Commit Graph

1243 Commits

Author SHA1 Message Date
John McCall c146582e60 When parsing an out-of-line member function declaration, we must delay
access-control diagnostics which arise from the portion of the declarator
following the scope specifier, just in case access is granted by
friending the individual method.  This can also happen with in-line
member function declarations of class templates due to templated-scope
friend declarations.

We were really playing fast-and-loose before with this sort of thing,
and it turned out to work because *most* friend functions are in file
scope.  Making us delay regardless of context exposed several bugs with
how we were manipulating delay.  I ended up needing a concept of a
context that's independent of the declarations in which it appears,
and then I actually had to make some things save contexts correctly,
but delay should be much cleaner now.

I also encapsulated all the delayed-diagnostics machinery in a single
subobject of Sema;  this is a pattern we might want to consider rolling
out to other components of Sema.

llvm-svn: 125485
2011-02-14 07:13:47 +00:00
John McCall 8322c3a197 Give some convenient idiomatic accessors to Stmt::child_range and
Stmt::const_child_range, then make a bunch of places use them instead
of the individual iterator accessors.

llvm-svn: 125450
2011-02-13 04:07:26 +00:00
Argyrios Kyrtzidis 7dd856a76b For -Woverloaded-virtual take into account canonical methods. Fixes rdar://8979966 & http://llvm.org/PR9182.
llvm-svn: 125296
2011-02-10 18:13:41 +00:00
Sebastian Redl 0890502f44 Basic implementation of inherited constructors. Only generates declarations, and probably only works for very basic use cases.
llvm-svn: 124970
2011-02-05 19:23:19 +00:00
Anders Carlsson 3c23948481 Re-land r124768, with a fix for PR9130.
We now emit everything except unused implicit virtual member functions when building the vtable.

llvm-svn: 124935
2011-02-05 04:35:53 +00:00
Argyrios Kyrtzidis 7272d9cf36 Implement -Woverloaded-virtual.
The difference with gcc is that it warns if you overload virtual methods only if
the method doesn't also override any method. This is to cut down on the number of warnings
and make it more useful like reported here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20423.
If we want to warn that not all overloads are overriden we can have an additional
warning like -Wpartial-override.

-Woverloaded-virtual, unlike gcc, is added to -Wmost. Addresses rdar://8757630.

llvm-svn: 124805
2011-02-03 18:01:15 +00:00
Rafael Espindola af5b92e8f7 Revert 124768.
This reopens PR99114, but that one at least can be avoided with an #include.
PR9130 cannot.

llvm-svn: 124780
2011-02-03 06:30:58 +00:00
Anders Carlsson bac84bef65 Don't try to mark virtual members referenced for classes where the key function
is not defined in the current translation unit. Doing so lead to compile errors
such as PR9114.

Instead, when CodeGen is building the vtable, don't try to emit a definition
for functions that aren't marked used in the current translation unit.
Fixes PR9114.

llvm-svn: 124768
2011-02-03 02:08:44 +00:00
Argyrios Kyrtzidis 83b797f490 Don't warn for -Wnon-virtual-dtor for dependent classes.
llvm-svn: 124735
2011-02-02 18:47:41 +00:00
Argyrios Kyrtzidis 33799caf6d Only warn for -Wnon-virtual-dtor for public destructors. Thanks to Benjamin Kramer for the hint!
llvm-svn: 124585
2011-01-31 17:10:25 +00:00
Argyrios Kyrtzidis 7f3986dc64 Warn if the class has virtual methods but non-virtual destructor. Addresses rdar://8756445.
llvm-svn: 124582
2011-01-31 07:05:00 +00:00
Anders Carlsson a03f3a85cb When building with optimizations, emit vtables where the key is not in the
current translation unit as available_externally. 

This helps devirtualize the second example in PR3100, comment 18:

struct S { S() {}; virtual void xyzzy(); };
inline void foo(S *s) { s->xyzzy(); }
void bar() { S s; foo(&s); }

This involved four major changes:

1. In DefineUsedVTables, always mark virtual member functions as referenced for
   non-template classes and class template specializations.
2. In CodeGenVTables::ShouldEmitVTableInThisTU return true if optimizations are
   enabled, even if the key function is not implemented in this translation 
   unit. We don't ever do this for code compiled with -fapple-kext, because we
   don't ever want to devirtualize virtual member function calls in that case.
3. Give the correct linkage for vtables where the key function is not defined.
4. Update the linkage for RTTI structures when necessary.

llvm-svn: 124565
2011-01-30 20:45:54 +00:00
Douglas Gregor 3fb22baddb Allow elision of invocations of move constructors from temporary objects.
llvm-svn: 124455
2011-01-27 23:24:55 +00:00
Douglas Gregor db9d664244 Rvalue references for *this:
- Add ref-qualifiers to the type system; they are part of the
  canonical type. Print & profile ref-qualifiers
  - Translate the ref-qualifier from the Declarator chunk for
    functions to the function type. 
  - Diagnose mis-uses of ref-qualifiers w.r.t. static member
  functions, free functions, constructors, destructors, etc.
  - Add serialization and deserialization of ref-qualifiers.

llvm-svn: 124281
2011-01-26 05:01:58 +00:00
Anders Carlsson e771e76326 Don't insert class templates into the DynamicClasses vector.
llvm-svn: 124201
2011-01-25 18:08:22 +00:00
Anders Carlsson 1eb95961d3 Use attributes for all the override control specifiers.
llvm-svn: 124122
2011-01-24 16:26:15 +00:00
Anders Carlsson 19588aa40b Get rid of the [[final]] C++0x attribute.
llvm-svn: 124083
2011-01-23 21:07:30 +00:00
Anders Carlsson 7d59a68330 Implement [class.derived]p8.
llvm-svn: 124047
2011-01-22 22:23:37 +00:00
Anders Carlsson fc1eef4898 Mark classes as final or explicit. Diagnose when a class marked 'final' is used as a base.
llvm-svn: 124039
2011-01-22 17:51:53 +00:00
Anders Carlsson 3f610c75de Diagnose when a virtual member function marked final is overridden.
llvm-svn: 123916
2011-01-20 16:25:36 +00:00
Anders Carlsson fa8e5d3d67 When checking for functions marked override, ignore dependent contexts.
llvm-svn: 123894
2011-01-20 06:33:26 +00:00
Anders Carlsson c87f861d0a Make CheckOverrideControl a member of Sema.
llvm-svn: 123893
2011-01-20 06:29:02 +00:00
Anders Carlsson fd83553733 Diagnose virtual member functions marked override but not overriding any virtual member functions.
llvm-svn: 123888
2011-01-20 05:57:14 +00:00
Anders Carlsson 13a69102d7 Only allow virtual member functions to be marked 'override' and 'final'.
llvm-svn: 123882
2011-01-20 04:34:22 +00:00
Anders Carlsson db36b8055e Pass the VirtSpecifiers along to Sema::ActOnCXXMemberDeclarator.
llvm-svn: 123878
2011-01-20 03:57:25 +00:00
Douglas Gregor 0231d8dac7 Implement support for non-type template parameter packs whose type is
a pack expansion, e.g., the parameter pack Values in:

  template<typename ...Types>
  struct Outer {
    template<Types ...Values>
    struct Inner;
  };

This new implementation approach introduces the notion of an
"expanded" non-type template parameter pack, for which we have already
expanded the types of the parameter pack (to, say, "int*, float*",
for Outer<int*, float*>) but have not yet expanded the values. Aside
from creating these expanded non-type template parameter packs, this
patch updates template argument checking and non-type template
parameter pack instantiation to make use of the appropriate types in
the parameter pack.

llvm-svn: 123845
2011-01-19 20:10:05 +00:00
John McCall 8b7fd8f156 When building the copy expression for a __block variable, make sure
there's a respectable point of instantiation.  Also, make sure we do
this operation even when instantiating a dependently-typed variable.

llvm-svn: 123818
2011-01-19 11:48:09 +00:00
John McCall 424cec97bd Change QualType::getTypePtr() to return a const pointer, then change a
thousand other things which were (generally inadvertantly) relying on that.

llvm-svn: 123814
2011-01-19 06:33:43 +00:00
Jeffrey Yasskin 8dfa5f1776 Fix warnings found by gcc-4.6, from -Wunused-but-set-variable and
-Wint-to-pointer-cast.

llvm-svn: 123719
2011-01-18 02:00:16 +00:00
Alexis Hunt 1d7926502f Renamed CXXBaseOrMemberInitializer to CXXCtorInitializer. This is both shorter,
more accurate, and makes it make sense for it to hold a delegating constructor
call.

llvm-svn: 123084
2011-01-08 20:30:50 +00:00
Alexis Hunt 4049b8d4cd Check for delegating constructors and (currently) return an error about them.
llvm-svn: 123076
2011-01-08 19:20:43 +00:00
Douglas Gregor 44e7df67d9 Implement pack expansion of base initializers, so that we can
initialize those lovely mixins that come from pack expansions of base
specifiers.

llvm-svn: 122793
2011-01-04 00:32:56 +00:00
Douglas Gregor 752a595655 Implement pack expansions whose pattern is a base-specifier.
llvm-svn: 122782
2011-01-03 22:36:02 +00:00
Douglas Gregor 7c84229db6 When checking a using declaration, make sure that the context we're
looking in is complete. Fixes PR8756.

llvm-svn: 122323
2010-12-21 07:41:49 +00:00
Douglas Gregor 72772f65ec Add tests checking for unexpanded parameter packs in declarations that
occur within statements. Teach Sema::ActOnExceptionDeclarator() to
check for unexpanded parameter packs in the exception type.

llvm-svn: 121984
2010-12-16 17:48:04 +00:00
Douglas Gregor 6ff1fbf65f Check for unexpanded parameter packs in default arguments.
llvm-svn: 121962
2010-12-16 08:48:57 +00:00
Douglas Gregor 71f39c96fb Check for unexpanded parameter packs within variable initializers.
llvm-svn: 121938
2010-12-16 01:31:22 +00:00
Douglas Gregor 6c110f385b Check for unexpanded parameter packs in friend declarations.
llvm-svn: 121934
2010-12-16 01:14:37 +00:00
Douglas Gregor c435653998 Check for unexpanded parameter packs in using declarations. As a
drive-by, make sure to check for unexpanded parameter packs within the
name of a declaration.

llvm-svn: 121930
2010-12-16 00:46:58 +00:00
Douglas Gregor ef68feee1d Check for unexpanded parameter packs in static assertion expressions.
llvm-svn: 121922
2010-12-15 23:55:21 +00:00
Argyrios Kyrtzidis 1cb0de1d4c Fix diagnostic pragmas.
Diagnostic pragmas are broken because we don't keep track of the diagnostic state changes and we only check the current/latest state.
Problems manifest if a diagnostic is emitted for a source line that has different diagnostic state than the current state; this can affect
a lot of places, like C++ inline methods, template instantiations, the lexer, etc.

Fix the issue by having the Diagnostic object keep track of the source location of the pragmas so that it is able to know what is the diagnostic state at any given source location.

