Renamed format "signed decimal" to be "decimal". "unsigned decimal" remains unchanged:
- the name "signed decimal" was interfering with symbol %S (use summary) in summary strings.
because of the way summary strings are implemented, this did not really lead to a bug, but
simply to performing more steps than necessary to display a summary. this is fixed.
Documentation improvements (more on synthetic children, some information on filters). This is still a WIP.
llvm-svn: 138384
e.g. you may get "foo_class @ 0x123456" when typing "type summary add -f ${var} foo_class"
- Added a new special formatting token %T for summaries. This shows the type of the object.
Using it, the new "type @ location" summary could be manually generated by writing ${var%T} @ ${var%L}
- Bits and pieces required to support "frame variable array[n-m]"
The feature is not enabled yet because some additional design and support code is required, but the basics
are getting there
- Fixed a potential issue where a ValueObjectSyntheticFilter was not holding on to its SyntheticChildrenSP
Because of the way VOSF are being built now, this has never been an actual issue, but it is still sensible for
a VOSF to hold on to the SyntheticChildrenSP as well as to its FrontEnd
llvm-svn: 138080
- reorganizing classes layout to have public part first
Typedefs that we want to keep private, but must be defined for some public code to work correctly are an exception
- avoiding methods in the form T foo() { code; } all on one-line
- moving method implementations from .h to .cpp whenever feasible
Templatized code is an exception and so are very small methods
- generally, adhering to coding conventions followed project-wide
Functional changes:
- fixed an issue where using ${var} in a summary for an aggregate, and then displaying a pointer-to-aggregate would lead to no summary being displayed
The issue was not a major one because all ${var} was meant to do in that context was display an error for invalid use of pointer
Accordingly fixed test cases and added a new test case
llvm-svn: 137944
- all instances of "vobj" have been renamed to "valobj"
- class Debugger::Formatting has been renamed to DataVisualization (defined in FormatManager.h/cpp)
The interface to this class has not changed
- FormatCategory now uses ConstString's as keys to the navigators instead of repeatedly casting
from ConstString to const char* and back all the time
Next step is making the same happen for categories themselves
- category gnu-libstdc++ is defined in the constructor for a FormatManager
The source code for it is defined in gnu_libstdcpp.py, drawn from examples/synthetic at compile time
All references to previous 'osxcpp' name have been removed from both code and file names
Functional changes:
- the name of the option to use a summary string for 'type summary add' has changed from the previous --format-string
to the new --summary-string. It is expected that the short option will change from -f to -s, and -s for --python-script
will become -o
llvm-svn: 137886
The category is enabled by default. If you run into issues with it, disable it and the previous behavior of LLDB is restored
** This is a temporary solution. The general solution to having formatters pulled in at startup should involve going through the Platform.
Fixed an issue in type synthetic list where a category with synthetic providers in it was not shown if all the providers were regex-based
llvm-svn: 137850
The converse is also true: an error is shown when the user tries to add a synthetic provider to a category that already has a filter for the same type
llvm-svn: 137493
*New setting target.max-children-count gives an upper-bound to the number of child objects that will be displayed at each depth-level
This might be a breaking change in some scenarios. To override the new limit you can use the --show-all-children (-A) option
to frame variable or increase the limit in your lldbinit file
*Command "type synthetic" has been split in two:
- "type synthetic" now only handles Python synthetic children providers
- the new command "type filter" handles filters
Because filters and synthetic providers are both ways to replace the children of a ValueObject, only one can be effective at any given time.
llvm-svn: 137416
Access to synthetic children by name:
if your object has a synthetic child named foo you can now type
frame variable object.foo (or ->foo if you have a pointer)
and that will print the value of the synthetic child
(if your object has an actual child named foo, the actual child prevails!)
this behavior should also work in summaries, and you should be able to use
${var.foo} and ${svar.foo} interchangeably
(but using svar.foo will mask an actual child named foo)
llvm-svn: 137314
This is helping us track down some extra references to ModuleSP objects that
are causing things to get kept around for too long.
