 8 (lldb) expr $1 + $0 $2 = (int) 14 As a bonus, this allowed us to move printing of expression results into the ClangPersistentVariable class. This code needs a bit of refactoring -- in particular, ClangExpressionDeclMap has eaten one too many bacteria and needs to undergo mitosis -- but the infrastructure appears to be holding up nicely. llvm-svn: 110896
2010-08-12 09:56:52 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
assert (m_parser_vars.get());
if (!m_struct_vars->m_result_name)
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
{
Target *target = m_parser_vars->GetTarget();
assert (target);
m_struct_vars->m_result_name = target->GetPersistentVariables().GetNextPersistentVariableName();
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
return m_struct_vars->m_result_name;
Added automatically generated result variables for each expression. It is now possible to do things like this: (lldb) expr int $i = 5; $i + 1 $0 = (int) 6 (lldb) expr $i + 3 $1 = (int) 8 (lldb) expr $1 + $0 $2 = (int) 14 As a bonus, this allowed us to move printing of expression results into the ClangPersistentVariable class. This code needs a bit of refactoring -- in particular, ClangExpressionDeclMap has eaten one too many bacteria and needs to undergo mitosis -- but the infrastructure appears to be holding up nicely. llvm-svn: 110896
2010-08-12 09:56:52 +08:00
}
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
lldb::ClangExpressionVariableSP
ClangExpressionDeclMap::BuildIntegerVariable (const ConstString &name,
lldb_private::TypeFromParser type,
const llvm::APInt& value)
{
assert (m_parser_vars.get());
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
ExecutionContext *exe_ctx = m_parser_vars->m_exe_ctx;
clang::ASTContext *context(exe_ctx->target->GetScratchClangASTContext()->getASTContext());
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
TypeFromUser user_type(ClangASTContext::CopyType(context,
type.GetASTContext(),
type.GetOpaqueQualType()),
context);
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
if (!m_parser_vars->m_persistent_vars->CreatePersistentVariable (exe_ctx->GetBestExecutionContextScope (),
name,
Fixed a problem where constant results of expressions were not being created in the proper way, meaning results were getting lost. llvm-svn: 122800
2011-01-04 10:41:41 +08:00
user_type,
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
exe_ctx->process->GetByteOrder(),
exe_ctx->process->GetAddressByteSize()))
Fixed a problem where constant results of expressions were not being created in the proper way, meaning results were getting lost. llvm-svn: 122800
2011-01-04 10:41:41 +08:00
return lldb::ClangExpressionVariableSP();
ClangExpressionVariableSP pvar_sp (m_parser_vars->m_persistent_vars->GetVariable(name));
if (!pvar_sp)
return lldb::ClangExpressionVariableSP();
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
Fixed a problem where constant results of expressions were not being created in the proper way, meaning results were getting lost. llvm-svn: 122800
2011-01-04 10:41:41 +08:00
uint8_t *pvar_data = pvar_sp->GetValueBytes();
if (pvar_data == NULL)
return lldb::ClangExpressionVariableSP();
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
uint64_t value64 = value.getLimitedValue();
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
ByteOrder byte_order = exe_ctx->process->GetByteOrder();
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
size_t num_val_bytes = sizeof(value64);
Fixed a problem where constant results of expressions were not being created in the proper way, meaning results were getting lost. llvm-svn: 122800
2011-01-04 10:41:41 +08:00
size_t num_data_bytes = pvar_sp->GetByteSize();
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
size_t num_bytes = num_val_bytes;
if (num_bytes > num_data_bytes)
num_bytes = num_data_bytes;
for (off_t byte_idx = 0;
byte_idx < num_bytes;
++byte_idx)
{
uint64_t shift = byte_idx * 8;
uint64_t mask = 0xffll << shift;
uint8_t cur_byte = (uint8_t)((value64 & mask) >> shift);
switch (byte_order)
{
Fixed a problem where constant results of expressions were not being created in the proper way, meaning results were getting lost. llvm-svn: 122800
2011-01-04 10:41:41 +08:00
case eByteOrderBig:
// High Low
// Original: |AABBCCDDEEFFGGHH|
// Target: |EEFFGGHH|
pvar_data[num_data_bytes - (1 + byte_idx)] = cur_byte;
break;
case eByteOrderLittle:
// Target: |HHGGFFEE|
pvar_data[byte_idx] = cur_byte;
break;
default:
return lldb::ClangExpressionVariableSP();
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
}
}
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
pvar_sp->m_flags |= ClangExpressionVariable::EVIsFreezeDried;
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
pvar_sp->m_flags |= ClangExpressionVariable::EVIsLLDBAllocated;
pvar_sp->m_flags |= ClangExpressionVariable::EVNeedsAllocation;
Fixed a problem where constant results of expressions were not being created in the proper way, meaning results were getting lost. llvm-svn: 122800
2011-01-04 10:41:41 +08:00
return pvar_sp;
Implemented a feature where the expression parser can avoid running the code in the target if the expression's result is known and the expression has no side effects. Right now this feature is quite conservative in its guess about side effects, and it only computes integer results, but the machinery to make it more sophisticated is there. llvm-svn: 121952
2010-12-16 11:17:46 +08:00
}
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
lldb::ClangExpressionVariableSP
ClangExpressionDeclMap::BuildCastVariable (const ConstString &name,
clang::VarDecl *decl,
lldb_private::TypeFromParser type)
{
assert (m_parser_vars.get());
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
ExecutionContext *exe_ctx = m_parser_vars->m_exe_ctx;
clang::ASTContext *context(exe_ctx->target->GetScratchClangASTContext()->getASTContext());
ClangExpressionVariableSP var_sp (m_found_entities.GetVariable(decl));
if (!var_sp)
var_sp = m_parser_vars->m_persistent_vars->GetVariable(decl);
if (!var_sp)
return ClangExpressionVariableSP();
TypeFromUser user_type(ClangASTContext::CopyType(context,
type.GetASTContext(),
type.GetOpaqueQualType()),
context);
TypeFromUser var_type = var_sp->GetTypeFromUser();
VariableSP var = FindVariableInScope (*exe_ctx->frame, var_sp->GetName(), &var_type);
if (!var)
return lldb::ClangExpressionVariableSP(); // but we should handle this; it may be a persistent variable
ValueObjectSP var_valobj = exe_ctx->frame->GetValueObjectForFrameVariable(var, lldb::eNoDynamicValues);
if (!var_valobj)
return lldb::ClangExpressionVariableSP();
ValueObjectSP var_casted_valobj = var_valobj->CastPointerType(name.GetCString(), user_type);
if (!var_casted_valobj)
return lldb::ClangExpressionVariableSP();
if (log)
{
StreamString my_stream_string;
ClangASTType::DumpTypeDescription (var_type.GetASTContext(),
var_type.GetOpaqueQualType(),
&my_stream_string);
log->Printf("Building cast variable to type: %s", my_stream_string.GetString().c_str());
}
ClangExpressionVariableSP pvar_sp = m_parser_vars->m_persistent_vars->CreatePersistentVariable (var_casted_valobj);
if (!pvar_sp)
return lldb::ClangExpressionVariableSP();
if (pvar_sp != m_parser_vars->m_persistent_vars->GetVariable(name))
return lldb::ClangExpressionVariableSP();
pvar_sp->m_flags |= ClangExpressionVariable::EVIsFreezeDried;
pvar_sp->m_flags |= ClangExpressionVariable::EVIsLLDBAllocated;
pvar_sp->m_flags |= ClangExpressionVariable::EVNeedsAllocation;
return pvar_sp;
}
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::AddPersistentVariable
(
const clang::NamedDecl *decl,
const ConstString &name,
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
TypeFromParser parser_type,
bool is_result,
bool is_lvalue
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
)
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_parser_vars.get());
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
ExecutionContext *exe_ctx = m_parser_vars->m_exe_ctx;
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
clang::ASTContext *context(exe_ctx->target->GetScratchClangASTContext()->getASTContext());
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
TypeFromUser user_type(ClangASTContext::CopyType(context,
Added automatically generated result variables for each expression. It is now possible to do things like this: (lldb) expr int $i = 5; $i + 1 $0 = (int) 6 (lldb) expr $i + 3 $1 = (int) 8 (lldb) expr $1 + $0 $2 = (int) 14 As a bonus, this allowed us to move printing of expression results into the ClangPersistentVariable class. This code needs a bit of refactoring -- in particular, ClangExpressionDeclMap has eaten one too many bacteria and needs to undergo mitosis -- but the infrastructure appears to be holding up nicely. llvm-svn: 110896
2010-08-12 09:56:52 +08:00
parser_type.GetASTContext(),
parser_type.GetOpaqueQualType()),
Fixed handling of explicitly-declared persistent variables. llvm-svn: 123398
2011-01-14 05:23:32 +08:00
context);
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
if (!m_parser_vars->m_persistent_vars->CreatePersistentVariable (exe_ctx->GetBestExecutionContextScope (),
name,
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
user_type,
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
exe_ctx->process->GetByteOrder(),
exe_ctx->process->GetAddressByteSize()))
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP var_sp (m_parser_vars->m_persistent_vars->GetVariable(name));
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!var_sp)
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
if (is_result)
var_sp->m_flags |= ClangExpressionVariable::EVNeedsFreezeDry;
else
var_sp->m_flags |= ClangExpressionVariable::EVKeepInTarget; // explicitly-declared persistent variables should persist
if (is_lvalue)
{
var_sp->m_flags |= ClangExpressionVariable::EVIsProgramReference;
}
else
{
var_sp->m_flags |= ClangExpressionVariable::EVIsLLDBAllocated;
var_sp->m_flags |= ClangExpressionVariable::EVNeedsAllocation;
}
if (log)
log->Printf("Created persistent variable with flags 0x%hx", var_sp->m_flags);
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
var_sp->EnableParserVars();
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
var_sp->m_parser_vars->m_named_decl = decl;
var_sp->m_parser_vars->m_parser_type = parser_type;
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return true;
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
}
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::AddValueToStruct
(
const clang::NamedDecl *decl,
const ConstString &name,
llvm::Value *value,
size_t size,
off_t alignment
)
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
assert (m_parser_vars.get());
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Fixed a bug where the parser-specific members of persistent variables were staying around too long. This caused the following problem: - A persistent result variable is created for the result of an expression. The pointer to the corresponding Decl is stored in the variable. - The persistent variable is looked up during struct generation (correctly) using its Decl. - Another expression defines a new result variable which happens to have a Decl in the same place as the original result variable. - The persistent variable is looked up during struct generation using its Decl, but the old result variable appears first in the list and has the same Decl pointer. The fix is to destroy parser-specific data when it is no longer valid. Also improved some logging as I diagnosed the bug. llvm-svn: 112540
2010-08-31 06:17:16 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_struct_vars->m_struct_laid_out = false;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
if (m_struct_members.GetVariable(decl))
return true;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP var_sp (m_found_entities.GetVariable(decl));
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!var_sp)
var_sp = m_parser_vars->m_persistent_vars->GetVariable(decl);
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!var_sp)
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
Fixed a bug where the parser-specific members of persistent variables were staying around too long. This caused the following problem: - A persistent result variable is created for the result of an expression. The pointer to the corresponding Decl is stored in the variable. - The persistent variable is looked up during struct generation (correctly) using its Decl. - Another expression defines a new result variable which happens to have a Decl in the same place as the original result variable. - The persistent variable is looked up during struct generation using its Decl, but the old result variable appears first in the list and has the same Decl pointer. The fix is to destroy parser-specific data when it is no longer valid. Also improved some logging as I diagnosed the bug. llvm-svn: 112540
2010-08-31 06:17:16 +08:00
if (log)
log->Printf("Adding value for decl %p [%s - %s] to the structure",
decl,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
name.GetCString(),
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
var_sp->GetName().GetCString());
Fixed a bug where the parser-specific members of persistent variables were staying around too long. This caused the following problem: - A persistent result variable is created for the result of an expression. The pointer to the corresponding Decl is stored in the variable. - The persistent variable is looked up during struct generation (correctly) using its Decl. - Another expression defines a new result variable which happens to have a Decl in the same place as the original result variable. - The persistent variable is looked up during struct generation using its Decl, but the old result variable appears first in the list and has the same Decl pointer. The fix is to destroy parser-specific data when it is no longer valid. Also improved some logging as I diagnosed the bug. llvm-svn: 112540
2010-08-31 06:17:16 +08:00
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
// We know entity->m_parser_vars is valid because we used a parser variable
// to find it
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
var_sp->m_parser_vars->m_llvm_value = value;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
var_sp->EnableJITVars();
var_sp->m_jit_vars->m_alignment = alignment;
var_sp->m_jit_vars->m_size = size;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
m_struct_members.AddVariable(var_sp);
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return true;
}
bool
ClangExpressionDeclMap::DoStructLayout ()
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
if (m_struct_vars->m_struct_laid_out)
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return true;
off_t cursor = 0;
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_struct_vars->m_struct_alignment = 0;
m_struct_vars->m_struct_size = 0;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
for (size_t member_index = 0, num_members = m_struct_members.GetSize();
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
member_index < num_members;
++member_index)
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP member_sp(m_struct_members.GetVariableAtIndex(member_index));
if (!member_sp)
return false;
if (!member_sp->m_jit_vars.get())
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
if (member_index == 0)
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
m_struct_vars->m_struct_alignment = member_sp->m_jit_vars->m_alignment;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (cursor % member_sp->m_jit_vars->m_alignment)
cursor += (member_sp->m_jit_vars->m_alignment - (cursor % member_sp->m_jit_vars->m_alignment));
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
member_sp->m_jit_vars->m_offset = cursor;
cursor += member_sp->m_jit_vars->m_size;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_struct_vars->m_struct_size = cursor;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_struct_vars->m_struct_laid_out = true;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return true;
}
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
bool ClangExpressionDeclMap::GetStructInfo
(
uint32_t &num_elements,
size_t &size,
off_t &alignment
)
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
if (!m_struct_vars->m_struct_laid_out)
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return false;
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
num_elements = m_struct_members.GetSize();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
size = m_struct_vars->m_struct_size;
alignment = m_struct_vars->m_struct_alignment;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return true;
}
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::GetStructElement
(
const clang::NamedDecl *&decl,
llvm::Value *&value,
off_t &offset,
ConstString &name,
uint32_t index
)
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
if (!m_struct_vars->m_struct_laid_out)
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return false;
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (index >= m_struct_members.GetSize())
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return false;
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP member_sp(m_struct_members.GetVariableAtIndex(index));
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!member_sp ||
!member_sp->m_parser_vars.get() ||
!member_sp->m_jit_vars.get())
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
decl = member_sp->m_parser_vars->m_named_decl;
value = member_sp->m_parser_vars->m_llvm_value;
offset = member_sp->m_jit_vars->m_offset;
name = member_sp->GetName();
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
"expr -i" now performs the required transforms to prepare the IR for JIT compilation. We still need to do the JIT compilation and move the arguments in/out of target memory. llvm-svn: 108279
2010-07-14 05:41:46 +08:00
return true;
}
Added support for calling functions from expressions. Right now we mock up the function as a variadic function when generating the IR for the call; we need to eventually make the function be the right type if the type is available. llvm-svn: 109543
2010-07-28 05:39:39 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::GetFunctionInfo
(
const clang::NamedDecl *decl,
llvm::Value**& value,
uint64_t &ptr
)
Added support for locating a function that is referenced in the IR. We don't yet support updating the call to that function. llvm-svn: 109483
2010-07-27 10:07:53 +08:00
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP entity_sp(m_found_entities.GetVariable(decl));
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!entity_sp)
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
Added support for locating a function that is referenced in the IR. We don't yet support updating the call to that function. llvm-svn: 109483
2010-07-27 10:07:53 +08:00
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
// We know m_parser_vars is valid since we searched for the variable by
// its NamedDecl
Added support for locating a function that is referenced in the IR. We don't yet support updating the call to that function. llvm-svn: 109483
2010-07-27 10:07:53 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
value = &entity_sp->m_parser_vars->m_llvm_value;
ptr = entity_sp->m_parser_vars->m_lldb_value->GetScalar().ULongLong();
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return true;
Added support for locating a function that is referenced in the IR. We don't yet support updating the call to that function. llvm-svn: 109483
2010-07-27 10:07:53 +08:00
}
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::GetFunctionAddress
(
const ConstString &name,
uint64_t &ptr
)
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_parser_vars.get());
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
// Back out in all cases where we're not fully initialized
Handle the case where you make a ClangExpressionDeclMap without a selected frame. llvm-svn: 121099
2010-12-07 09:56:02 +08:00
if (m_parser_vars->m_exe_ctx->target == NULL)
return false;
if (!m_parser_vars->m_sym_ctx.target_sp)
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
return false;
Did a lot of code cleanup. Fixed the DWARF plug-in such that when it gets all attributes for a DIE, that it omits the DW_AT_sibling and DW_AT_declaration when getting attributes from a DW_AT_abstract_origin or DW_AT_specification DIE. llvm-svn: 118654
2010-11-10 07:46:37 +08:00
SymbolContextList sc_list;
Changed the SymbolFile::FindFunction() function calls to only return lldb_private::Function objects. Previously the SymbolFileSymtab subclass would return lldb_private::Symbol objects when it was asked to find functions. The Module::FindFunctions (...) now take a boolean "bool include_symbols" so that the module can track down functions and symbols, yet functions are found by the SymbolFile plug-ins (through the SymbolVendor class), and symbols are gotten through the ObjectFile plug-ins. Fixed and issue where the DWARF parser might run into incomplete class member function defintions which would make clang mad when we tried to make certain member functions with invalid number of parameters (such as an operator= operator that had no parameters). Now we just avoid and don't complete these incomplete functions. llvm-svn: 124359
2011-01-27 14:44:37 +08:00
const bool include_symbols = true;
const bool append = false;
m_parser_vars->m_sym_ctx.FindFunctionsByName(name, include_symbols, append, sc_list);
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
Did a lot of code cleanup. Fixed the DWARF plug-in such that when it gets all attributes for a DIE, that it omits the DW_AT_sibling and DW_AT_declaration when getting attributes from a DW_AT_abstract_origin or DW_AT_specification DIE. llvm-svn: 118654
2010-11-10 07:46:37 +08:00
if (!sc_list.GetSize())
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
return false;
SymbolContext sym_ctx;
Did a lot of code cleanup. Fixed the DWARF plug-in such that when it gets all attributes for a DIE, that it omits the DW_AT_sibling and DW_AT_declaration when getting attributes from a DW_AT_abstract_origin or DW_AT_specification DIE. llvm-svn: 118654
2010-11-10 07:46:37 +08:00
sc_list.GetContextAtIndex(0, sym_ctx);