Fixes rdar://8365684.

llvm-svn: 121873
2010-12-15 18:44:22 +00:00
John McCall db40c7f573 Restore r121752 without modification.
llvm-svn: 121763
2010-12-14 08:05:40 +00:00
John McCall 5546da68bb Pull out r121752 in case it's causing the selfhost breakage.
llvm-svn: 121759
2010-12-14 07:30:51 +00:00
John McCall aa6d98c160 Factor out most of the extra state in a FunctionProtoType into a separate
class to be passed around.  The line between argument and return types and
everything else is kindof vague, but I think it's justifiable.

llvm-svn: 121752
2010-12-14 06:51:39 +00:00
Douglas Gregor 506bd56484 Variadic templates: extend Type, NestedNameSpecifier, TemplateName,
and TemplateArgument with an operation that determines whether there
are any unexpanded parameter packs within that construct. Use this
information to diagnose the appearance of the names of parameter packs
that have not been expanded (C++ [temp.variadic]p5). Since this
property is checked often (every declaration, ever expression
statement, etc.), we extend Type and Expr with a bit storing the
result of this computation, rather than walking the AST each time to
determine whether any unexpanded parameter packs occur.

This commit is deficient in several ways, which will be remedied with
future commits:
  - Expr has a bit to store the presence of an unexpanded parameter
  pack, but it is never set.
  - The error messages don't point out where the unexpanded parameter
  packs were named in the type/expression, but they should. 
  - We don't check for unexpanded parameter packs in all of the places
  where we should.
  - Testing is sparse, pending the resolution of the above three
  issues.

llvm-svn: 121724
2010-12-13 22:49:22 +00:00
Chandler Carruth 88bfa5ea51 Move the functionality to mark all vtables of key functions as used within
a translation unit to the ActOnEndOfTranslationUnit function instead of doing
it at the start of DefineUsedVTables. The latter is now called *recursively*
during template instantiation, which causes an absolutely insane number of
walks of every record decl in the translation unit.

After this patch, an extremely template instantiation heavy test case's compile
time drops by 10x, and we see between 15% and 20% improvement in average
compile times across a project. This is just recovering a regression, it
doesn't make anything faster than it was several weeks ago.

llvm-svn: 121644
2010-12-12 21:36:11 +00:00
Abramo Bagnara 924a8f3573 Added ParenType type node.
llvm-svn: 121488
2010-12-10 16:29:40 +00:00
John McCall 2faf32c4fd Treat visibility on an enclosing namespace as a non-explicit source of
visibility.  Fixes PR8713.

I've disabled a test which was testing that you can #pragma pop visibility
to get out of a namespace's visibility attribute.  We should probably just
diagnose that as an error unless it's instrumental to someone's system
headers.

llvm-svn: 121459
2010-12-10 02:59:44 +00:00
Jay Foad 6d4db0c885 PR5207: Change APInt methods trunc(), sext(), zext(), sextOrTrunc() and
zextOrTrunc(), and APSInt methods extend(), extOrTrunc() and new method
trunc(), to be const and to return a new value instead of modifying the
object in place.

llvm-svn: 121121
2010-12-07 08:25:34 +00:00
Douglas Gregor a40433a33a Use Sema::MaybeCreateExprWithCleanups() only after we've checked for a NULL/invalid expression
llvm-svn: 121081
2010-12-07 00:41:46 +00:00
Chandler Carruth d44c3105fb Un-templatetize this method. It's definition is out of line in the .cpp file,
so that's not a valid thing to do at all. Instead, switch to a ValueDecl
argument, the template isn't really necessary here.

When handling the types explicitly in the code, it becomes awkward to cerate
the CXXBaseOrMemberInitializer object in so many places. Re-flow the code to
calculate the Init expression first, and then create the initializer. If this
is too gross, we can factor the init expression logic into helper functions,
but it's not past my threshold yet.

llvm-svn: 120997
2010-12-06 09:23:57 +00:00
John McCall 5d41378146 Rename CXXExprWithTemporaries -> ExprWithCleanups; there's no theoretical
reason this is limited to C++, and it's certainly not limited to temporaries.

llvm-svn: 120996
2010-12-06 08:20:24 +00:00
Francois Pichet d583da04d0 More anonymous struct/union redesign. This one deals with anonymous field used in a constructor initializer list:
struct X {
  X() : au_i1(123) {}
  union {
    int au_i1;
    float au_f1;
  };
};

clang will now deal with au_i1 explicitly as an IndirectFieldDecl.

llvm-svn: 120900
2010-12-04 09:14:42 +00:00
John McCall 032092ffc2 Revert r120063, it was wrong.
llvm-svn: 120296
2010-11-29 18:01:58 +00:00
Nico Weber 20c9f1db48 Minor whitespace and comment fixes. No functionality change.
llvm-svn: 120266
2010-11-28 22:53:37 +00:00
John McCall 8a1013f8c3 Redeclarations of using declarations are not okay in function scopes.
Not sure what I was thinking before.

Fixes PR8668.

llvm-svn: 120063
2010-11-23 22:03:51 +00:00
Francois Pichet 783dd6ece4 Major anonymous union/struct redesign.
A new AST node is introduced:
   def IndirectField : DDecl<Value>;
IndirectFields are injected into the anonymous's parent scope and chain back to
the original field. Name lookup for anonymous entities now result in an
IndirectFieldDecl instead of a FieldDecl.
There is no functionality change, the code generated should be the same.

llvm-svn: 119919
2010-11-21 06:08:52 +00:00
Nick Lewycky 9331ed89f8 A bundle of whitespace changes, separated out from the functional changes.
llvm-svn: 119886
2010-11-20 01:29:55 +00:00
Argyrios Kyrtzidis 1865342442 Refactoring.
Move ErrorTrap from clang/Sema to clang/Basic as DiagnosticErrorTrap and use it in Scope.

llvm-svn: 119763
2010-11-19 00:19:12 +00:00
John McCall 4bc41aee2b Add an assertion, fix a whole bunch of bugs, comment the assertion
out because there are still bugs left.

llvm-svn: 119722
2010-11-18 19:01:18 +00:00
John McCall 7decc9e4ea Calculate the value kind of an expression when it's created and
store it on the expression node.  Also store an "object kind",
which distinguishes ordinary "addressed" l-values (like
variable references and pointer dereferences) and bitfield,
@property, and vector-component l-values.

Currently we're not using these for much, but I aim to switch
pretty much everything calculating l-valueness over to them.
For now they shouldn't necessarily be trusted.

llvm-svn: 119685
2010-11-18 06:31:45 +00:00
Ted Kremenek 5eec2b0bd3 Region-allocate all AttributeList objects from a factory object instead of manually managing them
using new/delete and OwningPtrs.  After memory profiling Clang, I witnessed periodic leaks of these
objects; digging deeper into the code, it was clear that our management of these objects was a mess.  The ownership rules were murky at best, and not always followed.  Worse, there are plenty of error paths where we could screw up.

This patch introduces AttributeList::Factory, which is a factory class that creates AttributeList
objects and then blows them away all at once.  While conceptually simple, most of the changes in
this patch just have to do with migrating over to the new interface.  Most of the changes have resulted in some nice simplifications.

This new strategy currently holds on to all AttributeList objects during the lifetime of the Parser
object.  This is easily tunable.  If we desire to have more bound the lifetime of AttributeList
objects more precisely, we can have the AttributeList::Factory object (in Parser) push/pop its
underlying allocator as we enter/leave key methods in the Parser.  This means that we get
simple memory management while still having the ability to finely control memory use if necessary.

Note that because AttributeList objects are now BumpPtrAllocated, we may reduce malloc() traffic
in many large files with attributes.

This fixes the leak reported in: <rdar://problem/8650003>

llvm-svn: 118675
2010-11-10 05:59:39 +00:00
John McCall 9b72f89f0f Diagnose attempst to template using declarations and using directives.
Recover from the latter and fail early for the former.  Fixes PR8022.

llvm-svn: 118669
2010-11-10 02:40:36 +00:00
Chris Lattner 8ea6442f92 tidy up
llvm-svn: 118626
2010-11-09 20:15:55 +00:00
Douglas Gregor 23ab745e79 Fix source locations in unnamed bitfield diagnostic, from Jakub
Wieczorek! Fixes PR8025.

llvm-svn: 118481
2010-11-09 03:31:16 +00:00
Douglas Gregor 266bb5fd5b Check for an invalid field earlier in a constructor's initialization
of that field. Otherwise, we can end up building and later trying to
instantiate a dependent member initializer that will fail at
instantiation time. 

Unfortunately, I've only managed to trigger this bug with very large
sources, so there's no test case :(

llvm-svn: 118306
2010-11-05 22:21:31 +00:00
John McCall c3007a2145 No really, we don't have a retain/release system for statements/expressions
anymore.

llvm-svn: 117357
2010-10-26 07:05:15 +00:00
Chandler Carruth 0171815ae1 Improve the tracking of source locations for parentheses in constructor calls.
This adds them where missing, and traces them through PCH. We fix at least one
bug in the extents found by the Index library, and make a lot of refactoring
tools which care about the exact formulation of a constructor call easier to
write. Also some minor cleanups to more consistently follow the friend pattern
instead of the setter pattern when rebuilding a serialized AST.

Patch originally by Samuel Benzaquen.

llvm-svn: 117254
2010-10-25 08:47:36 +00:00
Douglas Gregor 412c362d9e When performing name lookup for a namespace definition, only look into
the current context's redeclaration context, ignoring using
directives. Fixes PR8430.

llvm-svn: 117097
2010-10-22 15:24:46 +00:00
Craig Silverstein af8808dd28 Pass TInfo to CXXDestructorDecl::Create(), just like we do for other
function decls.