Added a module pointer accessor to target and change a lot of code to use
it where it would be more efficient.
"taret delete" can now specify "--clean=1" which will cleanup the global module
list for any orphaned module in the shared module cache which can save memory
and also help track down module reference leaks like we have now.
llvm-svn: 137294
- accordingly, the test cases for the synthetic providers for the std:: containers have been edited to use
${svar%#} instead of ${svar.len} to print out the count of elements ; the .len synthetic child has been
removed from the synthetic providers
The synthetic children providers for the std:: containers now return None when asked for children indexes >= num_children()
Basic code to support filter names based on regular expressions (WIP)
llvm-svn: 136862
Fixed a bug where Objective-C variables coming out of the expression parser could crash the Python synthetic providers:
- expression parser output has a "frozen data" component, which is a byte-exact copy of the value (in host memory),
if trying to read into memory based on the host address, LLDB would crash. we are now passing the correct (target)
pointer to the Python code
Objective-C "id" variables are now formatted according to their dynamic type, if the -d option to frame variable is used:
- Code based on the Objective-C 2.0 runtime is used to obtain this information without running code on the target
llvm-svn: 136695
- Completely new implementation of SBType
- Various enhancements in several other classes
Python synthetic children providers for std::vector<T>, std::list<T> and std::map<K,V>:
- these return the actual elements into the container as the children of the container
- basic template name parsing that works (hopefully) on both Clang and GCC
- find them in examples/synthetic and in the test suite in functionalities/data-formatter/data-formatter-python-synth
New summary string token ${svar :
- the syntax is just the same as in ${var but this new token lets you read the values
coming from the synthetic children provider instead of the actual children
- Python providers above provide a synthetic child len that returns the number of elements
into the container
Full bug fix for the issue in which getting byte size for a non-complete type would crash LLDB
Several other fixes, including:
- inverted the order of arguments in the ClangASTType constructor
- EvaluationPoint now only returns SharedPointer's to Target and Process
- the help text for several type subcommands now correctly indicates argument-less options as such
llvm-svn: 136504
- you can now define a Python class as a synthetic children producer for a type
the class must adhere to this "interface":
def __init__(self, valobj, dict):
def get_child_at_index(self, index):
def get_child_index(self, name):
then using type synth add -l className typeName
(e.g. type synth add -l fooSynthProvider foo)
(This is still WIP with lots to be added)
A small test case is available also as reference
llvm-svn: 135865
(e.g. ${var%S}). this might already be the default if your variable is of an aggregate type
new feature: synthetic filters. you can restrict the number of children for your variables to only a meaningful subset
- the restricted list of children obeys the typical rules (e.g. summaries prevail over children)
- one-line summaries show only the filtered (synthetic) children, if you type an expanded summary string, or you use Python scripts, all the real children are accessible
- to provide a synthetic children list use the "type synth add" command, as in:
type synth add foo_type --child varA --child varB[0] --child varC->packet->flags[1-4]
(you can use ., ->, single-item array operator [N] and bitfield operator [N-M]; array slice access is not supported, giving simplified names to expression paths is not supported)
- a new -S option to frame variable and target variable lets you override synthetic children and instead show real ones
llvm-svn: 135731
Code cleanup:
- The Format Manager implementation is now split between two files: FormatClasses.{h|cpp} where the
actual formatter classes (ValueFormat, SummaryFormat, ...) are implemented and
FormatManager.{h|cpp} where the infrastructure classes (FormatNavigator, FormatManager, ...)
are contained. The wrapper code always remains in Debugger.{h|cpp}
- Several leftover fields, methods and comments from previous design choices have been removed
type category subcommands (enable, disable, delete) now can take a list of category names as input
- for type category enable, saying "enable A B C" is the same as saying
enable C
enable B
enable A
(the ordering is relevant in enabling categories, and it is expected that a user typing
enable A B C wants to look into category A, then into B, then into C and not the other
way round)
- for the other two commands, the order is not really relevant (however, the same inverted ordering
is used for consistency)
llvm-svn: 135494
The "systemwide summaries" feature has been removed and replaced with a more general and
powerful mechanism.