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
const Address *fun_address;
if (sym_ctx.function)
fun_address = &sym_ctx.function->GetAddressRange().GetBaseAddress();
else if (sym_ctx.symbol)
fun_address = &sym_ctx.symbol->GetAddressRangeRef().GetBaseAddress();
else
return false;
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
ptr = fun_address->GetLoadAddress (m_parser_vars->m_exe_ctx->target);
Added support for rewriting objc_msgSend so we can call Objective-C methods from expressions. Also added some more logging to the function-calling thread plan so that we can see the registers when a function finishes. Also documented things maybe a bit better. llvm-svn: 109938
2010-07-31 09:32:05 +08:00
return true;
}
Added support for the fragile ivars provided by Apple's Objective-C 2.0 runtime. They are enabled if the Objective-C runtime has the proper version. llvm-svn: 123694
2011-01-18 07:42:46 +08:00
bool
ClangExpressionDeclMap::GetSymbolAddress
(
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
Target &target,
Added support for the fragile ivars provided by Apple's Objective-C 2.0 runtime. They are enabled if the Objective-C runtime has the proper version. llvm-svn: 123694
2011-01-18 07:42:46 +08:00
const ConstString &name,
uint64_t &ptr
)
{
SymbolContextList sc_list;
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
target.GetImages().FindSymbolsWithNameAndType(name, eSymbolTypeAny, sc_list);
Added support for the fragile ivars provided by Apple's Objective-C 2.0 runtime. They are enabled if the Objective-C runtime has the proper version. llvm-svn: 123694
2011-01-18 07:42:46 +08:00
if (!sc_list.GetSize())
return false;
SymbolContext sym_ctx;
sc_list.GetContextAtIndex(0, sym_ctx);
const Address *sym_address = &sym_ctx.symbol->GetAddressRangeRef().GetBaseAddress();
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
ptr = sym_address->GetLoadAddress(&target);
Added support for the fragile ivars provided by Apple's Objective-C 2.0 runtime. They are enabled if the Objective-C runtime has the proper version. llvm-svn: 123694
2011-01-18 07:42:46 +08:00
return true;
}
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
bool
ClangExpressionDeclMap::GetSymbolAddress
(
const ConstString &name,
uint64_t &ptr
)
{
assert (m_parser_vars.get());
if (!m_parser_vars->m_exe_ctx ||
!m_parser_vars->m_exe_ctx->target)
return false;
return GetSymbolAddress(*m_parser_vars->m_exe_ctx->target,
name,
ptr);
}
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
// Interface for CommandObjectExpression
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::Materialize
(
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
ExecutionContext &exe_ctx,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
lldb::addr_t &struct_address,
Error &err
)
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
EnableMaterialVars();
m_material_vars->m_process = exe_ctx.process;
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
bool result = DoMaterialize(false /* dematerialize */,
exe_ctx,
LLDB_INVALID_ADDRESS /* top of stack frame */,
LLDB_INVALID_ADDRESS /* bottom of stack frame */,
NULL, /* result SP */
err);
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
if (result)
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
struct_address = m_material_vars->m_materialized_location;
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
return result;
}
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::GetObjectPointer
(
lldb::addr_t &object_ptr,
Added support for generating expressions that have access to the members of the Objective-C self object. The approach we take is to generate the method as a @category on top of the self object, and to pass the "self" pointer to it. (_cmd is currently NULL.) Most changes are in ClangExpressionDeclMap, but the change that adds support to the ABIs to pass _cmd touches a fair amount of code. llvm-svn: 121722
2010-12-14 06:46:15 +08:00
ConstString &object_name,
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
ExecutionContext &exe_ctx,
Bugfixes for the new "self" pointer handling. Specifically, the code to pass the _cmd pointer has been improved, and _cmd is now set to the value of _cmd for the current context, as opposed to being simply NULL. llvm-svn: 121739
2010-12-14 08:42:36 +08:00
Error &err,
bool suppress_type_check
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
)
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
if (!exe_ctx.frame || !exe_ctx.target || !exe_ctx.process)
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
{
err.SetErrorString("Couldn't load 'this' because the context is incomplete");
return false;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (!m_struct_vars->m_object_pointer_type.GetOpaqueQualType())
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
{
err.SetErrorString("Couldn't load 'this' because its type is unknown");
return false;
}
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
VariableSP object_ptr_var = FindVariableInScope (*exe_ctx.frame,
object_name,
(suppress_type_check ? NULL : &m_struct_vars->m_object_pointer_type));
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
if (!object_ptr_var)
{
Added support for generating expressions that have access to the members of the Objective-C self object. The approach we take is to generate the method as a @category on top of the self object, and to pass the "self" pointer to it. (_cmd is currently NULL.) Most changes are in ClangExpressionDeclMap, but the change that adds support to the ABIs to pass _cmd touches a fair amount of code. llvm-svn: 121722
2010-12-14 06:46:15 +08:00
err.SetErrorStringWithFormat("Couldn't find '%s' with appropriate type in scope", object_name.GetCString());
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
return false;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
std::auto_ptr<lldb_private::Value> location_value(GetVariableValue(exe_ctx,
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
object_ptr_var,
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
NULL));
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
if (!location_value.get())
{
Added support for generating expressions that have access to the members of the Objective-C self object. The approach we take is to generate the method as a @category on top of the self object, and to pass the "self" pointer to it. (_cmd is currently NULL.) Most changes are in ClangExpressionDeclMap, but the change that adds support to the ABIs to pass _cmd touches a fair amount of code. llvm-svn: 121722
2010-12-14 06:46:15 +08:00
err.SetErrorStringWithFormat("Couldn't get the location for '%s'", object_name.GetCString());
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
return false;
}
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
switch (location_value->GetValueType())
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
{
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
default:
err.SetErrorStringWithFormat("'%s' is not in memory; LLDB must be extended to handle registers", object_name.GetCString());
return false;
case Value::eValueTypeLoadAddress:
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
{
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
lldb::addr_t value_addr = location_value->GetScalar().ULongLong();
uint32_t address_byte_size = exe_ctx.target->GetArchitecture().GetAddressByteSize();
lldb::ByteOrder address_byte_order = exe_ctx.process->GetByteOrder();
if (ClangASTType::GetClangTypeBitWidth(m_struct_vars->m_object_pointer_type.GetASTContext(),
m_struct_vars->m_object_pointer_type.GetOpaqueQualType()) != address_byte_size * 8)
{
err.SetErrorStringWithFormat("'%s' is not of an expected pointer size", object_name.GetCString());
return false;
}
DataBufferHeap data;
data.SetByteSize(address_byte_size);
Error read_error;
if (exe_ctx.process->ReadMemory (value_addr, data.GetBytes(), address_byte_size, read_error) != address_byte_size)
{
err.SetErrorStringWithFormat("Coldn't read '%s' from the target: %s", object_name.GetCString(), read_error.AsCString());
return false;
}
DataExtractor extractor(data.GetBytes(), data.GetByteSize(), address_byte_order, address_byte_size);
uint32_t offset = 0;
object_ptr = extractor.GetPointer(&offset);
return true;
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
}
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
case Value::eValueTypeScalar:
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
{
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
if (location_value->GetContextType() != Value::eContextTypeRegisterInfo)
{
StreamString ss;
location_value->Dump(&ss);
err.SetErrorStringWithFormat("%s is a scalar of unhandled type: %s", object_name.GetCString(), ss.GetString().c_str());
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
RegisterInfo *reg_info = location_value->GetRegisterInfo();
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (!reg_info)
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
{
err.SetErrorStringWithFormat("Couldn't get the register information for %s", object_name.GetCString());
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
RegisterContext *reg_ctx = exe_ctx.GetRegisterContext();
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (!reg_ctx)
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat("Couldn't read register context to read %s from %s", object_name.GetCString(), reg_info->name);
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
uint32_t register_number = reg_info->kinds[lldb::eRegisterKindLLDB];
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
object_ptr = reg_ctx->ReadRegisterAsUnsigned(register_number, 0x0);
Fixed a bug in the expression parser where the 'this' or 'self' variable was not properly read if the compiler optimized it into a register. llvm-svn: 126973
2011-03-04 08:23:47 +08:00
return true;
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
}
}
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::Dematerialize
(
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
ExecutionContext &exe_ctx,
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP &result_sp,
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
lldb::addr_t stack_frame_top,
lldb::addr_t stack_frame_bottom,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
Error &err
)
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
return DoMaterialize(true, exe_ctx, stack_frame_top, stack_frame_bottom, &result_sp, err);
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
DidDematerialize();
}
void
ClangExpressionDeclMap::DidDematerialize()
{
if (m_material_vars.get())
{
if (m_material_vars->m_materialized_location)
{
//#define SINGLE_STEP_EXPRESSIONS
#ifndef SINGLE_STEP_EXPRESSIONS
m_material_vars->m_process->DeallocateMemory(m_material_vars->m_materialized_location);
#endif
m_material_vars->m_materialized_location = 0;
}
DisableMaterialVars();
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
}
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::DumpMaterializedStruct
(
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
ExecutionContext &exe_ctx,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
Stream &s,
Error &err
)
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
assert (m_material_vars.get());
if (!m_struct_vars->m_struct_laid_out)
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
{
err.SetErrorString("Structure hasn't been laid out yet");
return false;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (!exe_ctx.process)
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
{
err.SetErrorString("Couldn't find the process");
return false;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (!exe_ctx.target)
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
{
err.SetErrorString("Couldn't find the target");
return false;
}
Logging improvements to help identify major events in LLDB expression execution. We also now print the argument structure after execution, to allow us to verify that the expression did indeed execute correctly. llvm-svn: 121126
2010-12-07 18:00:20 +08:00
if (!m_material_vars->m_materialized_location)
{
err.SetErrorString("No materialized location");
return false;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
lldb::DataBufferSP data(new DataBufferHeap(m_struct_vars->m_struct_size, 0));
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
Error error;
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (exe_ctx.process->ReadMemory (m_material_vars->m_materialized_location, data->GetBytes(), data->GetByteSize(), error) != data->GetByteSize())
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
{
err.SetErrorStringWithFormat ("Couldn't read struct from the target: %s", error.AsCString());
return false;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
DataExtractor extractor(data, exe_ctx.process->GetByteOrder(), exe_ctx.target->GetArchitecture().GetAddressByteSize());
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
for (size_t member_idx = 0, num_members = m_struct_members.GetSize();
member_idx < num_members;
++member_idx)
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP member_sp(m_struct_members.GetVariableAtIndex(member_idx));
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!member_sp)
return false;
s.Printf("[%s]\n", member_sp->GetName().GetCString());
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!member_sp->m_jit_vars.get())
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
extractor.Dump (&s, // stream
member_sp->m_jit_vars->m_offset, // offset
lldb::eFormatBytesWithASCII, // format
1, // byte size of individual entries
member_sp->m_jit_vars->m_size, // number of entries
16, // entries per line
m_material_vars->m_materialized_location + member_sp->m_jit_vars->m_offset, // address to print
0, // bit size (bitfields only; 0 means ignore)
0); // bit alignment (bitfields only; 0 means ignore)
Added logging: - When we JIT an expression, we print the disassembly of the generated code - When we put the structure into the target, we print the individual entries in the structure byte for byte. llvm-svn: 109278
2010-07-24 06:19:18 +08:00
s.PutChar('\n');
}
return true;
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::DoMaterialize
(
bool dematerialize,
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
ExecutionContext &exe_ctx,
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
lldb::addr_t stack_frame_top,
lldb::addr_t stack_frame_bottom,
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
lldb::ClangExpressionVariableSP *result_sp_ptr,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
Error &err
)
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (result_sp_ptr)
result_sp_ptr->reset();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Added automatically generated result variables for each expression. It is now possible to do things like this: (lldb) expr int $i = 5; $i + 1 $0 = (int) 6 (lldb) expr $i + 3 $1 = (int) 8 (lldb) expr $1 + $0 $2 = (int) 14 As a bonus, this allowed us to move printing of expression results into the ClangPersistentVariable class. This code needs a bit of refactoring -- in particular, ClangExpressionDeclMap has eaten one too many bacteria and needs to undergo mitosis -- but the infrastructure appears to be holding up nicely. llvm-svn: 110896
2010-08-12 09:56:52 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (!m_struct_vars->m_struct_laid_out)
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
{
err.SetErrorString("Structure hasn't been laid out yet");
return LLDB_INVALID_ADDRESS;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (!exe_ctx.frame)
Updated the IR converter for the target to eliminate spurious guard variables on expression statics. Updated the AST result synthesizer to eliminate the unneeded result pointer. Very rudimentary expressions now evaluate correctly in the target using the new JIT-based mechanism. llvm-svn: 109317
2010-07-24 09:37:44 +08:00
{
err.SetErrorString("Received null execution frame");
return LLDB_INVALID_ADDRESS;
}
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangPersistentVariables &persistent_vars = exe_ctx.target->GetPersistentVariables();
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (!m_struct_vars->m_struct_size)
Bugfixes to the expression parser. Fixes include: - If you put a semicolon at the end of an expression, this no longer causes the expression parser to error out. This was a two-part fix: first, ClangExpressionDeclMap::Materialize now handles an empty struct (such as when there is no return value); second, ASTResultSynthesizer walks backward from the end of the ASTs until it reaches something that's not a NullStmt. - ClangExpressionVariable now properly byte-swaps when printing itself. - ClangUtilityFunction now cleans up after itself when it's done compiling itself. - Utility functions can now use external functions just like user expressions. - If you end your expression with a statement that does not return a value, the expression now runs correctly anyway. Also, added the beginnings of an Objective-C object validator function, which is neither installed nor used as yet. llvm-svn: 113789
2010-09-14 05:34:21 +08:00
{
if (log)
log->PutCString("Not bothering to allocate a struct because no arguments are needed");
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_material_vars->m_allocated_area = NULL;
Bugfixes to the expression parser. Fixes include: - If you put a semicolon at the end of an expression, this no longer causes the expression parser to error out. This was a two-part fix: first, ClangExpressionDeclMap::Materialize now handles an empty struct (such as when there is no return value); second, ASTResultSynthesizer walks backward from the end of the ASTs until it reaches something that's not a NullStmt. - ClangExpressionVariable now properly byte-swaps when printing itself. - ClangUtilityFunction now cleans up after itself when it's done compiling itself. - Utility functions can now use external functions just like user expressions. - If you end your expression with a statement that does not return a value, the expression now runs correctly anyway. Also, added the beginnings of an Objective-C object validator function, which is neither installed nor used as yet. llvm-svn: 113789
2010-09-14 05:34:21 +08:00
return true;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
const SymbolContext &sym_ctx(exe_ctx.frame->GetSymbolContext(lldb::eSymbolContextEverything));
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
if (!dematerialize)
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (m_material_vars->m_materialized_location)
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
exe_ctx.process->DeallocateMemory(m_material_vars->m_materialized_location);
m_material_vars->m_materialized_location = 0;
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
}
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
Added extra logging, and made sure that the argument struct for expressions is deallocated when the ClangExpressionDeclMap is taken down. llvm-svn: 116028
2010-10-08 09:58:41 +08:00
if (log)
log->PutCString("Allocating memory for materialized argument struct");
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
lldb::addr_t mem = exe_ctx.process->AllocateMemory(m_struct_vars->m_struct_alignment + m_struct_vars->m_struct_size,
lldb::ePermissionsReadable | lldb::ePermissionsWritable,
err);
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
if (mem == LLDB_INVALID_ADDRESS)
return false;
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_material_vars->m_allocated_area = mem;
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
}
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_material_vars->m_materialized_location = m_material_vars->m_allocated_area;
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (m_material_vars->m_materialized_location % m_struct_vars->m_struct_alignment)
m_material_vars->m_materialized_location += (m_struct_vars->m_struct_alignment - (m_material_vars->m_materialized_location % m_struct_vars->m_struct_alignment));
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
for (uint64_t member_index = 0, num_members = m_struct_members.GetSize();
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
member_index < num_members;
++member_index)
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP member_sp(m_struct_members.GetVariableAtIndex(member_index));
Allowed ClangExpressionDeclMap to dematerialize persistent variables even after the parser has finished running. llvm-svn: 120521
2010-12-01 09:29:06 +08:00
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
if (m_found_entities.ContainsVariable (member_sp))
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
{
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
RegisterInfo *reg_info = member_sp->GetRegisterInfo ();
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (reg_info)
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
// This is a register variable
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
RegisterContext *reg_ctx = exe_ctx.GetRegisterContext();
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
if (!reg_ctx)
return false;
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!DoMaterializeOneRegister (dematerialize,
exe_ctx,
*reg_ctx,
*reg_info,
m_material_vars->m_materialized_location + member_sp->m_jit_vars->m_offset,
err))
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return false;
}
else
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!member_sp->m_jit_vars.get())
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return false;
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!DoMaterializeOneVariable (dematerialize,
exe_ctx,
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
sym_ctx,
member_sp,
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
m_material_vars->m_materialized_location + member_sp->m_jit_vars->m_offset,
err))
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return false;
}
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
}
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
else
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
{
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
// No need to look for presistent variables if the name doesn't start
// with with a '$' character...