Reviewed by rjmccall and nlewycky.

llvm-svn: 116979
2010-10-21 00:44:50 +00:00
John McCall ace48cd872 Redirect templated friend class decls to a new Sema callback and
construct an unsupported friend when there's a friend with a templated
scope specifier.  Fixes a consistency crash, rdar://problem/8540527

llvm-svn: 116786
2010-10-19 01:40:49 +00:00
John McCall c9739e30c6 Reformatting.
llvm-svn: 116662
2010-10-16 07:23:36 +00:00
John McCall 2c2eb12d9c White-listing templated-scope friend decls is a good idea, but doing it
by marking the decl invalid isn't.  Make some steps towards supporting these
and then hastily shut them down at the last second by marking them as
unsupported.

llvm-svn: 116661
2010-10-16 06:59:13 +00:00
Douglas Gregor 36c22a2335 Diagnose C++ [class.mem]p13-14, where a class member has the same name
as the class itself. Fixes PR7082.

llvm-svn: 116573
2010-10-15 13:21:21 +00:00
John McCall 1c9c3fd50a Death to blocks, or at least the word "block" in one particular obnoxiously
ambiguous context.

llvm-svn: 116567
2010-10-15 04:57:14 +00:00
John McCall f477659040 template-ids are looked up differently in friend declarations.
llvm-svn: 116529
2010-10-14 22:22:28 +00:00
Douglas Gregor a007d36c1b Generalize the checking for qualification of (non-friend) class
members. Provide a hard error when the qualification doesn't match the
current class type, or a warning + Fix-it if it does match the current
class type. Fixes PR8159.

llvm-svn: 116445
2010-10-13 22:19:53 +00:00
John McCall ccbc032146 Preserve the original scope when picking the right scope for a friend
declaration, because we'll need it later.  Hopefully fixed self-host.

llvm-svn: 116399
2010-10-13 06:22:15 +00:00
John McCall f7cfb2212c Support friend function declarations in local classes correctly.
Fixes a crash and diagnoses the error condition of an unqualified
friend which doesn't resolve to something.  I'm still not certain how
this is useful.

llvm-svn: 116393
2010-10-13 05:45:15 +00:00
John McCall de3fd2264c Handle dependent friends more explicitly and deal with the possibility
of templated-scope friends by marking them invalid and white-listing all
accesses until such time as we implement them.  Fixes a crash, this time
without a broken test case.

llvm-svn: 116364
2010-10-12 23:13:28 +00:00
Douglas Gregor 758cb67fcf If we end up instantiating a function parameter whose default argument
has not yet been parsed, note that the default argument hasn't been
parsed and keep track of all of the instantiations of that function
parameter. When its default argument does get parsed, imbue the
instantiations with that default argument. Fixes PR8245.

llvm-svn: 116324
2010-10-12 18:23:32 +00:00
John McCall acf0ee57b6 Track the location of the context requiring an implicit conversion and use it
to white-list conversions required by system headers.  rdar://problem/8232669

llvm-svn: 116029
2010-10-08 02:01:28 +00:00
Nick Lewycky 3005242401 Silence unused variable warning in Release builds.
llvm-svn: 115806
2010-10-06 18:37:39 +00:00
Anders Carlsson 0f7e94fd7b When checking for uninitialized fields in member initializers, special case static variables and enums. Fixes PR8075.
llvm-svn: 115732
2010-10-06 02:43:25 +00:00
Douglas Gregor 457104e98b Move the management of the set of conversion functions in a C++ class
into CXXRecordDecl. The only part that we do not handle this way are
using declarations, since that would require extra name lookup that we
don't currently want to pay for. This fixes <rdar://problem/8459981>,
so that LLDB can build a CXXRecordDecl and magically get all of the
right bits set.

llvm-svn: 115026
2010-09-29 04:25:11 +00:00
Douglas Gregor 8fb9512966 Move the maintenance of CXXRecordDecl::DefinitionData's Abstract bit
completely into CXXRecordDecl, by adding a new completeDefinition()
function. This required a little reshuffling of the final-overrider
checking code, since the "abstract" calculation in the presence of
abstract base classes needs to occur in
CXXRecordDecl::completeDefinition() but we don't want to compute final
overriders more than one in the common case.

llvm-svn: 115007
2010-09-29 00:15:42 +00:00
Douglas Gregor 7d9120c3ed Teach FunctionDecl::setPure() to (indirectly) mark the Abstract bit in
CXXRecordDecl::DefinitionData, rather than having Sema mark the bit.

llvm-svn: 114993
2010-09-28 21:55:22 +00:00
Douglas Gregor 11c024bbeb Reinstate r114925 and r114929, both steps toward
<rdar://problem/8459981>.

llvm-svn: 114984
2010-09-28 20:50:54 +00:00
Douglas Gregor 9d5938ae8a Centralize the management of CXXRecordDecl::DefinitionData's Empty bit
in CXXRecordDecl itself. Yes, this is also part of <rdar://problem/8459981>.

This reinstates r114924, with one crucial bug fix: we were ignoring
the implicit fields created by anonymous structs/unions when updating
the bits in CXXRecordDecl, which means that a class/struct containing
only an anonymous class/struct would be considered "empty". Hilarity
follows. 

llvm-svn: 114980
2010-09-28 20:38:10 +00:00
Douglas Gregor a832d3e2cc Reinstate r114921, which I've exonerated via a self-host build.
Centralize the management of CXXRecordDecl::DefinitionData's Aggregate
and PlainOldData bits in CXXRecordDecl itself. Another milepost on the
road toward <rdar://problem/8459981>.

llvm-svn: 114977
2010-09-28 19:45:33 +00:00
Bill Wendling 6d8c442e08 Temporarily revert 114929 114925 114924 114921. It looked like they (or at least
one of them) was causing a series of failures:

http://google1.osuosl.org:8011/builders/clang-x86_64-darwin10-selfhost/builds/4518

svn merge -c -114929 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114929 into '.':
U    include/clang/Sema/Sema.h
U    include/clang/AST/DeclCXX.h
U    lib/Sema/SemaDeclCXX.cpp
U    lib/Sema/SemaTemplateInstantiateDecl.cpp
U    lib/Sema/SemaDecl.cpp
U    lib/Sema/SemaTemplateInstantiate.cpp
U    lib/AST/DeclCXX.cpp
svn merge -c -114925 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114925 into '.':
G    include/clang/AST/DeclCXX.h
G    lib/Sema/SemaDeclCXX.cpp
G    lib/AST/DeclCXX.cpp
svn merge -c -114924 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114924 into '.':
G    include/clang/AST/DeclCXX.h
G    lib/Sema/SemaDeclCXX.cpp
G    lib/Sema/SemaDecl.cpp
G    lib/AST/DeclCXX.cpp
U    lib/AST/ASTContext.cpp
svn merge -c -114921 https://llvm.org/svn/llvm-project/cfe/trunk
--- Reverse-merging r114921 into '.':
G    include/clang/AST/DeclCXX.h
G    lib/Sema/SemaDeclCXX.cpp
G    lib/Sema/SemaDecl.cpp
G    lib/AST/DeclCXX.cpp

llvm-svn: 114933
2010-09-28 01:09:49 +00:00
Douglas Gregor 1f93ffb3e6 Centralize the management of CXXRecordDecl::DefinitionData's
HasTrivialConstructor, HasTrivialCopyConstructor,
HasTrivialCopyAssignment, and HasTrivialDestructor bits in
CXXRecordDecl's methods. This completes all but the Abstract bit and
the set of conversion functions, both of which will require a bit of
extra work. The majority of <rdar://problem/8459981> is now
implemented (but not all of it).

llvm-svn: 114929
2010-09-28 00:00:00 +00:00
Douglas Gregor d22553cf21 Centralize the management of CXXRecordDecl::DefinitionData's
Polymorphic bit in CXXRecordDecl itself. Yes, this is also part of
<rdar://problem/8459981>.

llvm-svn: 114925
2010-09-27 23:39:06 +00:00
Douglas Gregor 01daafc58a Centralize the management of CXXRecordDecl::DefinitionData's Empty bit
in CXXRecordDecl itself. Yes, this is also part of <rdar://problem/8459981>.

llvm-svn: 114924
2010-09-27 23:31:14 +00:00
Douglas Gregor 1be93f5143 Centralize the management of CXXRecordDecl::DefinitionData's Aggregate
and PlainOldData bits in CXXRecordDecl itself. Another milepost on the
road toward <rdar://problem/8459981>.

llvm-svn: 114921
2010-09-27 23:16:44 +00:00
Douglas Gregor 8f9ebe54b3 Centralize the handling of CXXRecordDecl::DefinitionData's
DeclaredDestructor and UserDeclaredDestructor bits in CXXRecordDecl
itself. Another step on the road to <rdar://problem/8459981>.

llvm-svn: 114918
2010-09-27 22:48:58 +00:00
Douglas Gregor ec3bec0c7a Kill FunctionDecl's IsCopyAssignment bit; it duplicated what could
already be determined by isCopyAssignmentOperator(), and was set too
late in the process for all clients to see the appropriate
value. Cleanup only; no functionality change.

llvm-svn: 114916
2010-09-27 22:37:28 +00:00
Douglas Gregor a1ce1f80cf Centralize the handling of
CXXRecordDecl::DefinitionData::DeclaredCopyAssignment, for
copy-assignment operators. Another step toward <rdar://problem/8459981>.

llvm-svn: 114899
2010-09-27 22:06:20 +00:00
Douglas Gregor d30e79f839 Clean up the handling of the DeclaredDefaultConstructor and
DeclaredCopyConstructor bits in CXXRecordDecl's DefinitionData
structure. Rather than having Sema call addedConstructor or set the
bits directly at semi-random places, move all of the logic for
managing these bits into CXXRecordDecl itself and tie the
addedConstructor call into DeclContext::addDecl().

This makes it easier for AST-building clients to get the right bits
set in DefinitionData, and is one small part of <rdar://problem/8459981>.

llvm-svn: 114889
2010-09-27 21:17:54 +00:00
Fariborz Jahanian 8fb87aec78 Patch implements passing arrays to functions expecting
vla. Implements pr7827.

llvm-svn: 114737
2010-09-24 17:30:16 +00:00
Argyrios Kyrtzidis 03f0e2b5b4 Do not warn with -Wuninitialized when the member is used in a sizeof or address-of expression.
Fixes rdar://8331312.

llvm-svn: 114426
2010-09-21 10:47:20 +00:00
Douglas Gregor 7319327007 Give implicitly-defined default constructors and destructors empty
bodies, from Martin Vejnar!

llvm-svn: 114329
2010-09-20 16:48:21 +00:00
John McCall 386dfc738e static local variables with destructors don't require a global destructor
unless we're on a platform without __cxa_atexit (or use thereof has been
disabled).  This patch actually just disables the check completely for
static locals, but I've filed http://llvm.org/bugs/show_bug.cgi?id=8176 to
track the platform-specific fix.

llvm-svn: 114269
2010-09-18 05:25:11 +00:00
John McCall 7a626f63f7 one piece of code is responsible for the lifetime of every aggregate
slot.  The easiest way to do that was to bundle up the information
we care about for aggregate slots into a new structure which demands
that its creators at least consider the question.