Categories:
- summaries can now be grouped into buckets, called "categories" (it is expected that categories
correspond to libraries and/or runtime environments)
- to add a summary to a category, you can use the -w option to type summary add and give
a category name (e.g. type summary add -f "foo" foo_t -w foo_category)
- categories are by default disabled, which means LLDB will not look into them for summaries,
to enable a category use "type category enable". once a category is enabled, LLDB will
look into that category for summaries. the rules are quite trivial: every enabled category
is searched for an exact match. if an exact match is nowhere to be found, any match is
searched for in every enabled category (whether it involves cascading, going to base classes,
...). categories are searched into the order in which they were enabled (the most recently
enabled category first, then the second most and so on..)
- by default, most commands that deal with summaries, use a category named "default" if no
explicit -w parameter is given (the observable behavior of LLDB should not change when
categories are not explicitly used)
- the systemwide summaries are now part of a "system" category
llvm-svn: 135463
- Summaries for char*, const char* and char[] are loaded at startup as
system-wide summaries. This means you cannot delete them unless you use
the -a option to type summary delete/clear
- You can add your own system-wide summaries by using the -w option to type
summary add
Several code improvements for the Python summaries feature
llvm-svn: 135326
- you can use a Python script to write a summary string for data-types, in one of
three ways:
-P option and typing the script a line at a time
-s option and passing a one-line Python script
-F option and passing the name of a Python function
these options all work for the "type summary add" command
your Python code (if provided through -P or -s) is wrapped in a function
that accepts two parameters: valobj (a ValueObject) and dict (an LLDB
internal dictionary object). if you use -F and give a function name,
you're expected to define the function on your own and with the right
prototype. your function, however defined, must return a Python string
- test case for the Python summary feature
- a few quirks:
Python summaries cannot have names, and cannot use regex as type names
both issues will be fixed ASAP
major redesign of type summary code:
- type summary working with strings and type summary working with Python code
are two classes, with a common base class SummaryFormat
- SummaryFormat classes now are able to actively format objects rather than
just aggregating data
- cleaner code to print descriptions for summaries
the public API now exports a method to easily navigate a ValueObject hierarchy
New InputReaderEZ and PriorityPointerPair classes
Several minor fixes and improvements
llvm-svn: 135238
- formats %s %char[] %c and %a now work to print 0-terminated c-strings if they are applied to a char* or char[] even without the [] operator (e.g. ${var%s})
- array formats (char[], intN[], ..) now work when applied to an array of a scalar type even without the [] operator (e.g. ${var%int32_t[]})
LLDB will not crash because of endless loop when trying to obtain a summary for an object that has no value and references itself in its summary string
In many cases, a wrong summary string will now display an "<error>" message instead of giving out an empty string
llvm-svn: 135007
- a new --name option for "type summary add" lets you give a name to a summary
- a new --summary option for "frame variable" lets you bind a named summary to one or more variables
${var%s} now works for printing the value of 0-terminated CStrings
type format test case now tests for cascading
- this is disabled on GCC because GCC may end up stripping typedef chains, basically breaking cascading
new design for the FormatNavigator class
new template class CleanUp2 meant to support cleanup routines with 1 additional parameter beyond resource handle
llvm-svn: 134943
new GetValueForExpressionPath() method in ValueObject to navigate expression paths in a more bitfield vs slices aware way
changes to the varformats.html document (WIP)
llvm-svn: 134679
instructions if they are conditional. Also fixed issues where the PC wasn't
getting bit zero stripped for ARM targets when a stack frame was thumb. We
now properly call through the GetOpcodeLoadAddress() functions to make sure
the addresses are properly stripped for any targets that may decorate up
their addresses.