if (member_sp->GetName().AsCString ("!")[0] == '$' && persistent_vars.ContainsVariable(member_sp))
Fixed a bug where the parser-specific members of persistent variables were staying around too long. This caused the following problem: - A persistent result variable is created for the result of an expression. The pointer to the corresponding Decl is stored in the variable. - The persistent variable is looked up during struct generation (correctly) using its Decl. - Another expression defines a new result variable which happens to have a Decl in the same place as the original result variable. - The persistent variable is looked up during struct generation using its Decl, but the old result variable appears first in the list and has the same Decl pointer. The fix is to destroy parser-specific data when it is no longer valid. Also improved some logging as I diagnosed the bug. llvm-svn: 112540
2010-08-31 06:17:16 +08:00
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
if (member_sp->GetName() == m_struct_vars->m_result_name)
{
if (log)
log->PutCString("Found result member in the struct");
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
if (result_sp_ptr)
*result_sp_ptr = member_sp;
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
}
if (!DoMaterializeOnePersistentVariable (dematerialize,
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
exe_ctx,
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
member_sp,
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
m_material_vars->m_materialized_location + member_sp->m_jit_vars->m_offset,
stack_frame_top,
stack_frame_bottom,
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
err))
return false;
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
}
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
else
{
err.SetErrorStringWithFormat("Unexpected variable %s", member_sp->GetName().GetCString());
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
return false;
Put more smarts into the RegisterContext base class. Now the base class has a method: void RegisterContext::InvalidateIfNeeded (bool force); Each time this function is called, when "force" is false, it will only call the pure virtual "virtual void RegisterContext::InvalideAllRegisters()" if the register context's stop ID doesn't match that of the process. When the stop ID doesn't match, or "force" is true, the base class will clear its cached registers and the RegisterContext will update its stop ID to match that of the process. This helps make it easier to correctly flush the register context (possibly from multiple locations depending on when and where new registers are availabe) without inadvertently clearing the register cache when it doesn't need to be. Modified the ProcessGDBRemote plug-in to be much more efficient when it comes to: - caching the expedited registers in the stop reply packets (we were ignoring these before and it was causing us to read at least three registers every time we stopped that were already supplied in the stop reply packet). - When a thread has no stop reason, don't keep asking for the thread stopped info. Prior to this fix we would continually send a qThreadStopInfo packet over and over when any thread stop info was requested. We now note the stop ID that the stop info was requested for and avoid multiple requests. Cleaned up some of the expression code to not look for ClangExpressionVariable objects up by name since they are now shared pointers and we can just look for the exact pointer match and avoid possible errors. Fixed an bug in the ValueObject code that would cause children to not be displayed. llvm-svn: 123127
2011-01-10 05:07:35 +08:00
}
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
}
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
return true;
}
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
static bool WriteAddressInto
(
ExecutionContext &exe_ctx,
lldb::addr_t target,
lldb::addr_t address,
Error &err
)
{
size_t pointer_byte_size = exe_ctx.process->GetAddressByteSize();
StreamString str (0 | Stream::eBinary,
pointer_byte_size,
exe_ctx.process->GetByteOrder());
switch (pointer_byte_size)
{
default:
assert(!"Unhandled byte size");
case 4:
{
uint32_t address32 = address & 0xffffffffll;
Endian patch from Kirk Beitz that allows better cross platform building. llvm-svn: 124643
2011-02-01 09:31:41 +08:00
str.PutRawBytes(&address32, sizeof(address32), endian::InlHostByteOrder(), eByteOrderInvalid);
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
}
break;
case 8:
{
uint64_t address64 = address;
Endian patch from Kirk Beitz that allows better cross platform building. llvm-svn: 124643
2011-02-01 09:31:41 +08:00
str.PutRawBytes(&address64, sizeof(address64), endian::InlHostByteOrder(), eByteOrderInvalid);
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
}
break;
}
return (exe_ctx.process->WriteMemory (target, str.GetData(), pointer_byte_size, err) == pointer_byte_size);
}
static lldb::addr_t ReadAddressFrom
(
ExecutionContext &exe_ctx,
lldb::addr_t source,
Error &err
)
{
size_t pointer_byte_size = exe_ctx.process->GetAddressByteSize();
DataBufferHeap *buf = new DataBufferHeap(pointer_byte_size, 0);
DataBufferSP buf_sp(buf);
if (exe_ctx.process->ReadMemory (source, buf->GetBytes(), pointer_byte_size, err) != pointer_byte_size)
return LLDB_INVALID_ADDRESS;
DataExtractor extractor (buf_sp, exe_ctx.process->GetByteOrder(), exe_ctx.process->GetAddressByteSize());
uint32_t offset = 0;
return (lldb::addr_t)extractor.GetPointer(&offset);
}
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::DoMaterializeOnePersistentVariable
(
bool dematerialize,
ExecutionContext &exe_ctx,
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP &var_sp,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
lldb::addr_t addr,
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
lldb::addr_t stack_frame_top,
lldb::addr_t stack_frame_bottom,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
Error &err
)
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (!var_sp)
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
err.SetErrorString("Invalid persistent variable");
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
return LLDB_INVALID_ADDRESS;
}
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
const size_t pvar_byte_size = var_sp->GetByteSize();
First step of refactoring variable handling in the expression parser. There shouldn't be four separate classes encapsulating a variable. ClangExpressionVariable is now meant to be the container for all variable information. It has several optional components that hold data for different subsystems. ClangPersistentVariable has been removed; we now use ClangExpressionVariable instead. llvm-svn: 111600
2010-08-20 09:02:30 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
uint8_t *pvar_data = var_sp->GetValueBytes();
if (pvar_data == NULL)
First step of refactoring variable handling in the expression parser. There shouldn't be four separate classes encapsulating a variable. ClangExpressionVariable is now meant to be the container for all variable information. It has several optional components that hold data for different subsystems. ClangPersistentVariable has been removed; we now use ClangExpressionVariable instead. llvm-svn: 111600
2010-08-20 09:02:30 +08:00
return false;
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
Error error;
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
lldb::addr_t mem; // The address of a spare memory area used to hold the persistent variable.
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
if (dematerialize)
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
if (log)
log->Printf("Dematerializing persistent variable with flags 0x%hx", var_sp->m_flags);
if ((var_sp->m_flags & ClangExpressionVariable::EVIsLLDBAllocated) ||
(var_sp->m_flags & ClangExpressionVariable::EVIsProgramReference))
{
// Get the location of the target out of the struct.
Error read_error;
mem = ReadAddressFrom(exe_ctx, addr, read_error);
if (mem == LLDB_INVALID_ADDRESS)
{
err.SetErrorStringWithFormat("Couldn't read address of %s from struct: %s", var_sp->GetName().GetCString(), error.AsCString());
return false;
}
if (var_sp->m_flags & ClangExpressionVariable::EVIsProgramReference &&
!var_sp->m_live_sp)
{
// If the reference comes from the program, then the ClangExpressionVariable's
// live variable data hasn't been set up yet. Do this now.
Fix up how the ValueObjects manage their life cycle so that you can hand out a shared pointer to a ValueObject or any of its dependent ValueObjects, and the whole cluster will stay around as long as that shared pointer stays around. llvm-svn: 130035
2011-04-23 07:53:53 +08:00
var_sp->m_live_sp = ValueObjectConstResult::Create (exe_ctx.GetBestExecutionContextScope (),
var_sp->GetTypeFromUser().GetASTContext(),
var_sp->GetTypeFromUser().GetOpaqueQualType(),
var_sp->GetName(),
mem,
eAddressTypeLoad,
pvar_byte_size);
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
}
if (!var_sp->m_live_sp)
{
err.SetErrorStringWithFormat("Couldn't find the memory area used to store %s", var_sp->GetName().GetCString());
return false;
}
Fixed the LLDB build so that we can have private types, private enums and public types and public enums. This was done to keep the SWIG stuff from parsing all sorts of enums and types that weren't needed, and allows us to abstract our API better. llvm-svn: 128239
2011-03-25 05:19:54 +08:00
if (var_sp->m_live_sp->GetValue().GetValueAddressType() != eAddressTypeLoad)
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
{
err.SetErrorStringWithFormat("The address of the memory area for %s is in an incorrect format", var_sp->GetName().GetCString());
return false;
}
Fixed handling of explicitly-declared persistent variables. llvm-svn: 123398
2011-01-14 05:23:32 +08:00
if (var_sp->m_flags & ClangExpressionVariable::EVNeedsFreezeDry ||
var_sp->m_flags & ClangExpressionVariable::EVKeepInTarget)
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
{
mem = var_sp->m_live_sp->GetValue().GetScalar().ULongLong();
if (log)
log->Printf("Dematerializing %s from 0x%llx", var_sp->GetName().GetCString(), (uint64_t)mem);
// Read the contents of the spare memory area
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
var_sp->ValueUpdated ();
if (exe_ctx.process->ReadMemory (mem, pvar_data, pvar_byte_size, error) != pvar_byte_size)
{
err.SetErrorStringWithFormat ("Couldn't read a composite type from the target: %s", error.AsCString());
return false;
}
Fixed a bug in which expression-local variables were treated as being permanently resident in target memory. In fact, since the expression's stack frame is deleted and potentially re-used after the expression completes, the variables need to be treated as being freeze-dried. llvm-svn: 131104
2011-05-10 06:04:36 +08:00
if (stack_frame_top != LLDB_INVALID_ADDRESS &&
stack_frame_bottom != LLDB_INVALID_ADDRESS &&
mem >= stack_frame_bottom &&
mem <= stack_frame_top)
{
// If the variable is resident in the stack frame created by the expression,
// then it cannot be relied upon to stay around. We treat it as needing
// reallocation.
var_sp->m_flags |= ClangExpressionVariable::EVIsLLDBAllocated;
var_sp->m_flags |= ClangExpressionVariable::EVNeedsAllocation;
var_sp->m_flags &= ~ClangExpressionVariable::EVIsProgramReference;
}
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
var_sp->m_flags &= ~ClangExpressionVariable::EVNeedsFreezeDry;
}
if (var_sp->m_flags & ClangExpressionVariable::EVNeedsAllocation &&
!(var_sp->m_flags & ClangExpressionVariable::EVKeepInTarget))
{
if (m_keep_result_in_memory)
{
var_sp->m_flags |= ClangExpressionVariable::EVKeepInTarget;
}
else
{
Error deallocate_error = exe_ctx.process->DeallocateMemory(mem);
if (!err.Success())
{
err.SetErrorStringWithFormat ("Couldn't deallocate memory for %s: %s", var_sp->GetName().GetCString(), deallocate_error.AsCString());
return false;
}
}
}
}
else
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
err.SetErrorStringWithFormat("Persistent variables without separate allocations are not currently supported.");
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
return false;
}
}
else
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
if (log)
log->Printf("Materializing persistent variable with flags 0x%hx", var_sp->m_flags);
if (var_sp->m_flags & ClangExpressionVariable::EVNeedsAllocation)
{
// Allocate a spare memory area to store the persistent variable's contents.
Error allocate_error;
mem = exe_ctx.process->AllocateMemory(pvar_byte_size,
lldb::ePermissionsReadable | lldb::ePermissionsWritable,
allocate_error);
if (mem == LLDB_INVALID_ADDRESS)
{
err.SetErrorStringWithFormat("Couldn't allocate a memory area to store %s: %s", var_sp->GetName().GetCString(), allocate_error.AsCString());
return false;
}
if (log)
log->Printf("Allocated %s (0x%llx) sucessfully", var_sp->GetName().GetCString(), mem);
// Put the location of the spare memory into the live data of the ValueObject.
Fix up how the ValueObjects manage their life cycle so that you can hand out a shared pointer to a ValueObject or any of its dependent ValueObjects, and the whole cluster will stay around as long as that shared pointer stays around. llvm-svn: 130035
2011-04-23 07:53:53 +08:00
var_sp->m_live_sp = ValueObjectConstResult::Create (exe_ctx.GetBestExecutionContextScope(),
var_sp->GetTypeFromUser().GetASTContext(),
var_sp->GetTypeFromUser().GetOpaqueQualType(),
var_sp->GetName(),
mem,
eAddressTypeLoad,
pvar_byte_size);
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
// Clear the flag if the variable will never be deallocated.
if (var_sp->m_flags & ClangExpressionVariable::EVKeepInTarget)
var_sp->m_flags &= ~ClangExpressionVariable::EVNeedsAllocation;
// Write the contents of the variable to the area.
if (exe_ctx.process->WriteMemory (mem, pvar_data, pvar_byte_size, error) != pvar_byte_size)
{
err.SetErrorStringWithFormat ("Couldn't write a composite type to the target: %s", error.AsCString());
return false;
}
}
if ((var_sp->m_flags & ClangExpressionVariable::EVIsProgramReference && var_sp->m_live_sp) ||
var_sp->m_flags & ClangExpressionVariable::EVIsLLDBAllocated)
{
mem = var_sp->m_live_sp->GetValue().GetScalar().ULongLong();
// Now write the location of the area into the struct.
Error write_error;
if (!WriteAddressInto(exe_ctx, addr, mem, write_error))
{
err.SetErrorStringWithFormat ("Couldn't write %s to the target: %s", var_sp->GetName().GetCString(), write_error.AsCString());
return false;
}
if (log)
log->Printf("Materialized %s into 0x%llx", var_sp->GetName().GetCString(), (uint64_t)mem);
}
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
else if (!(var_sp->m_flags & ClangExpressionVariable::EVIsProgramReference))
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
err.SetErrorStringWithFormat("Persistent variables without separate allocations are not currently supported.");
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
return false;
}
}
return true;
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
bool
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::DoMaterializeOneVariable
(
bool dematerialize,
ExecutionContext &exe_ctx,
const SymbolContext &sym_ctx,
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
ClangExpressionVariableSP &expr_var,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
lldb::addr_t addr,
Error &err
)
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
if (!exe_ctx.frame || !exe_ctx.process)
Added code to support use of "this" and "self" in expressions. This involved three main changes: - In ClangUserExpression::ClangUserExpression(), we now insert the following lines into the expression: #define this ___clang_this #define self ___clang_self - In ClangExpressionDeclMap::GetDecls(), we special-case ___clang_(this|self) and instead look up "this" or "self" - In ClangASTSource, we introduce the capability to generate Decls with a different, overridden, name from the one that was requested, e.g. this for ___clang_this. llvm-svn: 113866
2010-09-15 05:59:34 +08:00
return false;
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
// Vital information about the value
const ConstString &name(expr_var->GetName());
TypeFromUser type(expr_var->GetTypeFromUser());
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
VariableSP var = FindVariableInScope (*exe_ctx.frame, name, &type);
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
Symbol *sym = FindGlobalDataSymbol(*exe_ctx.target, name);
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
std::auto_ptr<lldb_private::Value> location_value;
if (var)
{
location_value.reset(GetVariableValue(exe_ctx,
var,
NULL));
}
else if (sym)
{
location_value.reset(new Value);
uint64_t location_load_addr;
if (!GetSymbolAddress(*exe_ctx.target, name, location_load_addr))
{
if (log)
err.SetErrorStringWithFormat("Couldn't find value for global symbol %s", name.GetCString());
}
location_value->SetValueType(Value::eValueTypeLoadAddress);
location_value->GetScalar() = location_load_addr;
}
else
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
err.SetErrorStringWithFormat("Couldn't find %s with appropriate type", name.GetCString());
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
return false;
}
Modified TaggedASTType to inherit from ClangASTType and moved it to its own header file for cleanliness. Added more logging to ClangFunction so that we can diagnose crashes in the executing expression. Added code to extract the result of the expression from the struct that is passed to the JIT-compiled code. llvm-svn: 109199
2010-07-23 08:16:21 +08:00
if (log)
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
log->Printf("%s %s with type %p", (dematerialize ? "Dematerializing" : "Materializing"), name.GetCString(), type.GetOpaqueQualType());
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
if (!location_value.get())
{
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
err.SetErrorStringWithFormat("Couldn't get value for %s", name.GetCString());
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
return false;
}
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
// The size of the type contained in addr
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
size_t value_bit_size = ClangASTType::GetClangTypeBitWidth(type.GetASTContext(), type.GetOpaqueQualType());
size_t value_byte_size = value_bit_size % 8 ? ((value_bit_size + 8) / 8) : (value_bit_size / 8);
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
Value::ValueType value_type = location_value->GetValueType();
switch (value_type)
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
default:
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
StreamString ss;
location_value->Dump(&ss);
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
err.SetErrorStringWithFormat("%s has a value of unhandled type: %s", name.GetCString(), ss.GetString().c_str());
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
}
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
break;
case Value::eValueTypeLoadAddress:
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
if (!dematerialize)
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
lldb::addr_t value_addr = location_value->GetScalar().ULongLong();
Error error;
if (!WriteAddressInto(exe_ctx,
addr,
value_addr,
error))
{
err.SetErrorStringWithFormat ("Couldn't write %s to the target: %s", name.GetCString(), error.AsCString());
return false;
}
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
}
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
break;
case Value::eValueTypeScalar:
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
{
Modified the lldb_private::Type clang type resolving code to handle three cases when getting the clang type: - need only a forward declaration - need a clang type that can be used for layout (members and args/return types) - need a full clang type This allows us to partially parse the clang types and be as lazy as possible. The first case is when we just need to declare a type and we will complete it later. The forward declaration happens only for class/union/structs and enums. The layout type allows us to resolve the full clang type _except_ if we have any modifiers on a pointer or reference (both R and L value). In this case when we are adding members or function args or return types, we only need to know how the type will be laid out and we can defer completing the pointee type until we later need it. The last type means we need a full definition for the clang type. Did some renaming of some enumerations to get rid of the old "DC" prefix (which stands for DebugCore which is no longer around). Modified the clang namespace support to be almost ready to be fed to the expression parser. I made a new ClangNamespaceDecl class that can carry around the AST and the namespace decl so we can copy it into the expression AST. I modified the symbol vendor and symbol file plug-ins to use this new class. llvm-svn: 118976
2010-11-13 11:52:47 +08:00
if (location_value->GetContextType() != Value::eContextTypeRegisterInfo)
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
{
StreamString ss;
location_value->Dump(&ss);
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
err.SetErrorStringWithFormat("%s is a scalar of unhandled type: %s", name.GetCString(), ss.GetString().c_str());
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
lldb::addr_t reg_addr = LLDB_INVALID_ADDRESS; // The address of a spare memory area aused to hold the variable.