I could probably be convinced that the ObjC 'needs GC' bit should
be rolled into this structure.
Implement generalized copy elision.  The main obstacle here is that
IR-generation must be much more careful about making sure that exactly

llvm-svn: 113962
2010-09-15 10:14:12 +00:00
Sebastian Redl c15c326b51 Remove CXXRecordDecl::getDefaultConstructor(), an inherently unsafe function due to lazy declaration of default constructors. Now that __has_nothrow_constructor doesn't use it anymore, part of PR8101 is fixed.
llvm-svn: 113794
2010-09-13 22:02:47 +00:00
Douglas Gregor e47191cfe1 Fix a typo noted by Gabor
llvm-svn: 113754
2010-09-13 16:44:26 +00:00
Douglas Gregor 6309e3d18e Don't complain about useless user-defined conversion functions when
they were instantiated from a template. In template metaprogramming,
stuff happens. Fixes PR8065.

llvm-svn: 113722
2010-09-12 07:22:28 +00:00
Douglas Gregor 9f0e1aa0f2 Simplify template instantiation for C++ exception declarations,
eliminating an unnecessary use of TemporaryBase in the process.

llvm-svn: 113500
2010-09-09 17:09:21 +00:00
Douglas Gregor ce5aa33385 Eliminate the comma locations from all of the Sema routines that deal
with comma-separated lists. We never actually used the comma
locations, nor did we store them in the AST, but we did manage to
waste time during template instantiation to produce fake locations.

llvm-svn: 113495
2010-09-09 16:33:13 +00:00
Sebastian Redl 02f1eebdc0 Don't give 'global constructor' warnings for function statics, even if they have a direct initializer. Fixes PR8095.
llvm-svn: 113344
2010-09-08 04:46:19 +00:00
John McCall 96326e447c Devirtualize Sema, kill off DeleteExpr and DeleteStmt, and reformat.
llvm-svn: 112945
2010-09-03 02:10:08 +00:00
Sebastian Redl b5c2baa2d7 Enable inline namespaces in the AST.
llvm-svn: 112564
2010-08-31 00:36:36 +00:00
Sebastian Redl 50c682585f Rename DeclContext::getLookupContext to getRedeclContext and change its semantics slightly. No functionality change in the absence of inline namespaces. Also, change a few places where inline namespaces actually make a difference to be prepared for them.
llvm-svn: 112563
2010-08-31 00:36:30 +00:00
Alexis Hunt 3b7918625c Revert my user-defined literal commits - r1124{58,60,67} pending
some issues being sorted out.

llvm-svn: 112493
2010-08-30 17:47:05 +00:00
Alexis Hunt 79eb5469e0 Implement C++0x user-defined string literals.
The extra data stored on user-defined literal Tokens is stored in extra
allocated memory, which is managed by the PreprocessorLexer because there isn't
a better place to put it that makes sure it gets deallocated, but only after
it's used up. My testing has shown no significant slowdown as a result, but
independent testing would be appreciated.

llvm-svn: 112458
2010-08-29 21:26:48 +00:00
Argyrios Kyrtzidis 43b205796f Fix the memory leak of FloatingLiteral/IntegerLiteral.
For large floats/integers, APFloat/APInt will allocate memory from the heap to represent these numbers.
Unfortunately, when we use a BumpPtrAllocator to allocate IntegerLiteral/FloatingLiteral nodes the memory associated with
the APFloat/APInt values will never get freed.
I introduce the class 'APNumericStorage' which uses ASTContext's allocator for memory allocation and is used internally by FloatingLiteral/IntegerLiteral.

Fixes rdar://7637185

llvm-svn: 112361
2010-08-28 09:06:06 +00:00
Sebastian Redl 6766794c0b Parser support for inline namespaces
llvm-svn: 112320
2010-08-27 23:12:46 +00:00
John McCall faf5fb4b78 One who seeks knowledge learns something new every day.
One who seeks the Tao unlearns something new every day.
Less and less remains until you arrive at non-action.
When you arrive at non-action,
nothing will be left undone.

llvm-svn: 112244
2010-08-26 23:41:50 +00:00
John McCall b1be523752 Move things around so that Sema.h no longer depends on even DeclBase.h.
It still depends on Type because DeclarationName.h does.

llvm-svn: 112177
2010-08-26 09:15:37 +00:00
John McCall 8e7d656a4a De-memberify the VarDecl and FunctionDecl StorageClass enums.
This lets us remove Sema.h's dependency on Expr.h and Decl.h.

llvm-svn: 112156
2010-08-26 03:08:43 +00:00
John McCall 8302463dc6 Split out a header to hold APIs meant for the Sema implementation from Sema.h.
Clients of Sema don't need to know (for example) the list of diagnostics we
support.

llvm-svn: 112093
2010-08-25 22:03:47 +00:00
John McCall e302792b61 GCC didn't care for my attempt at API compatibility, so brute-force everything
to the new constants.

llvm-svn: 112047
2010-08-25 11:45:40 +00:00
John McCall 2536c6da0e More incremental progress towards not including Expr.h in Sema.h.
llvm-svn: 112044
2010-08-25 10:28:54 +00:00
John McCall 28a0cf7825 Remove Sema.h's dependency on DeclCXX.h.
llvm-svn: 112032
2010-08-25 07:42:41 +00:00
John McCall a1e130be25 Remove the DenseSet dependency from Sema.h.
llvm-svn: 112030
2010-08-25 07:03:20 +00:00
John McCall cc14d1fd23 More header elimination. The goal of all this is to allow Parser to
#include Sema.h while keeping all the AST declarations opaque.  That may
not be reasonably attainable, though.

llvm-svn: 111907
2010-08-24 08:50:51 +00:00
John McCall bfd822ce64 Remove a header dependency from Sema.h at the cost of some type safety.
If someone wants to fix this some other way....

llvm-svn: 111905
2010-08-24 07:32:53 +00:00
John McCall dadc575b1e OwningExprResult -> ExprResult. This patch brought to you by
M-x query-replace-regexp
\(Sema::\|Action::\|Parser::\|\)Owning\(Expr\|Stmt\)Result -> \2Result

llvm-svn: 111903
2010-08-24 06:29:42 +00:00
John McCall ba7bf59503 Abstract out passing around types and kill off ActionBase.
llvm-svn: 111901
2010-08-24 05:47:05 +00:00
Douglas Gregor f0f8369912 Diagnose the presence of multiple initializations of static data
members, from Faisal Vali! Fixes PR6904.

llvm-svn: 111900
2010-08-24 05:27:49 +00:00
John McCall b268a282a4 Kill off ExprArg (now just Expr*) and StmtArg (now just Stmt*).
llvm-svn: 111863
2010-08-23 23:25:46 +00:00
John McCall 37ad551a25 Sundry incremental steps towards killing off Action.
llvm-svn: 111795
2010-08-23 06:44:23 +00:00
Douglas Gregor ec3a3f5257 Do not elide copy construction when we're performing base-class initialization
llvm-svn: 111783
2010-08-22 18:27:02 +00:00
John McCall 4887165193 DeclPtrTy -> Decl *
llvm-svn: 111733
2010-08-21 09:40:31 +00:00
Fariborz Jahanian a50b3a20f1 objective-c ivar refactoring patch. Iterations
over ivars for a varienty of puposes is now
consolidated into two small routines; DeepCollectObjCIvars
and ShallowCollectObjCIvars. 

llvm-svn: 111679
2010-08-20 21:21:08 +00:00
John McCall 8b0666cf79 Another step in the process of making the parser depend on Sema:
- move DeclSpec &c into the Sema library
  - move ParseAST into the Parse library
Reflect this change in a thousand different includes.
Reflect this change in the link orders.

llvm-svn: 111667
2010-08-20 18:27:03 +00:00
Douglas Gregor 5be1eb8daf Revert r111609, which is failing its new test.
llvm-svn: 111611
2010-08-20 03:26:10 +00:00
John McCall a189f2eb0a Detect efforts to declare a template member friend and explicitly ignore them.
Avoids a crash.

llvm-svn: 111609
2010-08-20 01:40:01 +00:00
Douglas Gregor 086cae6c1f Fix the source range of an anonymous namespace, from Jan Bierbaum
llvm-svn: 111561
2010-08-19 20:55:47 +00:00
Alexis Hunt dcfba7b35b Generate Attr subclasses with TableGen.
Now all classes derived from Attr are generated from TableGen.
Additionally, Attr* is no longer its own linked list; SmallVectors or
Attr* are used. The accompanying LLVM commit contains the updates to
TableGen necessary for this.

Some other notes about newly-generated attribute classes:

 - The constructor arguments are a SourceLocation and a Context&,
   followed by the attributes arguments in the order that they were
   defined in Attr.td

 - Every argument in Attr.td has an appropriate accessor named getFoo,
   and there are sometimes a few extra ones (such as to get the length
   of a variadic argument).

Additionally, specific_attr_iterator has been introduced, which will
iterate over an AttrVec, but only over attributes of a certain type. It
can be accessed through either Decl::specific_attr_begin/end or
the global functions of the same name.

llvm-svn: 111455
2010-08-18 23:23:40 +00:00
Chris Lattner 396639db95 zap dead code.
llvm-svn: 111365
2010-08-18 16:09:06 +00:00
John McCall 02db245d5e Rip out the existing retroactive abstract-class usage checker,
which in a fit of zeal wanted to walk the entire translation unit,
and replace it with a new checker that walks the types of declarations
nested within the class.  Also, look into templates when doing this.

llvm-svn: 111357
2010-08-18 09:41:07 +00:00
John McCall 3696dcb171 Don't try to initialize a reference with a constructed temporary if either
of the classes is invalid.  A class is invalid if a base is invalid.

llvm-svn: 111227
2010-08-17 07:23:57 +00:00
Argyrios Kyrtzidis a3ae3eb690 Fix initialization for members of anonymous struct in a union.
llvm-svn: 111159
2010-08-16 17:27:13 +00:00
Douglas Gregor c3a6ade0bb Move Sema's headers into include/clang/Sema, renaming a few along the way.
llvm-svn: 110945
2010-08-12 20:07:10 +00:00
Abramo Bagnara 8de74e9c81 Added locations and type source info for DeclarationName inside UsingDecl.
llvm-svn: 110912
2010-08-12 11:46:03 +00:00
Abramo Bagnara d6d2f18905 Added locations and type source info for DeclarationName.
llvm-svn: 110860
2010-08-11 22:01:17 +00:00
Eli Friedman aab1fda91f PR7800: both virtual and non-virtual bases must be marked as used for VTTs.
llvm-svn: 110526
2010-08-07 23:11:44 +00:00
John McCall cf14216509 Store inheritance paths after CastExprs instead of inside them.
This takes some trickery since CastExpr has subclasses (and indeed,
is abstract).

Also, smoosh the CastKind into the bitfield from Expr.