We now don't pass the SIGSTOP signals along. We can revisit this soon, but
currently this was interfering with debugging some older ARM targets that
don't have vCont support in the GDB server.
llvm-svn: 134461
- ${*expr} now simply means to dereference expr before actually using it
- bitfields, array ranges and pointer ranges now work in a (hopefully) more natural and language-compliant way
a new class TypeHierarchyNavigator replicates the behavior of the FormatManager in going through type hierarchies
when one-lining summary strings, children's summaries can be used as well as values
llvm-svn: 134458
- type names can now be regular expressions (exact matching is done first, and is faster)
- integral (and floating) types can be printed as bitfields, i.e. ${var[low-high]} will extract bits low thru high of the value and print them
- array subscripts are supported, both for arrays and for pointers. the syntax is ${*var[low-high]}, or ${*var[]} to print the whole array (the latter only works for statically sized arrays)
- summary is now printed by default when a summary string references a variable. if that variable's type has no summary, value is printed instead. to force value, you can use %V as a format specifier
- basic support for ObjectiveC:
- ObjectiveC inheritance chains are now walked through
- %@ can be specified as a summary format, to print the ObjectiveC runtime description for an object
- some bug fixes
llvm-svn: 134293
implements three commands:
type summary add <format> <typename1> [<typename2> ...]
type summary delete <typename1> [<typename2> ...]
type summary list [<typename1> [<typename2>] ...]
type summary clear
This allows you to specify the default format that will be used to display
summaries for variables, shown when you use "frame variable" or "expression", or the SBValue classes.
Examples:
type summary add "x = ${var.x}" Point
type summary list
type summary add --one-liner SimpleType
llvm-svn: 134108
the FormatManager class. Modified the format arguments in any commands to be
able to use a single character format, or a full format name, or a partial
format name if no full format names match.
Modified any code that was displaying formats to use the new FormatManager
calls so that our help text and errors never get out of date.
Modified the display of the "type format list" command to be a bit more
human readable by showing the format as a format string rather than the single
character format char.
llvm-svn: 133765
This commit adds a new top level command named "type". Currently this command
implements three commands:
type format add <format> <typename1> [<typename2> ...]
type format delete <typename1> [<typename2> ...]
type format list [<typename1> [<typename2>] ...]
This allows you to specify the default format that will be used to display
types when you use "frame variable" or "expression", or the SBValue classes.
Examples:
// Format uint*_t as hex
type format add x uint16_t uint32_t uint64_t
// Format intptr_t as a pointer
type format add p intptr_t
The format characters are the same as "printf" for the most part with many
additions. These format character specifiers are also used in many other
commands ("frame variable" for one). The current list of format characters
include:
a - char buffer
b - binary
B - boolean
c - char
C - printable char
d - signed decimal
e - float
f - float
g - float
i - signed decimal
I - complex integer
o - octal
O - OSType
p - pointer
s - c-string
u - unsigned decimal
x - hex
X - complex float
y - bytes
Y - bytes with ASCII
llvm-svn: 133728
not write output (prompts, instructions,etc.) if the CommandInterpreter
is in batch_mode.
Also, finish updating InputReaders to write to the asynchronous stream,
rather than using the Debugger's output file directly.
llvm-svn: 133162
(or anything running in a terminal) wants. Not what a UI (Xcode) would want
where it creates a debugger per debug window. The current code had an infinite
loop after a debug session ended.
llvm-svn: 132280
the appropriate registers for arm and x86_64. The register names for the
arguments that are the size of a pointer or less are all named "arg1", "arg2",
etc. This allows you to read these registers by name:
(lldb) register read arg1 arg2 arg3
...
You can also now specify you want to see alternate register names when executing
the read register command:
(lldb) register read --alternate
(lldb) register read -A
llvm-svn: 131376
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 ®_value) = 0;
virtual bool
WriteRegister (const RegisterInfo *reg_info, const RegisterValue ®_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
command line driver, including the lldb prompt being output by
editline, the asynchronous process output & error messages, and
asynchronous messages written by target stop-hooks.