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
RegisterInfo *reg_info = location_value->GetRegisterInfo();
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (!reg_info)
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
{
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
err.SetErrorStringWithFormat("Couldn't get the register information for %s", name.GetCString());
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
RegisterValue reg_value;
RegisterContext *reg_ctx = exe_ctx.GetRegisterContext();
if (!reg_ctx)
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat("Couldn't read register context to read %s from %s", name.GetCString(), reg_info->name);
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
uint32_t register_byte_size = reg_info->byte_size;
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
if (dematerialize)
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
// Get the location of the spare memory area out of the variable's live data.
if (!expr_var->m_live_sp)
{
err.SetErrorStringWithFormat("Couldn't find the memory area used to store %s", name.GetCString());
return false;
}
Fixed the LLDB build so that we can have private types, private enums and public types and public enums. This was done to keep the SWIG stuff from parsing all sorts of enums and types that weren't needed, and allows us to abstract our API better. llvm-svn: 128239
2011-03-25 05:19:54 +08:00
if (expr_var->m_live_sp->GetValue().GetValueAddressType() != eAddressTypeLoad)
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
{
err.SetErrorStringWithFormat("The address of the memory area for %s is in an incorrect format", name.GetCString());
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
reg_addr = expr_var->m_live_sp->GetValue().GetScalar().ULongLong();
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err = reg_ctx->ReadRegisterValueFromMemory (reg_info, reg_addr, value_byte_size, reg_value);
if (err.Fail())
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (!reg_ctx->WriteRegister (reg_info, reg_value))
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat("Couldn't write %s to register %s", name.GetCString(), reg_info->name);
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
}
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
// Deallocate the spare area and clear the variable's live data.
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
Error deallocate_error = exe_ctx.process->DeallocateMemory(reg_addr);
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
if (!deallocate_error.Success())
{
err.SetErrorStringWithFormat("Couldn't deallocate spare memory area for %s: %s", name.GetCString(), deallocate_error.AsCString());
return false;
}
expr_var->m_live_sp.reset();
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
}
else
{
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
// Allocate a spare memory area to place the register's contents into. This memory area will be pointed to by the slot in the
// struct.
Error allocate_error;
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
reg_addr = exe_ctx.process->AllocateMemory (value_byte_size,
lldb::ePermissionsReadable | lldb::ePermissionsWritable,
allocate_error);
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (reg_addr == LLDB_INVALID_ADDRESS)
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
{
err.SetErrorStringWithFormat("Couldn't allocate a memory area to store %s: %s", name.GetCString(), allocate_error.AsCString());
return false;
}
// Put the location of the spare memory into the live data of the ValueObject.
Fix up how the ValueObjects manage their life cycle so that you can hand out a shared pointer to a ValueObject or any of its dependent ValueObjects, and the whole cluster will stay around as long as that shared pointer stays around. llvm-svn: 130035
2011-04-23 07:53:53 +08:00
expr_var->m_live_sp = ValueObjectConstResult::Create (exe_ctx.GetBestExecutionContextScope(),
type.GetASTContext(),
type.GetOpaqueQualType(),
name,
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
reg_addr,
Fix up how the ValueObjects manage their life cycle so that you can hand out a shared pointer to a ValueObject or any of its dependent ValueObjects, and the whole cluster will stay around as long as that shared pointer stays around. llvm-svn: 130035
2011-04-23 07:53:53 +08:00
eAddressTypeLoad,
value_byte_size);
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
// Now write the location of the area into the struct.
Error write_error;
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (!WriteAddressInto(exe_ctx, addr, reg_addr, write_error))
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
{
err.SetErrorStringWithFormat ("Couldn't write %s to the target: %s", name.GetCString(), write_error.AsCString());
return false;
}
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
// Moving from a register into addr
//
// Case 1: addr_byte_size and register_byte_size are the same
//
// |AABBCCDD| Register contents
// |AABBCCDD| Address contents
//
// Case 2: addr_byte_size is bigger than register_byte_size
//
// Error! (The register should always be big enough to hold the data)
//
// Case 3: register_byte_size is bigger than addr_byte_size
//
// |AABBCCDD| Register contents
// |AABB| Address contents on little-endian hardware
// |CCDD| Address contents on big-endian hardware
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
if (value_byte_size > register_byte_size)
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat("%s is too big to store in %s", name.GetCString(), reg_info->name);
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
}
uint32_t register_offset;
switch (exe_ctx.process->GetByteOrder())
{
default:
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
err.SetErrorStringWithFormat("%s is stored with an unhandled byte order", name.GetCString());
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
case lldb::eByteOrderLittle:
register_offset = 0;
break;
case lldb::eByteOrderBig:
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
register_offset = register_byte_size - value_byte_size;
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
break;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
RegisterValue reg_value;
if (!reg_ctx->ReadRegister (reg_info, reg_value))
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat("Couldn't read %s from %s", name.GetCString(), reg_info->name);
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err = reg_ctx->WriteRegisterValueToMemory(reg_info, reg_addr, value_byte_size, reg_value);
if (err.Fail())
Added support for (de)materializing values in registers, so that expressions can use them. llvm-svn: 115658
2010-10-06 04:18:48 +08:00
return false;
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
}
}
return true;
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
}
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
bool
ClangExpressionDeclMap::DoMaterializeOneRegister
(
bool dematerialize,
ExecutionContext &exe_ctx,
RegisterContext &reg_ctx,
Fixed the LLDB build so that we can have private types, private enums and public types and public enums. This was done to keep the SWIG stuff from parsing all sorts of enums and types that weren't needed, and allows us to abstract our API better. llvm-svn: 128239
2011-03-25 05:19:54 +08:00
const RegisterInfo &reg_info,
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
lldb::addr_t addr,
Error &err
)
{
uint32_t register_byte_size = reg_info.byte_size;
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
RegisterValue reg_value;
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
if (dematerialize)
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
Error read_error (reg_ctx.ReadRegisterValueFromMemory(&reg_info, addr, register_byte_size, reg_value));
if (read_error.Fail())
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
Enabled extra warnings and fixed a bunch of small issues. llvm-svn: 124250
2011-01-26 07:55:37 +08:00
err.SetErrorStringWithFormat ("Couldn't read %s from the target: %s", reg_info.name, read_error.AsCString());
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (!reg_ctx.WriteRegister (&reg_info, reg_value))
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat("Couldn't write register %s (dematerialize)", reg_info.name);
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return false;
}
}
else
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
if (!reg_ctx.ReadRegister(&reg_info, reg_value))
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat("Couldn't read %s (materialize)", reg_info.name);
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return false;
}
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
Error write_error (reg_ctx.WriteRegisterValueToMemory(&reg_info, addr, register_byte_size, reg_value));
if (write_error.Fail())
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
While implementing unwind information using UnwindAssemblyInstEmulation I ran into some cleanup I have been wanting to do when reading/writing registers. Previously all RegisterContext subclasses would need to implement: virtual bool ReadRegisterBytes (uint32_t reg, DataExtractor &data); virtual bool WriteRegisterBytes (uint32_t reg, DataExtractor &data, uint32_t data_offset = 0); There is now a new class specifically designed to hold register values: lldb_private::RegisterValue The new register context calls that subclasses must implement are: virtual bool ReadRegister (const RegisterInfo *reg_info, RegisterValue &reg_value) = 0; virtual bool WriteRegister (const RegisterInfo *reg_info, const RegisterValue &reg_value) = 0; The RegisterValue class must be big enough to handle any register value. The class contains an enumeration for the value type, and then a union for the data value. Any integer/float values are stored directly in an appropriate host integer/float. Anything bigger is stored in a byte buffer that has a length and byte order. The RegisterValue class also knows how to copy register value bytes into in a buffer with a specified byte order which can be used to write the register value down into memory, and this does the right thing when not all bytes from the register values are needed (getting a uint8 from a uint32 register value..). All RegiterContext and other sources have been switched over to using the new regiter value class. llvm-svn: 131096
2011-05-10 04:18:18 +08:00
err.SetErrorStringWithFormat ("Couldn't write %s to the target: %s", write_error.AsCString());
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return false;
}
}
return true;
}
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
lldb::VariableSP
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::FindVariableInScope
(
StackFrame &frame,
const ConstString &name,
TypeFromUser *type
)
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
{
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Added code to support use of "this" and "self" in expressions. This involved three main changes: - In ClangUserExpression::ClangUserExpression(), we now insert the following lines into the expression: #define this ___clang_this #define self ___clang_self - In ClangExpressionDeclMap::GetDecls(), we special-case ___clang_(this|self) and instead look up "this" or "self" - In ClangASTSource, we introduce the capability to generate Decls with a different, overridden, name from the one that was requested, e.g. this for ___clang_this. llvm-svn: 113866
2010-09-15 05:59:34 +08:00
VariableList *var_list = frame.GetVariableList(true);
Fixed an expression parsing issue where if you were stopped somewhere without debug information and you evaluated an expression, a crash would occur as a result of an unchecked pointer. Added the ability to get the expression path for a ValueObject. For a rectangle point child "x" the expression path would be something like: "rect.top_left.x". This will allow GUI and command lines to get ahold of the expression path for a value object without having to explicitly know about the hierarchy. This means the ValueObject base class now has a "ValueObject *m_parent;" member. All ValueObject subclasses now correctly track their lineage and are able to provide value expression paths as well. Added a new "--flat" option to the "frame variable" to allow for flat variable output. An example of the current and new outputs: (lldb) frame variable argc = 1 argv = 0x00007fff5fbffe80 pt = { x = 2 y = 3 } rect = { bottom_left = { x = 1 y = 2 } top_right = { x = 3 y = 4 } } (lldb) frame variable --flat argc = 1 argv = 0x00007fff5fbffe80 pt.x = 2 pt.y = 3 rect.bottom_left.x = 1 rect.bottom_left.y = 2 rect.top_right.x = 3 rect.top_right.y = 4 As you can see when there is a lot of hierarchy it can help flatten things out. Also if you want to use a member in an expression, you can copy the text from the "--flat" output and not have to piece it together manually. This can help when you want to use parts of the STL in expressions: (lldb) frame variable --flat argc = 1 argv = 0x00007fff5fbffea8 hello_world._M_dataplus._M_p = 0x0000000000000000 (lldb) expr hello_world._M_dataplus._M_p[0] == '\0' llvm-svn: 116532
2010-10-15 06:52:14 +08:00
if (!var_list)
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
return lldb::VariableSP();
Fixed an expression parsing issue where if you were stopped somewhere without debug information and you evaluated an expression, a crash would occur as a result of an unchecked pointer. Added the ability to get the expression path for a ValueObject. For a rectangle point child "x" the expression path would be something like: "rect.top_left.x". This will allow GUI and command lines to get ahold of the expression path for a value object without having to explicitly know about the hierarchy. This means the ValueObject base class now has a "ValueObject *m_parent;" member. All ValueObject subclasses now correctly track their lineage and are able to provide value expression paths as well. Added a new "--flat" option to the "frame variable" to allow for flat variable output. An example of the current and new outputs: (lldb) frame variable argc = 1 argv = 0x00007fff5fbffe80 pt = { x = 2 y = 3 } rect = { bottom_left = { x = 1 y = 2 } top_right = { x = 3 y = 4 } } (lldb) frame variable --flat argc = 1 argv = 0x00007fff5fbffe80 pt.x = 2 pt.y = 3 rect.bottom_left.x = 1 rect.bottom_left.y = 2 rect.top_right.x = 3 rect.top_right.y = 4 As you can see when there is a lot of hierarchy it can help flatten things out. Also if you want to use a member in an expression, you can copy the text from the "--flat" output and not have to piece it together manually. This can help when you want to use parts of the STL in expressions: (lldb) frame variable --flat argc = 1 argv = 0x00007fff5fbffea8 hello_world._M_dataplus._M_p = 0x0000000000000000 (lldb) expr hello_world._M_dataplus._M_p[0] == '\0' llvm-svn: 116532
2010-10-15 06:52:14 +08:00
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
lldb::VariableSP var_sp (var_list->FindVariable(name));
const bool append = true;
const uint32_t max_matches = 1;
if (!var_sp)
Added code to support use of "this" and "self" in expressions. This involved three main changes: - In ClangUserExpression::ClangUserExpression(), we now insert the following lines into the expression: #define this ___clang_this #define self ___clang_self - In ClangExpressionDeclMap::GetDecls(), we special-case ___clang_(this|self) and instead look up "this" or "self" - In ClangASTSource, we introduce the capability to generate Decls with a different, overridden, name from the one that was requested, e.g. this for ___clang_this. llvm-svn: 113866
2010-09-15 05:59:34 +08:00
{
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
// Look for globals elsewhere in the module for the frame
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
ModuleSP module_sp (frame.GetSymbolContext(eSymbolContextModule).module_sp);
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
if (module_sp)
{
VariableList module_globals;
if (module_sp->FindGlobalVariables (name, append, max_matches, module_globals))
var_sp = module_globals.GetVariableAtIndex (0);
}
Added code to support use of "this" and "self" in expressions. This involved three main changes: - In ClangUserExpression::ClangUserExpression(), we now insert the following lines into the expression: #define this ___clang_this #define self ___clang_self - In ClangExpressionDeclMap::GetDecls(), we special-case ___clang_(this|self) and instead look up "this" or "self" - In ClangASTSource, we introduce the capability to generate Decls with a different, overridden, name from the one that was requested, e.g. this for ___clang_this. llvm-svn: 113866
2010-09-15 05:59:34 +08:00
}
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
if (!var_sp)
Added code to support use of "this" and "self" in expressions. This involved three main changes: - In ClangUserExpression::ClangUserExpression(), we now insert the following lines into the expression: #define this ___clang_this #define self ___clang_self - In ClangExpressionDeclMap::GetDecls(), we special-case ___clang_(this|self) and instead look up "this" or "self" - In ClangASTSource, we introduce the capability to generate Decls with a different, overridden, name from the one that was requested, e.g. this for ___clang_this. llvm-svn: 113866
2010-09-15 05:59:34 +08:00
{
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
// Look for globals elsewhere in the program (all images)
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
TargetSP target_sp (frame.GetSymbolContext(eSymbolContextTarget).target_sp);
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
if (target_sp)
{
VariableList program_globals;
if (target_sp->GetImages().FindGlobalVariables (name, append, max_matches, program_globals))
var_sp = program_globals.GetVariableAtIndex (0);
}
Added code to support use of "this" and "self" in expressions. This involved three main changes: - In ClangUserExpression::ClangUserExpression(), we now insert the following lines into the expression: #define this ___clang_this #define self ___clang_self - In ClangExpressionDeclMap::GetDecls(), we special-case ___clang_(this|self) and instead look up "this" or "self" - In ClangASTSource, we introduce the capability to generate Decls with a different, overridden, name from the one that was requested, e.g. this for ___clang_this. llvm-svn: 113866
2010-09-15 05:59:34 +08:00
}
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
if (var_sp && type)
{
if (type->GetASTContext() == var_sp->GetType()->GetClangAST())
{
Clean up a bit of the type getting code where lldb_private:Type now has clang_type_t GetClangFullType(); // Get a completely defined clang type clang_type_t GetClangLayoutType(); // Get a clang type that can be used for type layout clang_type_t GetClangForwardType(); // A type that can be completed if needed, but is more efficient. llvm-svn: 125691
2011-02-17 07:00:21 +08:00
if (!ClangASTContext::AreTypesSame(type->GetASTContext(), type->GetOpaqueQualType(), var_sp->GetType()->GetClangFullType()))
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
return lldb::VariableSP();
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
}
else
{
if (log)
log->PutCString("Skipping a candidate variable because of different AST contexts");
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
return lldb::VariableSP();
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
}
}
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
return var_sp;
Added code to support use of "this" and "self" in expressions. This involved three main changes: - In ClangUserExpression::ClangUserExpression(), we now insert the following lines into the expression: #define this ___clang_this #define self ___clang_self - In ClangExpressionDeclMap::GetDecls(), we special-case ___clang_(this|self) and instead look up "this" or "self" - In ClangASTSource, we introduce the capability to generate Decls with a different, overridden, name from the one that was requested, e.g. this for ___clang_this. llvm-svn: 113866
2010-09-15 05:59:34 +08:00
}
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
Symbol *
ClangExpressionDeclMap::FindGlobalDataSymbol
(
Target &target,
const ConstString &name
)
{
SymbolContextList sc_list;
target.GetImages().FindSymbolsWithNameAndType(name,
eSymbolTypeData,
sc_list);
if (sc_list.GetSize())
{
SymbolContext sym_ctx;
sc_list.GetContextAtIndex(0, sym_ctx);
return sym_ctx.symbol;
}
return NULL;
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
// Interface for ClangASTSource
void
Did a lot of code cleanup. Fixed the DWARF plug-in such that when it gets all attributes for a DIE, that it omits the DW_AT_sibling and DW_AT_declaration when getting attributes from a DW_AT_abstract_origin or DW_AT_specification DIE. llvm-svn: 118654
2010-11-10 07:46:37 +08:00
ClangExpressionDeclMap::GetDecls (NameSearchContext &context, const ConstString &name)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_struct_vars.get());
assert (m_parser_vars.get());
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
log->Printf("Hunting for a definition for '%s'", name.GetCString());
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