Drops two words of storage from Expr in the common case of expressions
which don't need inheritance paths.  Avoids a separate allocation and
another word of overhead in cases needing inheritance paths.  Also has
the advantage of not leaking memory, since destructors for AST nodes are
never run.

llvm-svn: 110507
2010-08-07 06:22:56 +00:00
Nick Lewycky 60ecc522e0 Remove the DeclaredInCondition bit now that it's no longer used.
llvm-svn: 110432
2010-08-06 05:43:55 +00:00
Eli Friedman 570024a8d9 Implement #pragma GCC visibility.
llvm-svn: 110315
2010-08-05 06:57:20 +00:00
Sebastian Redl e0351b970d Remove a redundant and broken check. Fixes PR7810.
llvm-svn: 110294
2010-08-05 00:45:34 +00:00
John McCall deb646ebb5 Only look up an 'operator delete' on the definition of a destructor, not on
a declaration.

llvm-svn: 110175
2010-08-04 01:04:25 +00:00
John McCall 8b0f4ff317 Further adjustments to -Wglobal-constructors; works for references and direct
initializations now.

llvm-svn: 110063
2010-08-02 21:13:48 +00:00
Argyrios Kyrtzidis 2d68810caf Read/write in PCH Sema's StdNamespace and StdBadAlloc and use a LazyDeclPtr for them that will deserialize them when needed.
llvm-svn: 110031
2010-08-02 07:14:54 +00:00
Argyrios Kyrtzidis 4f8e17379d Rename getStdNamespace -> getOrCreateStdNamespace, to better reflect its functionality.
No functionality change.

llvm-svn: 110030
2010-08-02 07:14:39 +00:00
John McCall 47e40931c9 Make a first pass at implementing -Wglobal-constructors. I'm worried that this
will end up bizarrely mirroring CGExprConstant, but that might be the hazard of
this feature.

llvm-svn: 109984
2010-08-01 20:20:59 +00:00
Abramo Bagnara ed5b6899ab Fixed typedef inside extern "C".
llvm-svn: 109865
2010-07-30 16:47:02 +00:00
Nick Lewycky 19b9f958e1 Add source location information to C++ base specifiers.
llvm-svn: 109396
2010-07-26 16:56:01 +00:00
Douglas Gregor b412e174db Remove the vast majority of the Destroy methods from the AST library,
since we aren't going to be calling them ever.

llvm-svn: 109377
2010-07-25 18:17:45 +00:00
Eli Friedman e4686d7acf Remove dead code.
llvm-svn: 109358
2010-07-24 21:35:09 +00:00
Eli Friedman 9255adfb7a PR7698: Make sure we correctly handle the initialization of an array with
dependent size.

llvm-svn: 109356
2010-07-24 21:19:15 +00:00
John McCall 2ca705eb13 Support catching Objective C pointers in C++ under the non-fragile NeXT runtime.
Diagnose attempts to do this under the GNU or fragile NeXT runtimes.

llvm-svn: 109298
2010-07-24 00:37:23 +00:00
Sebastian Redl c57d34bc35 Update ImplicitCastExpr to be able to represent an XValue.
llvm-svn: 108807
2010-07-20 04:20:21 +00:00
Douglas Gregor 1d9ef840fa A using declaration can redeclare a typedef to the same type. These
typedefs won't have the same canonical declaration (since they are
distinct), so we need to check for this case specifically. Fixes
<rdar://problem/8018262>.

llvm-svn: 107833
2010-07-07 23:08:52 +00:00
Argyrios Kyrtzidis 36ea322579 Introduce Decl::hasBody() and FunctionDecl::hasBody() and use them instead of getBody() when we are just checking the existence of a body, to avoid de-serialization of the body from PCH.
Makes de-serialization of the function body even more "lazier".

llvm-svn: 107768
2010-07-07 11:31:19 +00:00
Douglas Gregor 6191f6c607 Remove my egregious hack that forced deserialization of visible
declarations when implicitly declaring the default constructor, copy
constructor, destructor, and copy-assignment operators of a
class. Argiris fixed the underlying problem in r107596.

llvm-svn: 107681
2010-07-06 18:29:01 +00:00
John McCall 1e5d75d73f Mark the operator delete associated with a virtual destructor as referenced.
llvm-svn: 107573
2010-07-03 18:33:00 +00:00
Douglas Gregor 9672f92f7f Lazily declare default constructors. We now delay the construction of
declarations for implicit default constructors, copy constructors,
copy assignment operators, and destructors. On a "simple" translation
unit that includes a bunch of C++ standard library headers, we
generate relatively few of these implicit declarations now:

  4/159 implicit default constructors created
  18/236 implicit copy constructors created
  70/241 implicit copy assignment operators created
  0/173 implicit destructors created

And, on this translation unit, this optimization doesn't really
provide any benefit. I'll do some more performance measurements soon,
but this completes the implementation work for <rdar://problem/8151045>.

llvm-svn: 107551
2010-07-03 00:47:00 +00:00
Douglas Gregor a6d695057c Lazily declare implicit copy constructors.
llvm-svn: 107543
2010-07-02 23:41:54 +00:00
Douglas Gregor 330b9cff74 Lazily declare copy-assignment operators.
llvm-svn: 107521
2010-07-02 21:50:04 +00:00
Douglas Gregor 7454c563f1 Lazily declare the implicitly-declared destructor in a C++ class.
llvm-svn: 107510
2010-07-02 20:37:36 +00:00
Douglas Gregor 0be31a2eb7 Move the "current scope" state from the Parser into Action. This
allows Sema some limited access to the current scope, which we only
use in one way: when Sema is performing some kind of declaration that
is not directly driven by the parser (e.g., due to template
instantiatio or lazy declaration of a member), we can find the Scope
associated with a DeclContext, if that DeclContext is still in the
process of being parsed. 

Use this to make the implicit declaration of special member functions
in a C++ class more "scope-less", rather than using the NULL Scope hack.

llvm-svn: 107491
2010-07-02 17:43:08 +00:00
Douglas Gregor e71eddac59 Add a new routine Sema::LookupDestructor and make all destructor-lookup calls use that routine
llvm-svn: 107444
2010-07-01 22:47:18 +00:00
Douglas Gregor 6d880b1a83 Provide exception specifications for implicitly-declared default constructors.
llvm-svn: 107437
2010-07-01 22:31:05 +00:00
Douglas Gregor 4e8b5fb1eb Move the implicit declaration of a default constructor into a separate
routine; no functionality change.

llvm-svn: 107434
2010-07-01 22:02:46 +00:00
Douglas Gregor 8453ddb5fe Provide exception specifications for implicitly-declared copy constructors.
llvm-svn: 107429
2010-07-01 20:59:04 +00:00
Douglas Gregor f120304cc4 Move the implicit declaration of a constructor out to a separate
method. No functionality change.

llvm-svn: 107415
2010-07-01 19:09:28 +00:00
Douglas Gregor cfe682274c Teach DeclareImplicitCopyConstructor how to cope with virtual bases
and multi-dimensional array fields. Fixes several bugs found by
inspection.

llvm-svn: 107411
2010-07-01 18:27:03 +00:00
Douglas Gregor 54be33925a Move declaration of a class's implicit copy constructor into a
separate function. Aside from making the loops infinitely faster, no
functionality change.

llvm-svn: 107407
2010-07-01 17:57:27 +00:00
Douglas Gregor 68e1136585 Provide an exception-specification for an implicitly-declared
copy-assignment operator.

llvm-svn: 107406
2010-07-01 17:48:08 +00:00
Douglas Gregor f56ab7b6be Move the implicit declaration of a class's copy-assignment operator
into a separate routine. No functionality change.

llvm-svn: 107402
2010-07-01 16:36:15 +00:00
Douglas Gregor 3311ed4f0e Make loops infinitely faster. No functionality change.
llvm-svn: 107398
2010-07-01 15:29:53 +00:00
Douglas Gregor bac7490ffc Remove unnecessary ASTContext parameter from
CXXRecordDecl::getDestructor(); no functionality change.

llvm-svn: 107394
2010-07-01 14:13:13 +00:00
Douglas Gregor 957551609c Reinstate fix for PR7526, which was failing because, now that we
aren't dropping all exception specifications on destructors, the
exception specifications on implicitly-declared destructors were
detected as being wrong (which they were). 

Introduce logic to provide a proper exception-specification for
implicitly-declared destructors. This also fixes PR6972.

Note that the other implicitly-declared special member functions also
need to get exception-specifications. I'll deal with that in a
subsequent commit.

llvm-svn: 107385
2010-07-01 05:10:53 +00:00
Douglas Gregor 3671ad4571 Revert r107374, which broke bootstrap.
llvm-svn: 107378
2010-07-01 03:28:42 +00:00
Douglas Gregor c6e4c68371 When building the type of a destructor, make sure to keep the
exception specification. Fixes PR7526.

llvm-svn: 107374
2010-07-01 02:33:15 +00:00
Chandler Carruth 139e96216a Reapply r107235, this time with both my typo fixed, and a logical bug fixed.
Previously we relied on the presence of a member which needs no initialization
to prevent us from creating an additional initialization of the outer anonymous
union field. We have already correctly marked that field as initialized by the
member of the union (repeatedly due to the original bug this patch fixes) so we
simply need to bail out.

llvm-svn: 107242
2010-06-30 02:59:29 +00:00
Chandler Carruth 0e783c2e86 Revert r107235, it had a silly typo in it, and fixing the typo breaks something
else. Get the build bots happy while I debug. Very sorry for the delay fixing
this...

llvm-svn: 107239
2010-06-30 02:06:16 +00:00
Chandler Carruth abb04f730e Fix PR7402: We were creating implicit member initializers for every field in an
anonymous union under the presumption that they didn't do anything. While this
is true, our checks for redundant initialization of an anonymous union still
fire when these overlap with explicit user initialization. A cleaner approach
is to avoid initializing multiple members of a union altogether, but this still
is in a rather fuzzy are especially when C++0x allows non-POD types into
unions.

llvm-svn: 107235
2010-06-30 00:54:29 +00:00
Chandler Carruth 9c9286b531 Retain the source location of the constructor when building an implicit member
initialization. I tried several ideas but couldn't come up with a test case for
this that didn't rely on a Clang bug to report a diagnostic after template
instantiation of the constructor due to the implicit initializers. Suggestions
welcome. This fixes the source location aspect of PR7402.

llvm-svn: 107226
2010-06-29 23:50:44 +00:00
Douglas Gregor c048c52734 When typo correction produces a result that is not of the kind we're
looking for, reset the name within the LookupResult structure in
addition to clearing out the results. Fixes PR7508.

llvm-svn: 107197
2010-06-29 19:27:42 +00:00
Douglas Gregor 9629e9ac3e Typo correction for namespace alias definitions
llvm-svn: 107191
2010-06-29 18:55:19 +00:00
Douglas Gregor cdf87024ed Allow a using directive to refer to the implicitly-defined namespace
"std", with a warning, to improve GCC compatibility. Fixes PR7517.

As a drive-by, add typo correction for using directives.

llvm-svn: 107172
2010-06-29 17:53:46 +00:00
Douglas Gregor ebada077d9 Given Decl::isUsed() a flag indicating when to consider the "used"
attribute as part of the calculation. Sema::MarkDeclReferenced(), and
a few other places, want only to consider the "used" bit to determine,
e.g, whether to perform template instantiation. Fixes a linkage issue
with Boost.Serialization.

llvm-svn: 106252
2010-06-17 23:14:26 +00:00
Fariborz Jahanian 4a30307840 Fixed conflict between objc_memmove_collectable builtin
decl. and one ddefined in darwin header file.

llvm-svn: 106107
2010-06-16 16:22:04 +00:00
Douglas Gregor 1c69bf00ae If a non-dependent base class initializer fails to match any direct or
virtual base class, but the class still has dependent base classes,
then don't diagnose the failed match as an error: the right base class
might magically appear. Fixes PR7259.

llvm-svn: 106103
2010-06-16 16:03:14 +00:00
John McCall e9cccd86da Fix a point of semantics with using declaration hiding: method templates
introduced by using decls are hidden even if their template parameter lists
or return types differ from the "overriding" declaration.

Propagate using shadow declarations around more effectively when looking up
template-ids.  Reperform lookup for template-ids in member expressions so that
access control is properly set up.