As part of this it introduces a new Stream class,
StreamAsynchronousIO. A StreamAsynchronousIO object is created with a
broadcaster, who will eventually broadcast the stream's data for a
listener to handle, and an event type indicating what type of event
the broadcaster will broadcast. When the Write method is called on a
StreamAsynchronousIO object, the data is appended to an internal
string. When the Flush method is called on a StreamAsynchronousIO
object, it broadcasts it's data string and clears the string.
Anything in lldb-core that needs to generate asynchronous output for
the end-user should use the StreamAsynchronousIO objects.
I have also added a new notification type for InputReaders, to let
them know that a asynchronous output has been written. This is to
allow the input readers to, for example, refresh their prompts and
lines, if desired. I added the case statements to all the input
readers to catch this notification, but I haven't added any code for
handling them yet (except to the IOChannel input reader).
llvm-svn: 130721
set by default when dumping registers. If you want to see all of the register
sets you can use the "--all" option:
(lldb) register read --all
If you want to just see some register sets, you can currently specify them
by index:
(lldb) register read --set 0 --set 2
We need to get shorter register set names soon so we can specify the register
sets by name without having to type too much. I will make this change soon.
You can also have any integer encoded registers resolve the address values
back to any code or data from the object files using the "--lookup" option.
Below is sample output when stopped in the libc function "puts" with some
const strings in registers:
Process 8973 stopped
* thread #1: tid = 0x2c03, 0x00007fff828fa30f libSystem.B.dylib`puts + 1, stop reason = instruction step into
frame #0: 0x00007fff828fa30f libSystem.B.dylib`puts + 1
(lldb) register read --lookup
General Purpose Registers:
rax = 0x0000000100000e98 "----------------------------------------------------------------------"
rbx = 0x0000000000000000
rcx = 0x0000000000000001
rdx = 0x0000000000000000
rdi = 0x0000000100000e98 "----------------------------------------------------------------------"
rsi = 0x0000000100800000
rbp = 0x00007fff5fbff710
rsp = 0x00007fff5fbff280
r8 = 0x0000000000000040
r9 = 0x0000000000000000
r10 = 0x0000000000000000
r11 = 0x0000000000000246
r12 = 0x0000000000000000
r13 = 0x0000000000000000
r14 = 0x0000000000000000
r15 = 0x0000000000000000
rip = 0x00007fff828fa30f libSystem.B.dylib`puts + 1
rflags = 0x0000000000000246
cs = 0x0000000000000027
fs = 0x0000000000000000
gs = 0x0000000000000000
As we can see, we see two constant strings and the PC (register "rip") is
showing the code it resolves to.
I fixed the register "--format" option to work as expected.
Added a setting to disable skipping the function prologue when setting
breakpoints as a target settings variable:
(lldb) settings set target.skip-prologue false
Updated the user settings controller boolean value handler funciton to be able
to take the default value so it can correctly respond to the eVarSetOperationClear
operation.
Did some usability work on the OptionValue classes.
Fixed the "image lookup" command to correctly respond to the "--verbose"
option and display the detailed symbol context information when looking up
line table entries and functions by name. This previously was only working
for address lookups.
llvm-svn: 129977
the CommandInterpreter where it was always being used.
Make sure that Modules can track their object file offsets correctly to
allow opening of sub object files (like the "__commpage" on darwin).
Modified the Platforms to be able to launch processes. The first part of this
move is the platform soon will become the entity that launches your program
and when it does, it uses a new ProcessLaunchInfo class which encapsulates
all process launching settings. This simplifies the internal APIs needed for
launching. I want to slowly phase out process launching from the process
classes, so for now we can still launch just as we used to, but eventually
the platform is the object that should do the launching.
Modified the Host::LaunchProcess in the MacOSX Host.mm to correctly be able
to launch processes with all of the new eLaunchFlag settings. Modified any
code that was manually launching processes to use the Host::LaunchProcess
functions.
Fixed an issue where lldb_private::Args had implicitly defined copy
constructors that could do the wrong thing. This has now been fixed by adding
an appropriate copy constructor and assignment operator.
Make sure we don't add empty ModuleSP entries to a module list.