// Back out in all cases where we're not fully initialized
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (m_parser_vars->m_exe_ctx->frame == NULL)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
return;
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (m_parser_vars->m_ignore_lookups)
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
{
if (log)
log->Printf("Ignoring a query during an import");
return;
}
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
Did a lot of code cleanup. Fixed the DWARF plug-in such that when it gets all attributes for a DIE, that it omits the DW_AT_sibling and DW_AT_declaration when getting attributes from a DW_AT_abstract_origin or DW_AT_specification DIE. llvm-svn: 118654
2010-11-10 07:46:37 +08:00
SymbolContextList sc_list;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
const char *name_unique_cstr = name.GetCString();
if (name_unique_cstr == NULL)
return;
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
// Only look for functions by name out in our symbols if the function
// doesn't start with our phony prefix of '$'
Just like functions can have a basename and a mangled/demangled name, variable can too. So now the lldb_private::Variable class has support for this. Variables now have support for having a basename ("i"), and a mangled name ("_ZN12_GLOBAL__N_11iE"), and a demangled name ("(anonymous namespace)::i"). Nowwhen searching for a variable by name, users might enter the fully qualified name, or just the basename. So new test functions were added to the Variable and Mangled classes as: bool NameMatches (const ConstString &name); bool NameMatches (const RegularExpression &regex); I also modified "ClangExpressionDeclMap::FindVariableInScope" to also search for global variables that are not in the current file scope by first starting with the current module, then moving on to all modules. Fixed an issue in the DWARF parser that could cause a varaible to get parsed more than once. Now, once we have parsed a VariableSP for a DIE, we cache the result even if a variable wasn't made so we don't do any re-parsing. Some DW_TAG_variable DIEs don't have locations, or are missing vital info that stops a debugger from being able to display anything for it, we parse a NULL variable shared pointer for these DIEs so we don't keep trying to reparse it. llvm-svn: 119085
2010-11-15 06:13:40 +08:00
if (name_unique_cstr[0] != '$')
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
{
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
VariableSP var = FindVariableInScope(*m_parser_vars->m_exe_ctx->frame, name);
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
// If we found a variable in scope, no need to pull up function names
if (var != NULL)
Improved function lookup to avoid conflicts between symbols with the same name and no debug information. Also improved the way functions are called so we don't automatically define them as variadic functions in the IR. llvm-svn: 113290
2010-09-08 05:49:41 +08:00
{
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
AddOneVariable(context, var);
Improved function lookup to avoid conflicts between symbols with the same name and no debug information. Also improved the way functions are called so we don't automatically define them as variadic functions in the IR. llvm-svn: 113290
2010-09-08 05:49:41 +08:00
}
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
else
Improved function lookup to avoid conflicts between symbols with the same name and no debug information. Also improved the way functions are called so we don't automatically define them as variadic functions in the IR. llvm-svn: 113290
2010-09-08 05:49:41 +08:00
{
Changed the SymbolFile::FindFunction() function calls to only return lldb_private::Function objects. Previously the SymbolFileSymtab subclass would return lldb_private::Symbol objects when it was asked to find functions. The Module::FindFunctions (...) now take a boolean "bool include_symbols" so that the module can track down functions and symbols, yet functions are found by the SymbolFile plug-ins (through the SymbolVendor class), and symbols are gotten through the ObjectFile plug-ins. Fixed and issue where the DWARF parser might run into incomplete class member function defintions which would make clang mad when we tried to make certain member functions with invalid number of parameters (such as an operator= operator that had no parameters). Now we just avoid and don't complete these incomplete functions. llvm-svn: 124359
2011-01-27 14:44:37 +08:00
const bool include_symbols = true;
const bool append = false;
m_parser_vars->m_sym_ctx.FindFunctionsByName (name,
include_symbols,
append,
sc_list);
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
if (sc_list.GetSize())
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
{
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
bool found_specific = false;
Symbol *generic_symbol = NULL;
Symbol *non_extern_symbol = NULL;
for (uint32_t index = 0, num_indices = sc_list.GetSize();
index < num_indices;
++index)
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
{
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
SymbolContext sym_ctx;
sc_list.GetContextAtIndex(index, sym_ctx);
if (sym_ctx.function)
{
// TODO only do this if it's a C function; C++ functions may be
// overloaded
if (!found_specific)
AddOneFunction(context, sym_ctx.function, NULL);
found_specific = true;
}
else if (sym_ctx.symbol)
{
if (sym_ctx.symbol->IsExternal())
generic_symbol = sym_ctx.symbol;
else
non_extern_symbol = sym_ctx.symbol;
}
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
}
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
if (!found_specific)
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
{
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
if (generic_symbol)
AddOneFunction (context, NULL, generic_symbol);
else if (non_extern_symbol)
AddOneFunction (context, NULL, non_extern_symbol);
}
ClangNamespaceDecl namespace_decl (m_parser_vars->m_sym_ctx.FindNamespace(name));
if (namespace_decl)
{
clang::NamespaceDecl *clang_namespace_decl = AddNamespace(context, namespace_decl);
if (clang_namespace_decl)
clang_namespace_decl->setHasExternalLexicalStorage();
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
}
}
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
else
Modified the lldb_private::Type clang type resolving code to handle three cases when getting the clang type: - need only a forward declaration - need a clang type that can be used for layout (members and args/return types) - need a full clang type This allows us to partially parse the clang types and be as lazy as possible. The first case is when we just need to declare a type and we will complete it later. The forward declaration happens only for class/union/structs and enums. The layout type allows us to resolve the full clang type _except_ if we have any modifiers on a pointer or reference (both R and L value). In this case when we are adding members or function args or return types, we only need to know how the type will be laid out and we can defer completing the pointee type until we later need it. The last type means we need a full definition for the clang type. Did some renaming of some enumerations to get rid of the old "DC" prefix (which stands for DebugCore which is no longer around). Modified the clang namespace support to be almost ready to be fed to the expression parser. I made a new ClangNamespaceDecl class that can carry around the AST and the namespace decl so we can copy it into the expression AST. I modified the symbol vendor and symbol file plug-ins to use this new class. llvm-svn: 118976
2010-11-13 11:52:47 +08:00
{
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
// We couldn't find a variable or function for this. Now we'll hunt for a generic
// data symbol, and -- if it is found -- treat it as a variable.
Symbol *data_symbol = FindGlobalDataSymbol(*m_parser_vars->m_exe_ctx->target, name);
if (data_symbol)
AddOneGenericVariable(context, *data_symbol);
Modified the lldb_private::Type clang type resolving code to handle three cases when getting the clang type: - need only a forward declaration - need a clang type that can be used for layout (members and args/return types) - need a full clang type This allows us to partially parse the clang types and be as lazy as possible. The first case is when we just need to declare a type and we will complete it later. The forward declaration happens only for class/union/structs and enums. The layout type allows us to resolve the full clang type _except_ if we have any modifiers on a pointer or reference (both R and L value). In this case when we are adding members or function args or return types, we only need to know how the type will be laid out and we can defer completing the pointee type until we later need it. The last type means we need a full definition for the clang type. Did some renaming of some enumerations to get rid of the old "DC" prefix (which stands for DebugCore which is no longer around). Modified the clang namespace support to be almost ready to be fed to the expression parser. I made a new ClangNamespaceDecl class that can carry around the AST and the namespace decl so we can copy it into the expression AST. I modified the symbol vendor and symbol file plug-ins to use this new class. llvm-svn: 118976
2010-11-13 11:52:47 +08:00
}
Improved function lookup to avoid conflicts between symbols with the same name and no debug information. Also improved the way functions are called so we don't automatically define them as variadic functions in the IR. llvm-svn: 113290
2010-09-08 05:49:41 +08:00
}
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
}
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
else
Updated the expression parser to ignore non-external functions it finds in libraries unless it cannot find an external function with the desired name. llvm-svn: 115721
2010-10-06 08:10:07 +08:00
{
Fixed an issue in the DWARF parser that was causing forward declarations to not get resolved. Fixed the "void **isa_ptr" variable inside the objective C verifier to start with a '$' character so we don't go looking for it in our program. Moved the lookup for "$__lldb_class" into the part that knows we are looking for internal types that start with a '$'. llvm-svn: 118488
2010-11-09 12:42:43 +08:00
static ConstString g_lldb_class_name ("$__lldb_class");
if (name == g_lldb_class_name)
{
// Clang is looking for the type of "this"
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
VariableList *vars = m_parser_vars->m_exe_ctx->frame->GetVariableList(false);
Fixed an issue in the DWARF parser that was causing forward declarations to not get resolved. Fixed the "void **isa_ptr" variable inside the objective C verifier to start with a '$' character so we don't go looking for it in our program. Moved the lookup for "$__lldb_class" into the part that knows we are looking for internal types that start with a '$'. llvm-svn: 118488
2010-11-09 12:42:43 +08:00
if (!vars)
return;
lldb::VariableSP this_var = vars->FindVariable(ConstString("this"));
if (!this_var)
return;
Type *this_type = this_var->GetType();
if (!this_type)
return;
First attempt and getting "const" C++ method function signatures correct. It currently isn't working, but it should be close. I will work on this more when I figure out what I am not doing correctly. llvm-svn: 119324
2010-11-16 10:10:54 +08:00
if (log)
{
log->PutCString ("Type for \"this\" is: ");
StreamString strm;
this_type->Dump(&strm, true);
log->PutCString (strm.GetData());
}
Clean up a bit of the type getting code where lldb_private:Type now has clang_type_t GetClangFullType(); // Get a completely defined clang type clang_type_t GetClangLayoutType(); // Get a clang type that can be used for type layout clang_type_t GetClangForwardType(); // A type that can be completed if needed, but is more efficient. llvm-svn: 125691
2011-02-17 07:00:21 +08:00
TypeFromUser this_user_type(this_type->GetClangFullType(),
Fixed an issue in the DWARF parser that was causing forward declarations to not get resolved. Fixed the "void **isa_ptr" variable inside the objective C verifier to start with a '$' character so we don't go looking for it in our program. Moved the lookup for "$__lldb_class" into the part that knows we are looking for internal types that start with a '$'. llvm-svn: 118488
2010-11-09 12:42:43 +08:00
this_type->GetClangAST());
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_struct_vars->m_object_pointer_type = this_user_type;
Fixed an issue in the DWARF parser that was causing forward declarations to not get resolved. Fixed the "void **isa_ptr" variable inside the objective C verifier to start with a '$' character so we don't go looking for it in our program. Moved the lookup for "$__lldb_class" into the part that knows we are looking for internal types that start with a '$'. llvm-svn: 118488
2010-11-09 12:42:43 +08:00
void *pointer_target_type;
if (!ClangASTContext::IsPointerType(this_user_type.GetOpaqueQualType(),
&pointer_target_type))
return;
TypeFromUser class_user_type(pointer_target_type,
this_type->GetClangAST());
Added support for generating expressions that have access to the members of the Objective-C self object. The approach we take is to generate the method as a @category on top of the self object, and to pass the "self" pointer to it. (_cmd is currently NULL.) Most changes are in ClangExpressionDeclMap, but the change that adds support to the ABIs to pass _cmd touches a fair amount of code. llvm-svn: 121722
2010-12-14 06:46:15 +08:00
if (log)
{
StreamString type_stream;
class_user_type.DumpTypeCode(&type_stream);
type_stream.Flush();
log->Printf("Adding type for $__lldb_class: %s", type_stream.GetString().c_str());
}
Fixed an issue in the DWARF parser that was causing forward declarations to not get resolved. Fixed the "void **isa_ptr" variable inside the objective C verifier to start with a '$' character so we don't go looking for it in our program. Moved the lookup for "$__lldb_class" into the part that knows we are looking for internal types that start with a '$'. llvm-svn: 118488
2010-11-09 12:42:43 +08:00
AddOneType(context, class_user_type, true);
return;
}
Added support for generating expressions that have access to the members of the Objective-C self object. The approach we take is to generate the method as a @category on top of the self object, and to pass the "self" pointer to it. (_cmd is currently NULL.) Most changes are in ClangExpressionDeclMap, but the change that adds support to the ABIs to pass _cmd touches a fair amount of code. llvm-svn: 121722
2010-12-14 06:46:15 +08:00
static ConstString g_lldb_objc_class_name ("$__lldb_objc_class");
if (name == g_lldb_objc_class_name)
{
// Clang is looking for the type of "*self"
VariableList *vars = m_parser_vars->m_exe_ctx->frame->GetVariableList(false);
if (!vars)
return;
lldb::VariableSP self_var = vars->FindVariable(ConstString("self"));
if (!self_var)
return;
Type *self_type = self_var->GetType();
if (!self_type)
return;
Clean up a bit of the type getting code where lldb_private:Type now has clang_type_t GetClangFullType(); // Get a completely defined clang type clang_type_t GetClangLayoutType(); // Get a clang type that can be used for type layout clang_type_t GetClangForwardType(); // A type that can be completed if needed, but is more efficient. llvm-svn: 125691
2011-02-17 07:00:21 +08:00
TypeFromUser self_user_type(self_type->GetClangFullType(),
Added support for generating expressions that have access to the members of the Objective-C self object. The approach we take is to generate the method as a @category on top of the self object, and to pass the "self" pointer to it. (_cmd is currently NULL.) Most changes are in ClangExpressionDeclMap, but the change that adds support to the ABIs to pass _cmd touches a fair amount of code. llvm-svn: 121722
2010-12-14 06:46:15 +08:00
self_type->GetClangAST());
m_struct_vars->m_object_pointer_type = self_user_type;
void *pointer_target_type;
if (!ClangASTContext::IsPointerType(self_user_type.GetOpaqueQualType(),
&pointer_target_type))
return;
TypeFromUser class_user_type(pointer_target_type,
self_type->GetClangAST());
if (log)
{
StreamString type_stream;
class_user_type.DumpTypeCode(&type_stream);
type_stream.Flush();
log->Printf("Adding type for $__lldb_objc_class: %s", type_stream.GetString().c_str());
}
AddOneType(context, class_user_type, false);
return;
}
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP pvar_sp(m_parser_vars->m_persistent_vars->GetVariable(name));
Updated the expression parser to ignore non-external functions it finds in libraries unless it cannot find an external function with the desired name. llvm-svn: 115721
2010-10-06 08:10:07 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
if (pvar_sp)
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
AddOneVariable(context, pvar_sp);
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
return;
}
const char *reg_name(&name.GetCString()[1]);
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
if (m_parser_vars->m_exe_ctx->GetRegisterContext())
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
Fixed the LLDB build so that we can have private types, private enums and public types and public enums. This was done to keep the SWIG stuff from parsing all sorts of enums and types that weren't needed, and allows us to abstract our API better. llvm-svn: 128239
2011-03-25 05:19:54 +08:00
const RegisterInfo *reg_info(m_parser_vars->m_exe_ctx->GetRegisterContext()->GetRegisterInfoByName(reg_name));
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
if (reg_info)
AddOneRegister(context, reg_info);
}
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
}
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
lldb::TypeSP type_sp (m_parser_vars->m_sym_ctx.FindTypeByName (name));
Added type lookup, so variables with user-defined types can be allocated and manipulated. llvm-svn: 114928
2010-09-28 07:54:58 +08:00
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
if (type_sp)
{
if (log)
Added type lookup, so variables with user-defined types can be allocated and manipulated. llvm-svn: 114928
2010-09-28 07:54:58 +08:00
{
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
log->Printf ("Matching type found for \"%s\": ", name.GetCString());
StreamString strm;
type_sp->Dump(&strm, true);
log->PutCString (strm.GetData());
}
First attempt and getting "const" C++ method function signatures correct. It currently isn't working, but it should be close. I will work on this more when I figure out what I am not doing correctly. llvm-svn: 119324
2010-11-16 10:10:54 +08:00
Clean up a bit of the type getting code where lldb_private:Type now has clang_type_t GetClangFullType(); // Get a completely defined clang type clang_type_t GetClangLayoutType(); // Get a clang type that can be used for type layout clang_type_t GetClangForwardType(); // A type that can be completed if needed, but is more efficient. llvm-svn: 125691
2011-02-17 07:00:21 +08:00
TypeFromUser user_type (type_sp->GetClangFullType(),
A few of the issue I have been trying to track down and fix have been due to the way LLDB lazily gets complete definitions for types within the debug info. When we run across a class/struct/union definition in the DWARF, we will only parse the full definition if we need to. This works fine for top level types that are assigned directly to variables and arguments, but when we have a variable with a class, lets say "A" for this example, that has a member: "B *m_b". Initially we don't need to hunt down a definition for this class unless we are ever asked to do something with it ("expr m_b->getDecl()" for example). With my previous approach to lazy type completion, we would be able to take a "A *a" and get a complete type for it, but we wouldn't be able to then do an "a->m_b->getDecl()" unless we always expanded all types within a class prior to handing out the type. Expanding everything is very costly and it would be great if there were a better way. A few months ago I worked with the llvm/clang folks to have the ExternalASTSource class be able to complete classes if there weren't completed yet: class ExternalASTSource { .... virtual void CompleteType (clang::TagDecl *Tag); virtual void CompleteType (clang::ObjCInterfaceDecl *Class); }; This was great, because we can now have the class that is producing the AST (SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources and the object that creates the forward declaration types can now also complete them anywhere within the clang type system. This patch makes a few major changes: - lldb_private::Module classes now own the AST context. Previously the TypeList objects did. - The DWARF parsers now sign up as an external AST sources so they can complete types. - All of the pure clang type system wrapper code we have in LLDB (ClangASTContext, ClangASTType, and more) can now be iterating through children of any type, and if a class/union/struct type (clang::RecordType or ObjC interface) is found that is incomplete, we can ask the AST to get the definition. - The SymbolFileDWARFDebugMap class now will create and use a single AST that all child SymbolFileDWARF classes will share (much like what happens when we have a complete linked DWARF for an executable). We will need to modify some of the ClangUserExpression code to take more advantage of this completion ability in the near future. Meanwhile we should be better off now that we can be accessing any children of variables through pointers and always be able to resolve the clang type if needed. llvm-svn: 123613