Fix some number of latent bugs involving template-ids with totally invalid
base types.  You can only actually get these with a scope specifier, since
otherwise the template-id won't parse as a template-id.

Fixes PR7384.

llvm-svn: 106093
2010-06-16 08:42:20 +00:00
Fariborz Jahanian 00bdca561a Workaround a possible VS C++ bug.
llvm-svn: 106070
2010-06-16 00:16:38 +00:00
Fariborz Jahanian 021510e96f Patch adds support for copying of those
objective-c++ class objects which have GC'able objc object
pointers and need to use ObjC's objc_memmove_collectable
API (radar 8070772). 

llvm-svn: 106061
2010-06-15 22:44:06 +00:00
Nick Lewycky a2fb98bfda When analyzing for member self-assignment, don't attempt to dereference null
Stmt* such as those which occur in ?: . Fixes PR7378.

Also, generally whip the code into shape fixing several coding style violations.

llvm-svn: 105992
2010-06-15 07:32:55 +00:00
Abramo Bagnara d73405829b Added AccessSpecDecl node.
llvm-svn: 105525
2010-06-05 05:09:32 +00:00
John McCall 8cb7bdfd33 Alter the interface of GetTypeForDeclarator to return a TypeSourceInfo*.
This is never null, but the associated type might be.

llvm-svn: 105503
2010-06-04 23:28:52 +00:00
John McCall b1cd7dac3d Delay checking for mutable const fields until we're checking the field.
Allows this check to work properly for instantiated fields and removes
an unnecessary GetTypeForDeclarator call.

llvm-svn: 105463
2010-06-04 08:34:12 +00:00
Anders Carlsson a2f74f3c98 Add all final overriders to the map.
llvm-svn: 105374
2010-06-03 01:00:02 +00:00
John McCall 4583186b8b When filtering out previous declarations of friend functions, consider the
lookup context, not the direct semantic context.  Fixes PR7230.

llvm-svn: 104917
2010-05-28 01:41:47 +00:00
Douglas Gregor fd42e9542c Tweak the fix-it for a copy constructor declared to take its argument
by-value. We now only prepend a space if it is needed. 

llvm-svn: 104889
2010-05-27 21:28:21 +00:00
Fariborz Jahanian 1138ba6693 Fixes misc. flexible array bugs in c++ (PR7029).
llvm-svn: 104733
2010-05-26 20:19:07 +00:00
Abramo Bagnara 341d783488 Added source order to CXXBaseOrMemberInitializer.
llvm-svn: 104712
2010-05-26 18:09:23 +00:00
Daniel Dunbar 532177685a IRgen/C++: When mark vtables used, make sure to still append to the VTableUse array if we promote a vtable from being just used to having its definition required. This ensures that we properly inform the consumer about whether the vtable is required or not, previously we could fail to do so when the vtable was in the VTableUses array before the decl which marked it as required.
- I think this can be cleaned up, since this means we may notify the consumer about the vtable twice, but I didn't see an easy fix for this without more substantial refactoring.
 - Doug, please review!

llvm-svn: 104577
2010-05-25 00:33:13 +00:00
Daniel Dunbar 105ce6db87 Simplify.
llvm-svn: 104576
2010-05-25 00:32:58 +00:00
Douglas Gregor 024d80e571 Don't look for a destructor in a dependent type. Fixes PR7198.
llvm-svn: 104445
2010-05-22 17:12:29 +00:00
Douglas Gregor aabdfcb29e Fix a crasher in constructor-initializer reordering warnings (PR7179).
llvm-svn: 104299
2010-05-20 23:49:34 +00:00
Fariborz Jahanian 2792f30f13 Generate objc_memmove_collectable write-barrier for
classes whose base class have GC'able object pointers.

llvm-svn: 104296
2010-05-20 23:34:56 +00:00
John McCall bc83b3f011 Be sure to apply initializers to members of anonymous structs and unions
recursively, e.g. so that members of anonymous unions inside anonymous structs
still get initialized.  Also generate default constructor calls for anonymous
struct members when necessary.

llvm-svn: 104292
2010-05-20 23:23:51 +00:00
Douglas Gregor 3f4f03a235 Add a new failure kind, FK_Incomplete, to InitializationSequence, to
capture failures when we try to initialize an incomplete
type. Previously, we would (ab)use FK_ConversionFailed, then
occasionally dereference a null pointer when trying to diagnose the
failure. Fixes <rdar://problem/7959007>.

llvm-svn: 104286
2010-05-20 22:12:02 +00:00
Abramo Bagnara 1108e7b873 Renamed misleading getSourceRange -> getLocalSourceRange and getFullSourceRange -> getSourceRange for TypeLoc.
llvm-svn: 104220
2010-05-20 10:00:11 +00:00
Douglas Gregor 527786ea3a Various small fixes for construction/destruction of Objective-C++
instance variables:
  - Use isRecordType() rather than isa<RecordType>(), so that we see
  through typedefs in ivar types.
  - Mark the destructor as referenced
  - Perform C++ access control on the destructor

llvm-svn: 104206
2010-05-20 02:24:22 +00:00
Abramo Bagnara d754848f10 Added basic source locations to Elaborated and DependentName types.
llvm-svn: 104169
2010-05-19 21:37:53 +00:00
Fariborz Jahanian 16f94c6e8f Don't attempt to poke into an invalid field's class type
to mark its destructors as referenced which may cause 
a crash. Fixes radar 7896920

llvm-svn: 103953
2010-05-17 18:15:18 +00:00
Douglas Gregor 83de20f5e3 A vtable is used if the key function is defined... even if that key
function's definition is an out-of-class definition marked
"inline". Fixes an assertion in WebKit.

llvm-svn: 103763
2010-05-14 04:08:48 +00:00
Douglas Gregor 88d292ccb8 Rework when and how vtables are emitted, by tracking where vtables are
"used" (e.g., we will refer to the vtable in the generated code) and
when they are defined (i.e., because we've seen the key function
definition). Previously, we were effectively tracking "potential
definitions" rather than uses, so we were a bit too eager about emitting
vtables for classes without key functions. 

The new scheme:
  - For every use of a vtable, Sema calls MarkVTableUsed() to indicate
  the use. For example, this occurs when calling a virtual member
  function of the class, defining a constructor of that class type,
  dynamic_cast'ing from that type to a derived class, casting
  to/through a virtual base class, etc.
  - For every definition of a vtable, Sema calls MarkVTableUsed() to
  indicate the definition. This happens at the end of the translation
  unit for classes whose key function has been defined (so we can
  delay computation of the key function; see PR6564), and will also
  occur with explicit template instantiation definitions.
 - For every vtable defined/used, we mark all of the virtual member
 functions of that vtable as defined/used, unless we know that the key
 function is in another translation unit. This instantiates virtual
 member functions when needed.
  - At the end of the translation unit, Sema tells CodeGen (via the
  ASTConsumer) which vtables must be defined (CodeGen will define
  them) and which may be used (for which CodeGen will define the
  vtables lazily). 

From a language perspective, both the old and the new schemes are
permissible: we're allowed to instantiate virtual member functions
whenever we want per the standard. However, all other C++ compilers
were more lazy than we were, and our eagerness was both a performance
issue (we instantiated too much) and a portability problem (we broke
Boost test cases, which now pass).

Notes:
  (1) There's a ton of churn in the tests, because the order in which
  vtables get emitted to IR has changed. I've tried to isolate some of
  the larger tests from these issues.
  (2) Some diagnostics related to
  implicitly-instantiated/implicitly-defined virtual member functions
  have moved to the point of first use/definition. It's better this
  way.
  (3) I could use a review of the places where we MarkVTableUsed, to
  see if I missed any place where the language effectively requires a
  vtable.

Fixes PR7114 and PR6564.

llvm-svn: 103718
2010-05-13 16:44:06 +00:00
Chris Lattner 5c02c2af3f improve comments.
llvm-svn: 103670
2010-05-12 23:26:21 +00:00
Douglas Gregor 54818f0c37 When we emit an error during the implicit definition of a special
member function (default constructor, copy constructor, copy
assignment operator, destructor), emit a note showing where that
implicit definition was required.

llvm-svn: 103619
2010-05-12 16:39:35 +00:00
Abramo Bagnara 6150c884df Merged Elaborated and QualifiedName types.
llvm-svn: 103517
2010-05-11 21:36:43 +00:00
Daniel Dunbar 0547ad38e3 Speculatively revert r103497, "Do not mark the virtual members of an
implicitly-instantiated class as ...", which seems to have broken bootstrap.

llvm-svn: 103515
2010-05-11 21:32:35 +00:00
Douglas Gregor 0c4aad15c2 Do not mark the virtual members of an implicitly-instantiated class as
referenced unless we see one of them defined (or the key function
defined, if it as one) or if we need the vtable for something. Fixes
PR7114.

llvm-svn: 103497
2010-05-11 20:24:17 +00:00
Douglas Gregor 4b718ee691 It turns out that we should be allowing redeclarations within function
scope. Thanks to Steven Watanabe for correcting me.

llvm-svn: 103210
2010-05-06 23:31:27 +00:00
Douglas Gregor e902956f59 Partial and full specializations of a class template may have a
different tag kind ("struct" vs. "class") than the primary template,
which has an affect on access control.

Should fix the last remaining Boost.Accumulors failure.

llvm-svn: 103144
2010-05-06 00:28:52 +00:00
Douglas Gregor bf1fb44efa When implicit definition of the copy-assignment operator fails,
provide a note that shows where the copy-assignment operator was
needed. We used to have this, but I broke it during refactoring. 

Finishes PR6999.

llvm-svn: 103127
2010-05-05 22:38:15 +00:00
Douglas Gregor 94f9a4820a Reimplement code generation for copying fields in the
implicitly-generated copy constructor. Previously, Sema would perform
some checking and instantiation to determine which copy constructors,
etc., would be called, then CodeGen would attempt to figure out which
copy constructor to call... but would get it wrong, or poke at an
uninstantiated default argument, or fail in other ways.

The new scheme is similar to what we now do for the implicit
copy-assignment operator, where Sema performs all of the semantic
analysis and builds specific ASTs that look similar to the ASTs we'd
get from explicitly writing the copy constructor, so that CodeGen need
only do a direct translation.

However, it's not quite that simple because one cannot explicit write
elementwise copy-construction of an array. So, I've extended
CXXBaseOrMemberInitializer to contain a list of indexing variables
used to copy-construct the elements. For example, if we have:

  struct A { A(const A&); };
  
  struct B {
    A array[2][3];
  };

then we generate an implicit copy assignment operator for B that looks
something like this:

  B::B(const B &other) : array[i0][i1](other.array[i0][i1]) { }

CodeGen will loop over the invented variables i0 and i1 to visit all
elements in the array, so that each element in the destination array
will be copy-constructed from the corresponding element in the source
array. Of course, if we're dealing with arrays of scalars or class
types with trivial copy-assignment operators, we just generate a
memcpy rather than a loop.