Fixed the commpage module creation on MacOSX, but we still need to train
the MacOSX dynamic loader to not get rid of it when it doesn't have an entry
in the all image infos.
Abstracted many more calls from in ProcessGDBRemote down into the
GDBRemoteCommunicationClient subclass to make the classes cleaner and more
efficient.
Fixed the default iOS ARM register context to be correct and also added support
for targets that don't support the qThreadStopInfo packet by selecting the
current thread (only if needed) and then sending a stop reply packet.
Debugserver can now start up with a --unix-socket (-u for short) and can
then bind to port zero and send the port it bound to to a listening process
on the other end. This allows the GDB remote platform to spawn new GDB server
instances (debugserver) to allow platform debugging.
llvm-svn: 129351
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
platform status -- gets status information for the selected platform
platform create <platform-name> -- creates a new instance of a remote platform
platform list -- list all available platforms
platform select -- select a platform instance as the current platform (not working yet)
When using "platform create" it will create a remote platform and make it the
selected platform. For instances for iPhone OS debugging on Mac OS X one can
do:
(lldb) platform create remote-ios --sdk-version=4.0
Remote platform: iOS platform
SDK version: 4.0
SDK path: "/Developer/Platforms/iPhoneOS.platform/DeviceSupport/4.0"
Not connected to a remote device.
(lldb) file ~/Documents/a.out
Current executable set to '~/Documents/a.out' (armv6).
(lldb) image list
[ 0] /Volumes/work/gclayton/Documents/devb/attach/a.out
[ 1] /Developer/Platforms/iPhoneOS.platform/DeviceSupport/4.0/Symbols/usr/lib/dyld
[ 2] /Developer/Platforms/iPhoneOS.platform/DeviceSupport/4.0/Symbols/usr/lib/libSystem.B.dylib
Note that this is all happening prior to running _or_ connecting to a remote
platform. Once connected to a remote platform the OS version might change which
means we will need to update our dependecies. Also once we run, we will need
to match up the actualy binaries with the actualy UUID's to files in the
SDK, or download and cache them locally.
This is just the start of the remote platforms, but this modification is the
first iteration in getting the platforms really doing something.
llvm-svn: 127934
correct order. Previously this was tacitly implemented but not
enforced, so it was possible to accidentally do things in the wrong
order and cause problems. This fixes that problem.
llvm-svn: 127430
of Stephen Wilson's idea (thanks for the input Stephen!). What I ended up
doing was:
- Got rid of ArchSpec::CPU (which was a generic CPU enumeration that mimics
the contents of llvm::Triple::ArchType). We now rely upon the llvm::Triple
to give us the machine type from llvm::Triple::ArchType.
- There is a new ArchSpec::Core definition which further qualifies the CPU
core we are dealing with into a single enumeration. If you need support for
a new Core and want to debug it in LLDB, it must be added to this list. In
the future we can allow for dynamic core registration, but for now it is
hard coded.
- The ArchSpec can now be initialized with a llvm::Triple or with a C string
that represents the triple (it can just be an arch still like "i386").
- The ArchSpec can still initialize itself with a architecture type -- mach-o
with cpu type and subtype, or ELF with e_machine + e_flags -- and this will
then get translated into the internal llvm::Triple::ArchSpec + ArchSpec::Core.
The mach-o cpu type and subtype can be accessed using the getter functions:
uint32_t
ArchSpec::GetMachOCPUType () const;
uint32_t
ArchSpec::GetMachOCPUSubType () const;
But these functions are just converting out internal llvm::Triple::ArchSpec
+ ArchSpec::Core back into mach-o. Same goes for ELF.
All code has been updated to deal with the changes.
This should abstract us until later when the llvm::TargetSpec stuff gets
finalized and we can then adopt it.
llvm-svn: 126278
now, in addition to cpu type/subtype and architecture flavor, contains:
- byte order (big endian, little endian)
- address size in bytes
- llvm::Triple for true target triple support and for more powerful plug-in
selection.
llvm-svn: 125602
Johnny Chen