2011-01-17 11:46:26 +08:00
type_sp->GetClangAST());
Added type lookup, so variables with user-defined types can be allocated and manipulated. llvm-svn: 114928
2010-09-28 07:54:58 +08:00
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
AddOneType(context, user_type, false);
Added type lookup, so variables with user-defined types can be allocated and manipulated. llvm-svn: 114928
2010-09-28 07:54:58 +08:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
Value *
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
ClangExpressionDeclMap::GetVariableValue
(
ExecutionContext &exe_ctx,
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
VariableSP var,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
clang::ASTContext *parser_ast_context,
TypeFromUser *user_type,
TypeFromParser *parser_type
)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
{
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Updated the expression parser to use proper logging when looking for external variables. Also cleaned up the log messages coming from the DWARF interpreter. llvm-svn: 106613
2010-06-23 08:47:48 +08:00
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
Type *var_type = var->GetType();
if (!var_type)
{
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
log->PutCString("Skipped a definition because it has no type");
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
return NULL;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
Clean up a bit of the type getting code where lldb_private:Type now has clang_type_t GetClangFullType(); // Get a completely defined clang type clang_type_t GetClangLayoutType(); // Get a clang type that can be used for type layout clang_type_t GetClangForwardType(); // A type that can be completed if needed, but is more efficient. llvm-svn: 125691
2011-02-17 07:00:21 +08:00
clang_type_t var_opaque_type = var_type->GetClangFullType();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
if (!var_opaque_type)
{
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
log->PutCString("Skipped a definition because it has no Clang type");
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
return NULL;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
A few of the issue I have been trying to track down and fix have been due to the way LLDB lazily gets complete definitions for types within the debug info. When we run across a class/struct/union definition in the DWARF, we will only parse the full definition if we need to. This works fine for top level types that are assigned directly to variables and arguments, but when we have a variable with a class, lets say "A" for this example, that has a member: "B *m_b". Initially we don't need to hunt down a definition for this class unless we are ever asked to do something with it ("expr m_b->getDecl()" for example). With my previous approach to lazy type completion, we would be able to take a "A *a" and get a complete type for it, but we wouldn't be able to then do an "a->m_b->getDecl()" unless we always expanded all types within a class prior to handing out the type. Expanding everything is very costly and it would be great if there were a better way. A few months ago I worked with the llvm/clang folks to have the ExternalASTSource class be able to complete classes if there weren't completed yet: class ExternalASTSource { .... virtual void CompleteType (clang::TagDecl *Tag); virtual void CompleteType (clang::ObjCInterfaceDecl *Class); }; This was great, because we can now have the class that is producing the AST (SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources and the object that creates the forward declaration types can now also complete them anywhere within the clang type system. This patch makes a few major changes: - lldb_private::Module classes now own the AST context. Previously the TypeList objects did. - The DWARF parsers now sign up as an external AST sources so they can complete types. - All of the pure clang type system wrapper code we have in LLDB (ClangASTContext, ClangASTType, and more) can now be iterating through children of any type, and if a class/union/struct type (clang::RecordType or ObjC interface) is found that is incomplete, we can ask the AST to get the definition. - The SymbolFileDWARFDebugMap class now will create and use a single AST that all child SymbolFileDWARF classes will share (much like what happens when we have a complete linked DWARF for an executable). We will need to modify some of the ClangUserExpression code to take more advantage of this completion ability in the near future. Meanwhile we should be better off now that we can be accessing any children of variables through pointers and always be able to resolve the clang type if needed. llvm-svn: 123613
2011-01-17 11:46:26 +08:00
clang::ASTContext *ast = var_type->GetClangASTContext().getASTContext();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
A few of the issue I have been trying to track down and fix have been due to the way LLDB lazily gets complete definitions for types within the debug info. When we run across a class/struct/union definition in the DWARF, we will only parse the full definition if we need to. This works fine for top level types that are assigned directly to variables and arguments, but when we have a variable with a class, lets say "A" for this example, that has a member: "B *m_b". Initially we don't need to hunt down a definition for this class unless we are ever asked to do something with it ("expr m_b->getDecl()" for example). With my previous approach to lazy type completion, we would be able to take a "A *a" and get a complete type for it, but we wouldn't be able to then do an "a->m_b->getDecl()" unless we always expanded all types within a class prior to handing out the type. Expanding everything is very costly and it would be great if there were a better way. A few months ago I worked with the llvm/clang folks to have the ExternalASTSource class be able to complete classes if there weren't completed yet: class ExternalASTSource { .... virtual void CompleteType (clang::TagDecl *Tag); virtual void CompleteType (clang::ObjCInterfaceDecl *Class); }; This was great, because we can now have the class that is producing the AST (SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources and the object that creates the forward declaration types can now also complete them anywhere within the clang type system. This patch makes a few major changes: - lldb_private::Module classes now own the AST context. Previously the TypeList objects did. - The DWARF parsers now sign up as an external AST sources so they can complete types. - All of the pure clang type system wrapper code we have in LLDB (ClangASTContext, ClangASTType, and more) can now be iterating through children of any type, and if a class/union/struct type (clang::RecordType or ObjC interface) is found that is incomplete, we can ask the AST to get the definition. - The SymbolFileDWARFDebugMap class now will create and use a single AST that all child SymbolFileDWARF classes will share (much like what happens when we have a complete linked DWARF for an executable). We will need to modify some of the ClangUserExpression code to take more advantage of this completion ability in the near future. Meanwhile we should be better off now that we can be accessing any children of variables through pointers and always be able to resolve the clang type if needed. llvm-svn: 123613
2011-01-17 11:46:26 +08:00
if (!ast)
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
{
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
log->PutCString("There is no AST context for the current execution context");
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
return NULL;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
DWARFExpression &var_location_expr = var->LocationExpression();
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
std::auto_ptr<Value> var_location(new Value);
Looking at some of the test suite failures in DWARF in .o files with the debug map showed that the location lists in the .o files needed some refactoring in order to work. The case that was failing was where a function that was in the "__TEXT.__textcoal_nt" in the .o file, and in the "__TEXT.__text" section in the main executable. This made symbol lookup fail due to the way we were finding a real address in the debug map which was by finding the section that the function was in in the .o file and trying to find this in the main executable. Now the section list supports finding a linked address in a section or any child sections. After fixing this, we ran into issue that were due to DWARF and how it represents locations lists. DWARF makes a list of address ranges and expressions that go along with those address ranges. The location addresses are expressed in terms of a compile unit address + offset. This works fine as long as nothing moves around. When stuff moves around and offsets change between the remapped compile unit base address and the new function address, then we can run into trouble. To deal with this, we now store supply a location list slide amount to any location list expressions that will allow us to make the location list addresses into zero based offsets from the object that owns the location list (always a function in our case). With these fixes we can now re-link random address ranges inside the debugger for use with our DWARF + debug map, incremental linking, and more. Another issue that arose when doing the DWARF in the .o files was that GCC 4.2 emits a ".debug_aranges" that only mentions functions that are externally visible. This makes .debug_aranges useless to us and we now generate a real address range lookup table in the DWARF parser at the same time as we index the name tables (that are needed because .debug_pubnames is just as useless). llvm-gcc doesn't generate a .debug_aranges section, though this could be fixed, we aren't going to rely upon it. Renamed a bunch of "UINT_MAX" to "UINT32_MAX". llvm-svn: 113829
2010-09-14 10:20:48 +08:00
lldb::addr_t loclist_base_load_addr = LLDB_INVALID_ADDRESS;
if (var_location_expr.IsLocationList())
{
SymbolContext var_sc;
var->CalculateSymbolContext (&var_sc);
Moved the section load list up into the target so we can use the target to symbolicate things without the need for a valid process subclass. llvm-svn: 113895
2010-09-15 07:36:40 +08:00
loclist_base_load_addr = var_sc.function->GetAddressRange().GetBaseAddress().GetLoadAddress (exe_ctx.target);
Looking at some of the test suite failures in DWARF in .o files with the debug map showed that the location lists in the .o files needed some refactoring in order to work. The case that was failing was where a function that was in the "__TEXT.__textcoal_nt" in the .o file, and in the "__TEXT.__text" section in the main executable. This made symbol lookup fail due to the way we were finding a real address in the debug map which was by finding the section that the function was in in the .o file and trying to find this in the main executable. Now the section list supports finding a linked address in a section or any child sections. After fixing this, we ran into issue that were due to DWARF and how it represents locations lists. DWARF makes a list of address ranges and expressions that go along with those address ranges. The location addresses are expressed in terms of a compile unit address + offset. This works fine as long as nothing moves around. When stuff moves around and offsets change between the remapped compile unit base address and the new function address, then we can run into trouble. To deal with this, we now store supply a location list slide amount to any location list expressions that will allow us to make the location list addresses into zero based offsets from the object that owns the location list (always a function in our case). With these fixes we can now re-link random address ranges inside the debugger for use with our DWARF + debug map, incremental linking, and more. Another issue that arose when doing the DWARF in the .o files was that GCC 4.2 emits a ".debug_aranges" that only mentions functions that are externally visible. This makes .debug_aranges useless to us and we now generate a real address range lookup table in the DWARF parser at the same time as we index the name tables (that are needed because .debug_pubnames is just as useless). llvm-gcc doesn't generate a .debug_aranges section, though this could be fixed, we aren't going to rely upon it. Renamed a bunch of "UINT_MAX" to "UINT32_MAX". llvm-svn: 113829
2010-09-14 10:20:48 +08:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
Error err;
Enabled extra warnings and fixed a bunch of small issues. llvm-svn: 124250
2011-01-26 07:55:37 +08:00
if (!var_location_expr.Evaluate(&exe_ctx, ast, NULL, NULL, NULL, loclist_base_load_addr, NULL, *var_location.get(), &err))
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
{
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
log->Printf("Error evaluating location: %s", err.AsCString());
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
return NULL;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
A few of the issue I have been trying to track down and fix have been due to the way LLDB lazily gets complete definitions for types within the debug info. When we run across a class/struct/union definition in the DWARF, we will only parse the full definition if we need to. This works fine for top level types that are assigned directly to variables and arguments, but when we have a variable with a class, lets say "A" for this example, that has a member: "B *m_b". Initially we don't need to hunt down a definition for this class unless we are ever asked to do something with it ("expr m_b->getDecl()" for example). With my previous approach to lazy type completion, we would be able to take a "A *a" and get a complete type for it, but we wouldn't be able to then do an "a->m_b->getDecl()" unless we always expanded all types within a class prior to handing out the type. Expanding everything is very costly and it would be great if there were a better way. A few months ago I worked with the llvm/clang folks to have the ExternalASTSource class be able to complete classes if there weren't completed yet: class ExternalASTSource { .... virtual void CompleteType (clang::TagDecl *Tag); virtual void CompleteType (clang::ObjCInterfaceDecl *Class); }; This was great, because we can now have the class that is producing the AST (SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources and the object that creates the forward declaration types can now also complete them anywhere within the clang type system. This patch makes a few major changes: - lldb_private::Module classes now own the AST context. Previously the TypeList objects did. - The DWARF parsers now sign up as an external AST sources so they can complete types. - All of the pure clang type system wrapper code we have in LLDB (ClangASTContext, ClangASTType, and more) can now be iterating through children of any type, and if a class/union/struct type (clang::RecordType or ObjC interface) is found that is incomplete, we can ask the AST to get the definition. - The SymbolFileDWARFDebugMap class now will create and use a single AST that all child SymbolFileDWARF classes will share (much like what happens when we have a complete linked DWARF for an executable). We will need to modify some of the ClangUserExpression code to take more advantage of this completion ability in the near future. Meanwhile we should be better off now that we can be accessing any children of variables through pointers and always be able to resolve the clang type if needed. llvm-svn: 123613
2011-01-17 11:46:26 +08:00
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
void *type_to_use;
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
if (parser_ast_context)
{
A few of the issue I have been trying to track down and fix have been due to the way LLDB lazily gets complete definitions for types within the debug info. When we run across a class/struct/union definition in the DWARF, we will only parse the full definition if we need to. This works fine for top level types that are assigned directly to variables and arguments, but when we have a variable with a class, lets say "A" for this example, that has a member: "B *m_b". Initially we don't need to hunt down a definition for this class unless we are ever asked to do something with it ("expr m_b->getDecl()" for example). With my previous approach to lazy type completion, we would be able to take a "A *a" and get a complete type for it, but we wouldn't be able to then do an "a->m_b->getDecl()" unless we always expanded all types within a class prior to handing out the type. Expanding everything is very costly and it would be great if there were a better way. A few months ago I worked with the llvm/clang folks to have the ExternalASTSource class be able to complete classes if there weren't completed yet: class ExternalASTSource { .... virtual void CompleteType (clang::TagDecl *Tag); virtual void CompleteType (clang::ObjCInterfaceDecl *Class); }; This was great, because we can now have the class that is producing the AST (SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources and the object that creates the forward declaration types can now also complete them anywhere within the clang type system. This patch makes a few major changes: - lldb_private::Module classes now own the AST context. Previously the TypeList objects did. - The DWARF parsers now sign up as an external AST sources so they can complete types. - All of the pure clang type system wrapper code we have in LLDB (ClangASTContext, ClangASTType, and more) can now be iterating through children of any type, and if a class/union/struct type (clang::RecordType or ObjC interface) is found that is incomplete, we can ask the AST to get the definition. - The SymbolFileDWARFDebugMap class now will create and use a single AST that all child SymbolFileDWARF classes will share (much like what happens when we have a complete linked DWARF for an executable). We will need to modify some of the ClangUserExpression code to take more advantage of this completion ability in the near future. Meanwhile we should be better off now that we can be accessing any children of variables through pointers and always be able to resolve the clang type if needed. llvm-svn: 123613
2011-01-17 11:46:26 +08:00
type_to_use = GuardedCopyType(parser_ast_context, ast, var_opaque_type);
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
Made GetVariableValue() more robust in the face of failures in the AST importer. Also ensured that a variable will not be blindly added if GetVariableValue() returns an error. llvm-svn: 119889
2010-11-20 10:19:29 +08:00
if (!type_to_use)
{
if (log)
log->Printf("Couldn't copy a variable's type into the parser's AST context");
return NULL;
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
if (parser_type)
*parser_type = TypeFromParser(type_to_use, parser_ast_context);
}
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
else
type_to_use = var_opaque_type;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
if (var_location.get()->GetContextType() == Value::eContextTypeInvalid)
Modified the lldb_private::Type clang type resolving code to handle three cases when getting the clang type: - need only a forward declaration - need a clang type that can be used for layout (members and args/return types) - need a full clang type This allows us to partially parse the clang types and be as lazy as possible. The first case is when we just need to declare a type and we will complete it later. The forward declaration happens only for class/union/structs and enums. The layout type allows us to resolve the full clang type _except_ if we have any modifiers on a pointer or reference (both R and L value). In this case when we are adding members or function args or return types, we only need to know how the type will be laid out and we can defer completing the pointee type until we later need it. The last type means we need a full definition for the clang type. Did some renaming of some enumerations to get rid of the old "DC" prefix (which stands for DebugCore which is no longer around). Modified the clang namespace support to be almost ready to be fed to the expression parser. I made a new ClangNamespaceDecl class that can carry around the AST and the namespace decl so we can copy it into the expression AST. I modified the symbol vendor and symbol file plug-ins to use this new class. llvm-svn: 118976
2010-11-13 11:52:47 +08:00