Fixes PR6928, PR5989, and PR6887. Boost.Regex now compiles and passes
all of its regression tests.

Conspicuously missing from this patch is handling for the exceptional
case, where we need to destruct those objects that we have
constructed. I'll address that case separately.

llvm-svn: 103079
2010-05-05 05:51:00 +00:00
Douglas Gregor 40c92bbe4f When creating a call to a base subobject's operator= in an
implicitly-defined copy assignment operator, suppress the protected
access check. This eliminates the remaining failure in the
Boost.SmartPtr library (that was a product of the copy-assignment
generation rewrite) and, presumably, the Boost.TR1 library as well.

llvm-svn: 103010
2010-05-04 15:20:55 +00:00
John McCall e61b02bcaf When inheriting a default argument expression, inherit the full expression,
not just the inner expression.  This is important if the expression has any
temporaries.  Fixes PR 7028.

Basically a symptom of really tragic method names.

llvm-svn: 102998
2010-05-04 01:53:42 +00:00
Douglas Gregor 3f324d569b Diagnose unused exception parameters under a different warning group
(-Wunused-exception-parameter) than normal variables, since it's more
common to name and then ignore an exception parameter. This warning is
neither enabled by default nor by -Wall. Fixes <rdar://problem/7931045>.

llvm-svn: 102931
2010-05-03 18:51:14 +00:00
Douglas Gregor ed088b7fe7 When creating the declaration reference for implicit copy-construction
of a base class, give it real source-location information. Fixes PR7017.

llvm-svn: 102916
2010-05-03 15:43:53 +00:00
Douglas Gregor 5cf8d67bc9 When declaring a namespace alias, ignore previous declarations that
aren't in scope. Fixes PR7014.

llvm-svn: 102915
2010-05-03 15:37:31 +00:00
Benjamin Kramer bebee84281 Replace a char*/size pair with stringref.
llvm-svn: 102902
2010-05-03 13:08:54 +00:00
Anders Carlsson bcc066b659 Add an enum to CXXConstructExpr so we can determine if the construction expression constructs a non-virtual or virtual base.
llvm-svn: 102879
2010-05-02 22:54:08 +00:00
Douglas Gregor b139cd5843 Complete reimplementation of the synthesis for implicitly-defined copy
assignment operators. 

Previously, Sema provided type-checking and template instantiation for
copy assignment operators, then CodeGen would synthesize the actual
body of the copy constructor. Unfortunately, the two were not in sync,
and CodeGen might pick a copy-assignment operator that is different
from what Sema chose, leading to strange failures, e.g., link-time
failures when CodeGen called a copy-assignment operator that was not
instantiation, run-time failures when copy-assignment operators were
overloaded for const/non-const references and the wrong one was
picked, and run-time failures when by-value copy-assignment operators
did not have their arguments properly copy-initialized.

This implementation synthesizes the implicitly-defined copy assignment
operator bodies in Sema, so that the resulting ASTs encode exactly
what CodeGen needs to do; there is no longer any special code in
CodeGen to synthesize copy-assignment operators. The synthesis of the
body is relatively simple, and we generate one of three different
kinds of copy statements for each base or member:

  - For a class subobject, call the appropriate copy-assignment
    operator, after overload resolution has determined what that is.
  - For an array of scalar types or an array of class types that have
    trivial copy assignment operators, construct a call to
    __builtin_memcpy.
  - For an array of class types with non-trivial copy assignment
    operators, synthesize a (possibly nested!) for loop whose inner
    statement calls the copy constructor.
  - For a scalar type, use built-in assignment.

This patch fixes at least a few tests cases in Boost.Spirit that were
failing because CodeGen picked the wrong copy-assignment operator
(leading to link-time failures), and I suspect a number of undiagnosed
problems will also go away with this change.

Some of the diagnostics we had previously have gotten worse with this
change, since we're going through generic code for our
type-checking. I will improve this in a subsequent patch.

llvm-svn: 102853
2010-05-01 20:49:11 +00:00
Anders Carlsson 7911150a3d When defining implicit copy constructors, use SetBaseOrMemberInitializers to initialize the bases.
llvm-svn: 102842
2010-05-01 16:39:01 +00:00
Douglas Gregor a57478e8f6 Added an RAII object that helps set up/tear down the Sema context
information required to implicitly define a C++ special member
function. Use it rather than explicitly setting CurContext on entry
and exit, which is fragile. 

Use this RAII object for the implicitly-defined default constructor,
copy constructor, copy assignment operator, and destructor.

llvm-svn: 102840
2010-05-01 15:04:51 +00:00
John McCall 0b66eb38c7 It turns out that basically every caller to RequireCompleteDeclContext
already knows what context it's looking in.  Just pass that context in
instead of (questionably) recalculating it.

llvm-svn: 102818
2010-05-01 00:40:08 +00:00
Fariborz Jahanian c83726e64e More of Sema to implement initialization of
ivar of c++ object types.

llvm-svn: 102500
2010-04-28 16:11:27 +00:00
John McCall 1e67dd6b2f Improve the diagnostic you get when making a qualified member access
with a qualifier referencing a different type.

llvm-svn: 102409
2010-04-27 01:43:38 +00:00
Anders Carlsson 53e1ba948d Revert enough of my patches to fix self-host again :(
llvm-svn: 102289
2010-04-25 00:52:09 +00:00
Anders Carlsson 36db0d9c47 Add an explicit UsuallyTinyPtrVector that takes a single element.
llvm-svn: 102283
2010-04-24 22:54:32 +00:00
Anders Carlsson 90235beb55 DefineImplicitCopyConstructor now uses SetBaseOrMemberInitializers to create implicit base initializers. (Member initializers are still handled by CodeGenFunction::SynthesizeCXXCopyConstructor for now).
llvm-svn: 102279
2010-04-24 22:25:18 +00:00
Anders Carlsson af13c7bea9 When building implicit base initializers, add a derived-to-base cast that points to the exact base specifier.
llvm-svn: 102277
2010-04-24 22:02:54 +00:00
Anders Carlsson a70cff624e Actually produce base paths for CastExprs of kind CK_DerivedToBase.
llvm-svn: 102259
2010-04-24 19:06:50 +00:00
Anders Carlsson 7afe4245e2 Pass the base specifiers through to CheckDerivedToBaseConversion. No functionality change yet.
llvm-svn: 102250
2010-04-24 17:11:09 +00:00
Douglas Gregor bbdf20acd0 Keep track of when DependentNameTypes have no associated keyword
(e.g., no typename, enum, class, etc.), e.g., because the context is
one that is known to refer to a type. Patch from Enea Zaffanella!

llvm-svn: 102243
2010-04-24 15:35:55 +00:00
Anders Carlsson 7a0ffdbe53 Cleanup.
llvm-svn: 102179
2010-04-23 16:24:12 +00:00
Anders Carlsson 423f5d85fd Handle copy initialization in BuildImplicitMemberInitializer. Not used yet.
llvm-svn: 102178
2010-04-23 16:04:08 +00:00
Anders Carlsson 1b00e24d13 Add an ImplicitInitializerKind enum and pass it to BuildImplicitBaseInitializer and BuildImplicitMemberInitializer.
llvm-svn: 102166
2010-04-23 03:10:23 +00:00
Anders Carlsson dca6be04db Fix a think-o that broke self-host.
llvm-svn: 102165
2010-04-23 03:07:47 +00:00
Anders Carlsson 3c1db570b2 Factor code to initialize an implicit member out into a separate function.
llvm-svn: 102162
2010-04-23 02:15:47 +00:00
Anders Carlsson 6bd91c387b Remove calls to isDependentContext, since we handle that case earlier in the code. Make BuildImplicitBaseInitializer return a boolean instead.
llvm-svn: 102159
2010-04-23 02:00:02 +00:00
Anders Carlsson 05bf00938b Re-land the patch that merges two diagnostics into one now that it passes self-host :)
llvm-svn: 102050
2010-04-22 05:40:53 +00:00
Anders Carlsson 3eeb741e6c Remove an unused declaration.
llvm-svn: 102037
2010-04-21 23:55:31 +00:00
Anders Carlsson 43c64af5f0 Keep tack of whether a base in an InitializedEntity is an inherited virtual base or not. Use this in CheckConstructorAccess.
llvm-svn: 102020
2010-04-21 19:52:01 +00:00
Anders Carlsson cedc0a469f Factor some common code out into a separate function.
llvm-svn: 101952
2010-04-20 23:11:20 +00:00
Douglas Gregor c4df407604 Keep track of the actual storage specifier written on a variable or
function declaration, since it may end up being changed (e.g.,
"extern" can become "static" if a prior declaration was static). Patch
by Enea Zaffanella and Paolo Bolzoni.

llvm-svn: 101826
2010-04-19 22:54:31 +00:00
Anders Carlsson 11e5140db9 Vtable -> VTable renames across the board.
llvm-svn: 101666
2010-04-17 20:15:18 +00:00
Douglas Gregor b33eed0ced Collapse the three separate initialization paths in
TryStaticImplicitCast (for references, class types, and everything
else, respectively) into a single invocation of
InitializationSequence.

One of the paths (for class types) was the only client of
Sema::TryInitializationByConstructor, which I have eliminated. This
also simplified the interface for much of the cast-checking logic,
eliminating yet more code.

I've kept the representation of C++ functional casts with <> 1
arguments the same, despite the fact that I hate it. That fix will
come soon. To satisfy my paranoia, I've bootstrapped + tested Clang
with these changes.

llvm-svn: 101549
2010-04-16 22:09:46 +00:00
Benjamin Kramer 659d7fc003 Silence warning.
llvm-svn: 101495
2010-04-16 17:43:15 +00:00
Douglas Gregor b8eaf2944b Audit uses of Sema::LookupSingleName for those lookups that are
intended for redeclarations, fixing those that need it. Fixes PR6831.

This uncovered an issue where the C++ type-specifier-seq parsing logic
would try to perform name lookup on an identifier after it already had
a type-specifier, which could also lead to spurious ambiguity errors
(as in PR6831, but with a different test case).

llvm-svn: 101419
2010-04-15 23:40:53 +00:00
Douglas Gregor b2ccf010fb Feed proper source-location information into Sema::LookupSingleResult,
in case it ends up doing something that might trigger diagnostics
(template instantiation, ambiguity reporting, access
reporting). Noticed while working on PR6831.

llvm-svn: 101412
2010-04-15 22:33:43 +00:00
Douglas Gregor 454a5b65d4 Warn about non-aggregate classes with no user-declared constructors
that have reference or const scalar members, since those members can
never be initializer or modified. Fixes <rdar://problem/7804350>.

llvm-svn: 101316
2010-04-15 00:00:53 +00:00
Douglas Gregor 280e1ee0ae Teach typo correction about various language keywords. We can't
generally recover from typos in keywords (since we would effectively
have to mangle the token stream). However, there are still benefits to
typo-correcting with keywords:
  - We don't make stupid suggestions when the user typed something
  that is similar to a keyword. 
  - We can suggest the keyword in a diagnostic (did you mean
  "static_cast"?), even if we can't recover and therefore don't have
  a fix-it.

llvm-svn: 101274
2010-04-14 20:04:41 +00:00
Douglas Gregor 38ae6ab6bc Refactor and simplify the computation of implicit conversion sequences
for reference binding. The code attempted to handle both the
computation of the ICS and the actual conversion, but the latter is an
anachronism: we now use InitializationSequence for that. 