var_location.get()->SetContext(Value::eContextTypeClangType, type_to_use);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
if (var_location.get()->GetValueType() == Value::eValueTypeFileAddress)
{
SymbolContext var_sc;
var->CalculateSymbolContext(&var_sc);
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
if (!var_sc.module_sp)
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
return NULL;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
ObjectFile *object_file = var_sc.module_sp->GetObjectFile();
if (!object_file)
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
return NULL;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
Address so_addr(var_location->GetScalar().ULongLong(), object_file->GetSectionList());
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
lldb::addr_t load_addr = so_addr.GetLoadAddress(exe_ctx.target);
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
var_location->GetScalar() = load_addr;
var_location->SetValueType(Value::eValueTypeLoadAddress);
}
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
if (user_type)
A few of the issue I have been trying to track down and fix have been due to the way LLDB lazily gets complete definitions for types within the debug info. When we run across a class/struct/union definition in the DWARF, we will only parse the full definition if we need to. This works fine for top level types that are assigned directly to variables and arguments, but when we have a variable with a class, lets say "A" for this example, that has a member: "B *m_b". Initially we don't need to hunt down a definition for this class unless we are ever asked to do something with it ("expr m_b->getDecl()" for example). With my previous approach to lazy type completion, we would be able to take a "A *a" and get a complete type for it, but we wouldn't be able to then do an "a->m_b->getDecl()" unless we always expanded all types within a class prior to handing out the type. Expanding everything is very costly and it would be great if there were a better way. A few months ago I worked with the llvm/clang folks to have the ExternalASTSource class be able to complete classes if there weren't completed yet: class ExternalASTSource { .... virtual void CompleteType (clang::TagDecl *Tag); virtual void CompleteType (clang::ObjCInterfaceDecl *Class); }; This was great, because we can now have the class that is producing the AST (SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources and the object that creates the forward declaration types can now also complete them anywhere within the clang type system. This patch makes a few major changes: - lldb_private::Module classes now own the AST context. Previously the TypeList objects did. - The DWARF parsers now sign up as an external AST sources so they can complete types. - All of the pure clang type system wrapper code we have in LLDB (ClangASTContext, ClangASTType, and more) can now be iterating through children of any type, and if a class/union/struct type (clang::RecordType or ObjC interface) is found that is incomplete, we can ask the AST to get the definition. - The SymbolFileDWARFDebugMap class now will create and use a single AST that all child SymbolFileDWARF classes will share (much like what happens when we have a complete linked DWARF for an executable). We will need to modify some of the ClangUserExpression code to take more advantage of this completion ability in the near future. Meanwhile we should be better off now that we can be accessing any children of variables through pointers and always be able to resolve the clang type if needed. llvm-svn: 123613
2011-01-17 11:46:26 +08:00
*user_type = TypeFromUser(var_opaque_type, ast);
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
return var_location.release();
}
void
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
ClangExpressionDeclMap::AddOneVariable (NameSearchContext &context, VariableSP var)
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_parser_vars.get());
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
Added functionality to dematerialize values that were used by the JIT compiled expression, including the result of the expression. Also added a new class, ASTType, which encapsulates an opaque Clang type and its associated AST context. Refactored ClangExpressionDeclMap to use ASTTypes, significantly reducing the possibility of mixups of types from different AST contexts. llvm-svn: 108965
2010-07-21 07:31:16 +08:00
TypeFromUser ut;
TypeFromParser pt;
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
Value *var_location = GetVariableValue (*m_parser_vars->m_exe_ctx,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
var,
context.GetASTContext(),
&ut,
&pt);
Added the necessary code to copy variables used by an expression into the struct prepared for the JIT compiled code to use. llvm-svn: 108596
2010-07-17 08:43:37 +08:00
Made GetVariableValue() more robust in the face of failures in the AST importer. Also ensured that a variable will not be blindly added if GetVariableValue() returns an error. llvm-svn: 119889
2010-11-20 10:19:29 +08:00
if (!var_location)
return;
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
NamedDecl *var_decl = context.AddVarDecl(ClangASTContext::CreateLValueReferenceType(pt.GetASTContext(), pt.GetOpaqueQualType()));
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
std::string decl_name(context.m_decl_name.getAsString());
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ConstString entity_name(decl_name.c_str());
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
ClangExpressionVariableSP entity(m_found_entities.CreateVariable (m_parser_vars->m_exe_ctx->GetBestExecutionContextScope (),
entity_name,
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ut,
m_parser_vars->m_exe_ctx->process->GetByteOrder(),
m_parser_vars->m_exe_ctx->process->GetAddressByteSize()));
assert (entity.get());
entity->EnableParserVars();
entity->m_parser_vars->m_parser_type = pt;
entity->m_parser_vars->m_named_decl = var_decl;
entity->m_parser_vars->m_llvm_value = NULL;
entity->m_parser_vars->m_lldb_value = var_location;
Made expressions that are just casts of pointer variables be evaluated statically. Also fixed a bug that caused the results of statically-evaluated expressions to be materialized improperly. This bug also removes some duplicate code. llvm-svn: 131042
2011-05-07 09:06:41 +08:00
entity->m_parser_vars->m_lldb_var = var;
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
{
Improved logging and cleaned up a redundant return statement. Now when ClangExpressionDeclMap returns a variable for a name, it pretty-prints that variable to the log instead of printing a (fairly useless) NamedDecl pointer. llvm-svn: 117972
2010-11-02 07:22:47 +08:00
std::string var_decl_print_string;
llvm::raw_string_ostream var_decl_print_stream(var_decl_print_string);
var_decl->print(var_decl_print_stream);
var_decl_print_stream.flush();
log->Printf("Found variable %s, returned %s", decl_name.c_str(), var_decl_print_string.c_str());
Added a new class, ASTDumper, that provides verbose diagnostics of Clang AST classes for the purpose of debugging the types LLDB produces for DWARF objects. The ASTDumper is currently only used in log output if you enable verbose mode in the expression log: log enable -v lldb expr Its output then appears in the log for external variables used by the expr command. llvm-svn: 124703
2011-02-02 07:43:26 +08:00
if (log->GetVerbose())
{
StreamString var_decl_dump_string;
ASTDumper::DumpDecl(var_decl_dump_string, var_decl);
log->Printf("%s\n", var_decl_dump_string.GetData());
}
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
}
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
}
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
void
ClangExpressionDeclMap::AddOneVariable(NameSearchContext &context,
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionVariableSP &pvar_sp)
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
{
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Fixed a bug where the parser-specific members of persistent variables were staying around too long. This caused the following problem: - A persistent result variable is created for the result of an expression. The pointer to the corresponding Decl is stored in the variable. - The persistent variable is looked up during struct generation (correctly) using its Decl. - Another expression defines a new result variable which happens to have a Decl in the same place as the original result variable. - The persistent variable is looked up during struct generation using its Decl, but the old result variable appears first in the list and has the same Decl pointer. The fix is to destroy parser-specific data when it is no longer valid. Also improved some logging as I diagnosed the bug. llvm-svn: 112540
2010-08-31 06:17:16 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
TypeFromUser user_type (pvar_sp->GetTypeFromUser());
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
TypeFromParser parser_type (GuardedCopyType(context.GetASTContext(),
user_type.GetASTContext(),
user_type.GetOpaqueQualType()),
context.GetASTContext());
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
Implemented a major overhaul of the way variables are handled by LLDB. Instead of being materialized into the input structure passed to the expression, variables are left in place and pointers to them are materialzied into the structure. Variables not resident in memory (notably, registers) get temporary memory regions allocated for them. Persistent variables are the most complex part of this, because they are made in various ways and there are different expectations about their lifetime. Persistent variables now have flags indicating their status and what the expectations for longevity are. They can be marked as residing in target memory permanently -- this is the default for result variables from expressions entered on the command line and for explicitly declared persistent variables (but more on that below). Other result variables have their memory freed. Some major improvements resulting from this include being able to properly take the address of variables, better and cleaner support for functions that return references, and cleaner C++ support in general. One problem that remains is the problem of explicitly declared persistent variables; I have not yet implemented the code that makes references to them into indirect references, so currently materialization and dematerialization of these variables is broken. llvm-svn: 123371
2011-01-13 16:53:35 +08:00
NamedDecl *var_decl = context.AddVarDecl(ClangASTContext::CreateLValueReferenceType(parser_type.GetASTContext(), parser_type.GetOpaqueQualType()));
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
pvar_sp->EnableParserVars();
pvar_sp->m_parser_vars->m_parser_type = parser_type;
pvar_sp->m_parser_vars->m_named_decl = var_decl;
pvar_sp->m_parser_vars->m_llvm_value = NULL;
pvar_sp->m_parser_vars->m_lldb_value = NULL;
Fixed a bug where the parser-specific members of persistent variables were staying around too long. This caused the following problem: - A persistent result variable is created for the result of an expression. The pointer to the corresponding Decl is stored in the variable. - The persistent variable is looked up during struct generation (correctly) using its Decl. - Another expression defines a new result variable which happens to have a Decl in the same place as the original result variable. - The persistent variable is looked up during struct generation using its Decl, but the old result variable appears first in the list and has the same Decl pointer. The fix is to destroy parser-specific data when it is no longer valid. Also improved some logging as I diagnosed the bug. llvm-svn: 112540
2010-08-31 06:17:16 +08:00
if (log)
Improved logging and cleaned up a redundant return statement. Now when ClangExpressionDeclMap returns a variable for a name, it pretty-prints that variable to the log instead of printing a (fairly useless) NamedDecl pointer. llvm-svn: 117972
2010-11-02 07:22:47 +08:00
{
std::string var_decl_print_string;
llvm::raw_string_ostream var_decl_print_stream(var_decl_print_string);
var_decl->print(var_decl_print_stream);
var_decl_print_stream.flush();
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
log->Printf("Added pvar %s, returned %s", pvar_sp->GetName().GetCString(), var_decl_print_string.c_str());
Improved logging and cleaned up a redundant return statement. Now when ClangExpressionDeclMap returns a variable for a name, it pretty-prints that variable to the log instead of printing a (fairly useless) NamedDecl pointer. llvm-svn: 117972
2010-11-02 07:22:47 +08:00
}
Added support for persistent variables to the expression parser. It is now possible to type: (lldb) expr int $i = 5; $i + 1 (int) 6 (lldb) expr $i + 2 (int) 7 The skeleton for automatic result variables is also implemented. The changes affect: - the process, which now contains a ClangPersistentVariables object that holds persistent variables associated with it - the expression parser, which now uses the persistent variables during variable lookup - TaggedASTType, where I loaded some commonly used tags into a header so that they are interchangeable between different clients of the class llvm-svn: 110777
2010-08-11 11:57:18 +08:00
}
Added support for reading untyped symbols. Right now they are treated as pointers of type (void*). This allows reading of environ, for instance. llvm-svn: 131063
2011-05-08 10:21:26 +08:00
void
ClangExpressionDeclMap::AddOneGenericVariable(NameSearchContext &context,
Symbol &symbol)
{
assert(m_parser_vars.get());
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
clang::ASTContext *scratch_ast_context = m_parser_vars->m_exe_ctx->target->GetScratchClangASTContext()->getASTContext();
TypeFromUser user_type (ClangASTContext::GetVoidPtrType(scratch_ast_context, false),
scratch_ast_context);
TypeFromParser parser_type (ClangASTContext::GetVoidPtrType(context.GetASTContext(), false),
context.GetASTContext());
NamedDecl *var_decl = context.AddVarDecl(ClangASTContext::CreateLValueReferenceType(parser_type.GetASTContext(), parser_type.GetOpaqueQualType()));
std::string decl_name(context.m_decl_name.getAsString());
ConstString entity_name(decl_name.c_str());
ClangExpressionVariableSP entity(m_found_entities.CreateVariable (m_parser_vars->m_exe_ctx->GetBestExecutionContextScope (),
entity_name,
user_type,
m_parser_vars->m_exe_ctx->process->GetByteOrder(),
m_parser_vars->m_exe_ctx->process->GetAddressByteSize()));
assert (entity.get());
entity->EnableParserVars();
std::auto_ptr<Value> symbol_location(new Value);
AddressRange &symbol_range = symbol.GetAddressRangeRef();
Address &symbol_address = symbol_range.GetBaseAddress();
lldb::addr_t symbol_load_addr = symbol_address.GetLoadAddress(m_parser_vars->m_exe_ctx->target);
symbol_location->SetContext(Value::eContextTypeClangType, user_type.GetOpaqueQualType());
symbol_location->GetScalar() = symbol_load_addr;
symbol_location->SetValueType(Value::eValueTypeLoadAddress);
entity->m_parser_vars->m_parser_type = parser_type;
entity->m_parser_vars->m_named_decl = var_decl;
entity->m_parser_vars->m_llvm_value = NULL;
entity->m_parser_vars->m_lldb_value = symbol_location.release();
entity->m_parser_vars->m_lldb_sym = &symbol;
if (log)
{
std::string var_decl_print_string;
llvm::raw_string_ostream var_decl_print_stream(var_decl_print_string);
var_decl->print(var_decl_print_stream);
var_decl_print_stream.flush();
log->Printf("Found variable %s, returned %s", decl_name.c_str(), var_decl_print_string.c_str());
if (log->GetVerbose())
{
StreamString var_decl_dump_string;
ASTDumper::DumpDecl(var_decl_dump_string, var_decl);
log->Printf("%s\n", var_decl_dump_string.GetData());
}
}
}
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
void
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
ClangExpressionDeclMap::AddOneRegister (NameSearchContext &context,
const RegisterInfo *reg_info)
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
{
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
void *ast_type = ClangASTContext::GetBuiltinTypeForEncodingAndBitSize(context.GetASTContext(),
reg_info->encoding,
reg_info->byte_size * 8);
if (!ast_type)
{
log->Printf("Tried to add a type for %s, but couldn't get one", context.m_decl_name.getAsString().c_str());
return;
}
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
TypeFromParser parser_type (ast_type,
context.GetASTContext());
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
NamedDecl *var_decl = context.AddVarDecl(parser_type.GetOpaqueQualType());
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
ClangExpressionVariableSP entity(m_found_entities.CreateVariable (m_parser_vars->m_exe_ctx->GetBestExecutionContextScope(),
m_parser_vars->m_exe_ctx->process->GetByteOrder(),
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
m_parser_vars->m_exe_ctx->process->GetAddressByteSize()));
assert (entity.get());
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
std::string decl_name(context.m_decl_name.getAsString());
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
entity->SetName (ConstString (decl_name.c_str()));
entity->SetRegisterInfo (reg_info);
entity->EnableParserVars();
entity->m_parser_vars->m_parser_type = parser_type;
entity->m_parser_vars->m_named_decl = var_decl;
entity->m_parser_vars->m_llvm_value = NULL;
entity->m_parser_vars->m_lldb_value = NULL;
Added a feature where registers can be referred to using special $-variables from expressions. (lldb) expr $rip These variables are available for reading and writing. llvm-svn: 120367
2010-11-30 08:27:43 +08:00
if (log)
{
std::string var_decl_print_string;
llvm::raw_string_ostream var_decl_print_stream(var_decl_print_string);
var_decl->print(var_decl_print_stream);
var_decl_print_stream.flush();
log->Printf("Added register %s, returned %s", context.m_decl_name.getAsString().c_str(), var_decl_print_string.c_str());
}
}
Modified the lldb_private::Type clang type resolving code to handle three cases when getting the clang type: - need only a forward declaration - need a clang type that can be used for layout (members and args/return types) - need a full clang type This allows us to partially parse the clang types and be as lazy as possible. The first case is when we just need to declare a type and we will complete it later. The forward declaration happens only for class/union/structs and enums. The layout type allows us to resolve the full clang type _except_ if we have any modifiers on a pointer or reference (both R and L value). In this case when we are adding members or function args or return types, we only need to know how the type will be laid out and we can defer completing the pointee type until we later need it. The last type means we need a full definition for the clang type. Did some renaming of some enumerations to get rid of the old "DC" prefix (which stands for DebugCore which is no longer around). Modified the clang namespace support to be almost ready to be fed to the expression parser. I made a new ClangNamespaceDecl class that can carry around the AST and the namespace decl so we can copy it into the expression AST. I modified the symbol vendor and symbol file plug-ins to use this new class. llvm-svn: 118976
2010-11-13 11:52:47 +08:00
clang::NamespaceDecl *
ClangExpressionDeclMap::AddNamespace (NameSearchContext &context, const ClangNamespaceDecl &namespace_decl)
{
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
clang::Decl *copied_decl = ClangASTContext::CopyDecl (context.GetASTContext(),
namespace_decl.GetASTContext(),
namespace_decl.GetNamespaceDecl());
return dyn_cast<clang::NamespaceDecl>(copied_decl);
}
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
void
ClangExpressionDeclMap::AddOneFunction(NameSearchContext &context,
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
Function* fun,
Symbol* symbol)
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_parser_vars.get());
Modified all logging calls to hand out shared pointers to make sure we don't crash if we disable logging when some code already has a copy of the logger. Prior to this fix, logs were handed out as pointers and if they were held onto while a log got disabled, then it could cause a crash. Now all logs are handed out as shared pointers so this problem shouldn't happen anymore. We are also using our new shared pointers that put the shared pointer count and the object into the same allocation for a tad better performance. llvm-svn: 118319
2010-11-06 09:53:30 +08:00
lldb::LogSP log(lldb_private::GetLogIfAllCategoriesSet (LIBLLDB_LOG_EXPRESSIONS));
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
NamedDecl *fun_decl;
std::auto_ptr<Value> fun_location(new Value);
const Address *fun_address;
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
// only valid for Functions, not for Symbols
void *fun_opaque_type = NULL;
clang::ASTContext *fun_ast_context = NULL;
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
if (fun)
{
Type *fun_type = fun->GetType();
if (!fun_type)
{
if (log)
log->PutCString("Skipped a function because it has no type");
return;
}
Clean up a bit of the type getting code where lldb_private:Type now has clang_type_t GetClangFullType(); // Get a completely defined clang type clang_type_t GetClangLayoutType(); // Get a clang type that can be used for type layout clang_type_t GetClangForwardType(); // A type that can be completed if needed, but is more efficient. llvm-svn: 125691
2011-02-17 07:00:21 +08:00
fun_opaque_type = fun_type->GetClangFullType();
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
if (!fun_opaque_type)
{
if (log)
log->PutCString("Skipped a function because it has no Clang type");
return;
}
fun_address = &fun->GetAddressRange().GetBaseAddress();
A few of the issue I have been trying to track down and fix have been due to the way LLDB lazily gets complete definitions for types within the debug info. When we run across a class/struct/union definition in the DWARF, we will only parse the full definition if we need to. This works fine for top level types that are assigned directly to variables and arguments, but when we have a variable with a class, lets say "A" for this example, that has a member: "B *m_b". Initially we don't need to hunt down a definition for this class unless we are ever asked to do something with it ("expr m_b->getDecl()" for example). With my previous approach to lazy type completion, we would be able to take a "A *a" and get a complete type for it, but we wouldn't be able to then do an "a->m_b->getDecl()" unless we always expanded all types within a class prior to handing out the type. Expanding everything is very costly and it would be great if there were a better way. A few months ago I worked with the llvm/clang folks to have the ExternalASTSource class be able to complete classes if there weren't completed yet: class ExternalASTSource { .... virtual void CompleteType (clang::TagDecl *Tag); virtual void CompleteType (clang::ObjCInterfaceDecl *Class); }; This was great, because we can now have the class that is producing the AST (SymbolFileDWARF and SymbolFileDWARFDebugMap) sign up as external AST sources and the object that creates the forward declaration types can now also complete them anywhere within the clang type system. This patch makes a few major changes: - lldb_private::Module classes now own the AST context. Previously the TypeList objects did. - The DWARF parsers now sign up as an external AST sources so they can complete types. - All of the pure clang type system wrapper code we have in LLDB (ClangASTContext, ClangASTType, and more) can now be iterating through children of any type, and if a class/union/struct type (clang::RecordType or ObjC interface) is found that is incomplete, we can ask the AST to get the definition. - The SymbolFileDWARFDebugMap class now will create and use a single AST that all child SymbolFileDWARF classes will share (much like what happens when we have a complete linked DWARF for an executable). We will need to modify some of the ClangUserExpression code to take more advantage of this completion ability in the near future. Meanwhile we should be better off now that we can be accessing any children of variables through pointers and always be able to resolve the clang type if needed. llvm-svn: 123613
2011-01-17 11:46:26 +08:00
fun_ast_context = fun_type->GetClangASTContext().getASTContext();
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
void *copied_type = GuardedCopyType(context.GetASTContext(), fun_ast_context, fun_opaque_type);
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
fun_decl = context.AddFunDecl(copied_type);
}
else if (symbol)
{
fun_address = &symbol->GetAddressRangeRef().GetBaseAddress();
fun_decl = context.AddGenericFunDecl();
}
else
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
{
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
Changed SymbolContext so when you search for functions it returns a list of functions as a SymbolContextList. Rewrote the clients of SymbolContext to use this SymbolContextList. Rewrote some of the providers of the data to SymbolContext to make them respect preferences as to whether the list should be cleared first; propagated that change out. ClangExpressionDeclMap and ClangASTSource use this new function list to properly generate function definitions - even for functions that don't have a prototype in the debug information. llvm-svn: 109476
2010-07-27 08:55:47 +08:00
log->PutCString("AddOneFunction called with no function and no symbol");
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
return;
}
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
lldb::addr_t load_addr = fun_address->GetLoadAddress(m_parser_vars->m_exe_ctx->target);
Added support to the expression parser for locating externally-defined functions. llvm-svn: 106606
2010-06-23 07:46:24 +08:00
fun_location->SetValueType(Value::eValueTypeLoadAddress);
fun_location->GetScalar() = load_addr;
Convert ValueObject to explicitly maintain the Execution Context in which they were created, and then use that when they update themselves. That means all the ValueObject evaluate me type functions that used to require a Frame object now do not. I didn't remove the SBValue API's that take this now useless frame, but I added ones that don't require the frame, and marked the SBFrame taking ones as deprecated. llvm-svn: 128593
2011-03-31 08:19:25 +08:00
ClangExpressionVariableSP entity(m_found_entities.CreateVariable (m_parser_vars->m_exe_ctx->GetBestExecutionContextScope (),
m_parser_vars->m_exe_ctx->process->GetByteOrder(),
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
m_parser_vars->m_exe_ctx->process->GetAddressByteSize()));
assert (entity.get());
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
std::string decl_name(context.m_decl_name.getAsString());
Modified LLDB expressions to not have to JIT and run code just to see variable values or persistent expression variables. Now if an expression consists of a value that is a child of a variable, or of a persistent variable only, we will create a value object for it and make a ValueObjectConstResult from it to freeze the value (for program variables only, not persistent variables) and avoid running JITed code. For everything else we still parse up and JIT code and run it in the inferior. There was also a lot of clean up in the expression code. I made the ClangExpressionVariables be stored in collections of shared pointers instead of in collections of objects. This will help stop a lot of copy constructors on these large objects and also cleans up the code considerably. The persistent clang expression variables were moved over to the Target to ensure they persist across process executions. Added the ability for lldb_private::Target objects to evaluate expressions. We want to evaluate expressions at the target level in case we aren't running yet, or we have just completed running. We still want to be able to access the persistent expression variables between runs, and also evaluate constant expressions. Added extra logging to the dynamic loader plug-in for MacOSX. ModuleList objects can now dump their contents with the UUID, arch and full paths being logged with appropriate prefix values. Thread hardened the Communication class a bit by making the connection auto_ptr member into a shared pointer member and then making a local copy of the shared pointer in each method that uses it to make sure another thread can't nuke the connection object while it is being used by another thread. Added a new file to the lldb/test/load_unload test that causes the test a.out file to link to the libd.dylib file all the time. This will allow us to test using the DYLD_LIBRARY_PATH environment variable after moving libd.dylib somewhere else. llvm-svn: 121745
2010-12-14 10:59:59 +08:00
entity->SetName(ConstString(decl_name.c_str()));
entity->SetClangType (fun_opaque_type);
entity->SetClangAST (fun_ast_context);
entity->EnableParserVars();
entity->m_parser_vars->m_named_decl = fun_decl;
entity->m_parser_vars->m_llvm_value = NULL;
entity->m_parser_vars->m_lldb_value = fun_location.release();
Refactored ClangExpressionDeclMap to use ClangExpressionVariables for found external variables as well as for struct members, replacing the Tuple and StructMember data structures. llvm-svn: 111859
2010-08-24 07:09:38 +08:00
Wrote the code that looks at a context to see if the variables in that context allow a particular JIT compiled expression to run in that context. llvm-svn: 108485
2010-07-16 08:09:46 +08:00
if (log)
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
{
Improved logging and cleaned up a redundant return statement. Now when ClangExpressionDeclMap returns a variable for a name, it pretty-prints that variable to the log instead of printing a (fairly useless) NamedDecl pointer. llvm-svn: 117972
2010-11-02 07:22:47 +08:00
std::string fun_decl_print_string;
llvm::raw_string_ostream fun_decl_print_stream(fun_decl_print_string);
fun_decl->print(fun_decl_print_stream);
fun_decl_print_stream.flush();
log->Printf("Found %s function %s, returned %s", (fun ? "specific" : "generic"), decl_name.c_str(), fun_decl_print_string.c_str());
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
}
Initial checkin of lldb code from internal Apple repo. llvm-svn: 105619
2010-06-09 00:52:24 +08:00
}
Added support for accessing members of C++ objects, including superclass members. This involved ensuring that access control was ignored, and ensuring that the operands of BitCasts were properly scanned for variables that needed importing. Also laid the groundwork for declaring objects of custom types; however, this functionality is disabled for now because of a potential loop in ASTImporter. llvm-svn: 110174
2010-08-04 09:02:13 +08:00
void
ClangExpressionDeclMap::AddOneType(NameSearchContext &context,
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
TypeFromUser &ut,
bool add_method)
Added support for accessing members of C++ objects, including superclass members. This involved ensuring that access control was ignored, and ensuring that the operands of BitCasts were properly scanned for variables that needed importing. Also laid the groundwork for declaring objects of custom types; however, this functionality is disabled for now because of a potential loop in ASTImporter. llvm-svn: 110174
2010-08-04 09:02:13 +08:00
{
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
clang::ASTContext *parser_ast_context = context.GetASTContext();
clang::ASTContext *user_ast_context = ut.GetASTContext();
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
void *copied_type = GuardedCopyType(parser_ast_context, user_ast_context, ut.GetOpaqueQualType());
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
TypeFromParser parser_type(copied_type, parser_ast_context);
Added support for accessing members of C++ objects, including superclass members. This involved ensuring that access control was ignored, and ensuring that the operands of BitCasts were properly scanned for variables that needed importing. Also laid the groundwork for declaring objects of custom types; however, this functionality is disabled for now because of a potential loop in ASTImporter. llvm-svn: 110174
2010-08-04 09:02:13 +08:00
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
if (add_method && ClangASTContext::IsAggregateType(copied_type))
{
void *args[1];
args[0] = ClangASTContext::GetVoidPtrType(parser_ast_context, false);
void *method_type = ClangASTContext::CreateFunctionType (parser_ast_context,
ClangASTContext::GetBuiltInType_void(parser_ast_context),
args,
1,
false,
ClangASTContext::GetTypeQualifiers(copied_type));
Make C++ constructors and destructors correctly within the clang types we generate from DWARF. llvm-svn: 115268
2010-10-01 10:31:07 +08:00
Added the ability to create an objective C method for an objective C interface in ClangASTContext. Also added two bool returning functions that indicated if an opaque clang qual type is a CXX class type, and if it is an ObjC class type. Objective C classes now will get their methods added lazily as they are encountered. The reason for this is currently, unlike C++, the DW_TAG_structure_type and owns the ivars, doesn't not also contain the member functions. This means when we parse the objective C class interface we either need to find all functions whose names start with "+[CLASS_NAME" or "-[CLASS_NAME" and add them all to the class, or when we parse each objective C function, we slowly add it to the class interface definition. Since objective C's class doesn't change internal bits according to whether it has certain types of member functions (like C++ does if it has virtual functions, or if it has user ctors/dtors), I currently chose to lazily populate the class when each functions is parsed. Another issue we run into with ObjC method declarations is the "self" and "_cmd" implicit args are not marked as artificial in the DWARF (DW_AT_artifical), so we currently have to look for the parameters by name if we are trying to omit artificial function args if the language of the compile unit is ObjC or ObjC++. llvm-svn: 114722
2010-09-24 13:15:53 +08:00
const bool is_virtual = false;
const bool is_static = false;
const bool is_inline = false;
Make C++ constructors and destructors correctly within the clang types we generate from DWARF. llvm-svn: 115268
2010-10-01 10:31:07 +08:00
const bool is_explicit = false;
Added the ability to create an objective C method for an objective C interface in ClangASTContext. Also added two bool returning functions that indicated if an opaque clang qual type is a CXX class type, and if it is an ObjC class type. Objective C classes now will get their methods added lazily as they are encountered. The reason for this is currently, unlike C++, the DW_TAG_structure_type and owns the ivars, doesn't not also contain the member functions. This means when we parse the objective C class interface we either need to find all functions whose names start with "+[CLASS_NAME" or "-[CLASS_NAME" and add them all to the class, or when we parse each objective C function, we slowly add it to the class interface definition. Since objective C's class doesn't change internal bits according to whether it has certain types of member functions (like C++ does if it has virtual functions, or if it has user ctors/dtors), I currently chose to lazily populate the class when each functions is parsed. Another issue we run into with ObjC method declarations is the "self" and "_cmd" implicit args are not marked as artificial in the DWARF (DW_AT_artifical), so we currently have to look for the parameters by name if we are trying to omit artificial function args if the language of the compile unit is ObjC or ObjC++. llvm-svn: 114722
2010-09-24 13:15:53 +08:00
ClangASTContext::AddMethodToCXXRecordType (parser_ast_context,
copied_type,
Made many ConstString functions inlined in the header file. Changed all of our synthesized "___clang" functions, types and variables that get used in expressions over to have a prefix of "$_lldb". Now when we do name lookups we can easily switch off of the first '$' character to know if we should look through only our internal (when first char is '$') stuff, or when we should look through program variables, functions and types. Converted all of the clang expression code over to using "const ConstString&" values for names instead of "const char *" since there were many places that were converting the "const char *" names into ConstString names and them throwing them away. We now avoid making a lot of ConstString conversions and benefit from the quick comparisons in a few extra spots. Converted a lot of code from LLVM coding conventions into LLDB coding conventions. llvm-svn: 116634
2010-10-16 06:48:33 +08:00
"$__lldb_expr",
Added the ability to create an objective C method for an objective C interface in ClangASTContext. Also added two bool returning functions that indicated if an opaque clang qual type is a CXX class type, and if it is an ObjC class type. Objective C classes now will get their methods added lazily as they are encountered. The reason for this is currently, unlike C++, the DW_TAG_structure_type and owns the ivars, doesn't not also contain the member functions. This means when we parse the objective C class interface we either need to find all functions whose names start with "+[CLASS_NAME" or "-[CLASS_NAME" and add them all to the class, or when we parse each objective C function, we slowly add it to the class interface definition. Since objective C's class doesn't change internal bits according to whether it has certain types of member functions (like C++ does if it has virtual functions, or if it has user ctors/dtors), I currently chose to lazily populate the class when each functions is parsed. Another issue we run into with ObjC method declarations is the "self" and "_cmd" implicit args are not marked as artificial in the DWARF (DW_AT_artifical), so we currently have to look for the parameters by name if we are trying to omit artificial function args if the language of the compile unit is ObjC or ObjC++. llvm-svn: 114722
2010-09-24 13:15:53 +08:00
method_type,
lldb::eAccessPublic,
is_virtual,
is_static,
Make C++ constructors and destructors correctly within the clang types we generate from DWARF. llvm-svn: 115268
2010-10-01 10:31:07 +08:00
is_inline,
is_explicit);
Removed the hacky "#define this ___clang_this" handler for C++ classes. Replaced it with a less hacky approach: - If an expression is defined in the context of a method of class A, then that expression is wrapped as ___clang_class::___clang_expr(void*) { ... } instead of ___clang_expr(void*) { ... }. - ___clang_class is resolved as the type of the target of the "this" pointer in the method the expression is defined in. - When reporting the type of ___clang_class, a method with the signature ___clang_expr(void*) is added to that class, so that Clang doesn't complain about a method being defined without a corresponding declaration. - Whenever the expression gets called, "this" gets looked up, type-checked, and then passed in as the first argument. This required the following changes: - The ABIs were changed to support passing of the "this" pointer as part of trivial calls. - ThreadPlanCallFunction and ClangFunction were changed to support passing of an optional "this" pointer. - ClangUserExpression was extended to perform the wrapping described above. - ClangASTSource was changed to revert the changes required by the hack. - ClangExpressionParser, IRForTarget, and ClangExpressionDeclMap were changed to handle different manglings of ___clang_expr flexibly. This meant no longer searching for a function called ___clang_expr, but rather looking for a function whose name *contains* ___clang_expr. - ClangExpressionParser and ClangExpressionDeclMap now remember whether "this" is required, and know how to look it up as necessary. A few inheritance bugs remain, and I'm trying to resolve these. But it is now possible to use "this" as well as refer implicitly to member variables, when in the proper context. llvm-svn: 114384
2010-09-21 08:44:12 +08:00
}
Added support for accessing members of C++ objects, including superclass members. This involved ensuring that access control was ignored, and ensuring that the operands of BitCasts were properly scanned for variables that needed importing. Also laid the groundwork for declaring objects of custom types; however, this functionality is disabled for now because of a potential loop in ASTImporter. llvm-svn: 110174
2010-08-04 09:02:13 +08:00
context.AddTypeDecl(copied_type);
}
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
void *
ClangExpressionDeclMap::GuardedCopyType (ASTContext *dest_context,
ASTContext *source_context,
void *clang_type)
{
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
assert (m_parser_vars.get());
m_parser_vars->m_ignore_lookups = true;
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
void *ret = ClangASTContext::CopyType (dest_context,
source_context,
clang_type);
Fixed object lifetimes in ClangExpressionDeclMap so that it is not referring to potentially stale state during IR execution. This was done by introducing modular state (like ClangExpressionVariable) where groups of state variables have well-defined lifetimes: - m_parser_vars are specific to parsing, and only exist between calls to WillParse() and DidParse(). - m_struct_vars survive for the entire execution of the ClangExpressionDeclMap because they provide the template for a materialized set of expression variables. - m_material_vars are specific to a single instance of materialization, and only exist between calls to Materialize() and Dematerialize(). I also removed unnecessary references to long- lived state that really didn't need to be referred to at all, and also introduced several assert()s that helped me diagnose a few bugs (fixed too). llvm-svn: 120778
2010-12-03 09:38:59 +08:00
m_parser_vars->m_ignore_lookups = false;
Modifications to type handling logic. We no longer perform recursive type lookups, because these are not required for full type fidelity. We also make the SelectorTable last for the full lifetime of the Clang compiler; this was the source of many bugs. llvm-svn: 119835
2010-11-20 04:20:02 +08:00
return ret;
}