Sema::CheckReferenceInit is now a static function TryReferenceInit
that's only use within overload resolution, and has been simplified
slightly. It still needs to be updated per C++ [over.ics.ref], by
eliminating more of the lvalue/rvalue checks.

llvm-svn: 101136
2010-04-13 16:31:36 +00:00
Douglas Gregor 538be4622c During referencing binding, only consider conversion functions for
direct reference binding when the source and target types are not
reference-related. Fixes PR6066.

llvm-svn: 101132
2010-04-13 15:07:45 +00:00
John McCall 43314ab3ae Allow classes to befriend implicitly-declared members. Fixes PR6207 for
members of non-templated classes.

llvm-svn: 101122
2010-04-13 07:45:41 +00:00
John McCall 212fa2efb7 Diagnose declarations of conversion functions with declarators other than '()'.
llvm-svn: 101098
2010-04-13 00:04:31 +00:00
Douglas Gregor b93b606759 When creating the implicitly-declared special member functions, be
sure to introduce them into the current Scope (when we have one) in
addition to the DeclContext for the class, so that they can be found
by name lookup for inline members of the class. Fixes PR6570.

llvm-svn: 101047
2010-04-12 17:09:20 +00:00
Anders Carlsson 1c0f8bb4ce Have the CXXBaseOrMemberInitializer keep track of whether an initializer initializes a virtual base or not.
llvm-svn: 101004
2010-04-12 00:51:03 +00:00
John McCall 23eebd9c1e Diagnose more cases of initializing distinct members of an anonymous union
member.  Use a better diagnostic for this case.  Also fix a bug with nested
anonymous structs/unions for -Wreorder;  this last was PR6575.

llvm-svn: 100923
2010-04-10 09:28:51 +00:00
John McCall bb7b658ab5 Diagnose misordered initializers in constructor templates immediately instead of
when they're instantiated.  Merge the note into the -Wreorder warning;  it
doesn't really contribute much, and it was splitting a thought across diagnostics
anyway.  Don't crash in the parser when a constructor's initializers end in a
comma and there's no body;  the recovery here is still terrible, but anything's
better than a crash.

llvm-svn: 100922
2010-04-10 07:37:23 +00:00
John McCall da6841bade Suppress access control diagnostics when looking up a base or member name
fails to find a type.  There are no cases where it's valid for this to produce
an error.

llvm-svn: 100878
2010-04-09 19:01:14 +00:00
Jeffrey Yasskin c76498d409 Make CXXScopeSpec invalid when incomplete, and propagate that into any
Declarator that depends on it.  This fixes several redundant errors and bad
recoveries.

llvm-svn: 100779
2010-04-08 16:38:48 +00:00
John McCall b9c7848ba7 Implement dependent friend function template specializations.
llvm-svn: 100753
2010-04-08 09:05:18 +00:00
Alexis Hunt 7dd2617e9c Implement checking for template literal operator functions. This
code won't actually get used yet because we don't handle non-type
parameter packs, but when we do, this code should jump in and work.

llvm-svn: 100716
2010-04-07 23:11:06 +00:00
Alexis Hunt 079a6f7fe5 Updated comment to reflect changes made in the most recent draft.
llvm-svn: 100707
2010-04-07 22:57:35 +00:00
Douglas Gregor 3b4abb6792 Improve handling of friend types in several ways:
- When instantiating a friend type template, perform semantic
  analysis on the resulting type.
  - Downgrade the errors concerning friend type declarations that do
  not refer to classes to ExtWarns in C++98/03. C++0x allows
  practically any type to be befriended, and ignores the friend
  declaration if the type is not a class.

llvm-svn: 100635
2010-04-07 17:57:12 +00:00
Douglas Gregor afb9bc10fa Split Sema::ActOnFriendTypeDecl into Sema::CheckFriendTypeDecl (for
semantic analysis) and Sema::ActOnFriendTypeDecl (the action
callback). This is a prerequisite for improving template instantiation
of friend type declarations.

llvm-svn: 100633
2010-04-07 16:53:43 +00:00
Douglas Gregor 45cf7e3d2a Rework our handling of copy construction of temporaries, which was a
poor (and wrong) approximation of the actual rules governing when to
build a copy and when it can be elided.

The correct implementation is actually simpler than the
approximation. When we only enumerate constructors as part of
initialization (e.g., for direct initialization or when we're copying
from a class type or one of its derived classes), we don't create a
copy. When we enumerate all conversion functions, we do create a
copy. Before, we created some extra copies and missed some
others. The new test copy-initialization.cpp shows a case where we
missed creating a (required, non-elidable) copy as part of a
user-defined conversion, which resulted in a miscompile. This commit
also fixes PR6757, where the missing copy made us reject well-formed
code in the ternary operator.

This commit also cleans up our handling of copy elision in the case
where we create an extra copy of a temporary object, which became
necessary now that we produce the right copies. The code that seeks to
find the temporary object being copied has moved into
Expr::getTemporaryObject(); it used to have two different
not-quite-the-same implementations, one in Sema and one in CodeGen.

Note that we still do not attempt to perform the named return value
optimization, so we miss copy elisions for return values and throw
expressions.

llvm-svn: 100196
2010-04-02 18:24:57 +00:00
Anders Carlsson db0a965eca If a constructor is a dependent context, just set the base and member initializers as they are written. Fixes a bug where we wouldn't show initialization order warnings when instantiating.
llvm-svn: 100180
2010-04-02 06:26:44 +00:00
Anders Carlsson 7b3f2788a1 Diagnose multiple base and member initializers in class templates.
llvm-svn: 100179
2010-04-02 05:42:15 +00:00
Anders Carlsson 4c8cb01b5a More cleanup.
llvm-svn: 100175
2010-04-02 03:43:34 +00:00
Anders Carlsson 96b8fc664a Constify.
llvm-svn: 100174
2010-04-02 03:38:04 +00:00
Anders Carlsson e857b29e0c Minor cleanup.
llvm-svn: 100173
2010-04-02 03:37:03 +00:00
Douglas Gregor e677dafce1 Change the representation of dependent elaborated-type-specifiers
(such as "class T::foo") from an ElaboratedType of a TypenameType to a
DependentNameType, which more accurately models the underlying
concept.

Improve template instantiation for DependentNameType nodes that
represent nested-name-specifiers, by performing tag name lookup and
checking the resulting tag appropriately. Fixes PR5681.

There is still much testing and cleanup to do in this area.

llvm-svn: 100054
2010-03-31 22:19:08 +00:00
Douglas Gregor a771f46c82 Reinstate my CodeModificationHint -> FixItHint renaming patch, without
the C-only "optimization".

llvm-svn: 100022
2010-03-31 17:46:05 +00:00
Douglas Gregor 30e631862f Revert r100008, which inexplicably breaks the clang-i686-darwin10 builder
llvm-svn: 100018
2010-03-31 17:25:35 +00:00
Douglas Gregor 3baad0d4f7 Rename CodeModificationHint to FixItHint, since we've been using the
term "fix-it" everywhere and even *I* get tired of long names
sometimes. No functionality change.

llvm-svn: 100008
2010-03-31 15:31:50 +00:00
John McCall da4458e98f Regularize support for naming conversion functions in using decls.
llvm-svn: 99979
2010-03-31 01:36:47 +00:00
John McCall 16df1e59f2 Propagate the "found declaration" (i.e. the using declaration instead of
the underlying/instantiated decl) through a lot of API, including "intermediate"
MemberExprs required for (e.g.) template instantiation.  This is necessary
because of the access semantics of member accesses to using declarations:
only the base class *containing the using decl* need be accessible from the
naming class.

This allows us to complete an access-controlled selfhost, if there are no
recent regressions.

llvm-svn: 99936
2010-03-30 21:47:33 +00:00
Rafael Espindola c50c27cca8 the big refactoring bits of PR3782.
This introduces FunctionType::ExtInfo to hold the calling convention and the
noreturn attribute. The next patch will extend it to include the regparm
attribute and fix the bug.

llvm-svn: 99920
2010-03-30 20:24:48 +00:00
Anders Carlsson 83ac312965 Fix a bug where we would incorrectly report an error about initializing two fields in an anonymous struct.
llvm-svn: 99891
2010-03-30 16:19:37 +00:00
Anders Carlsson a942dcd92b Reduce nesting.
llvm-svn: 99889
2010-03-30 15:39:27 +00:00
Douglas Gregor 8933623b91 Optimize PartialDiagnostic's memory-allocation behavior by placing a
cache of PartialDiagnostic::Storage objects into an allocator within
the ASTContext. This eliminates a significant amount of malloc
traffic, for a 10% performance improvement in -fsyntax-only wall-clock
time with 403.gcc's combine.c.

Also, eliminate the RequireNonAbstractType hack I put in earlier,
which was but a symptom of this larger problem.

Fixes <rdar://problem/7806091>.

llvm-svn: 99849
2010-03-29 23:34:08 +00:00
Douglas Gregor 7ec3d11d16 Exit early from the simple form of Sema::RequireNonAbstractType(), for
a 2.47% speedup in 403.gcc.

llvm-svn: 99830
2010-03-29 20:28:52 +00:00
Douglas Gregor c48a10d652 Support __attribute__((packed)) (along with other attributes) at the
end of a struct/class/union in C++, from Justin Bogner!

llvm-svn: 99811
2010-03-29 14:42:08 +00:00
Douglas Gregor 4667effa8b Compare namespaces properly when looking for redeclarations of
namespace aliases. Fixes PR6341.

llvm-svn: 99664
2010-03-26 22:59:39 +00:00
Douglas Gregor c14922f14a When adding initializers to a constructor, be sure that we are looking
through the bases and fields of the definition of the class in which
the constructor is declared, rather than some other declaration of
that class.

llvm-svn: 99661
2010-03-26 22:43:07 +00:00
Rafael Espindola 5b33408481 Mark virtual methods that are used in tables included in VTTs as used.
Fixes PR6706.

llvm-svn: 99582
2010-03-26 00:36:59 +00:00
John McCall 15ad0962dc Preserve type-source information in friend declarations.
llvm-svn: 99525
2010-03-25 18:04:51 +00:00
Douglas Gregor f9f54eab43 Make sure to properly track the anonymous namespace that lives inside
each namespace, even when the outer namespace has multiple
definitions. As part of this, collapsed two pointers worth of storage
(original namespace and inner anonymous namespace) into a single
pointer with a distinguishing bit, since the two are mutually
exclusive, saving a pointer per NamespaceDecl. Fixes PR6620.

llvm-svn: 99368
2010-03-24 00:46:35 +00:00
Douglas Gregor 4165bd6772 Implement computation of the final overriders for each virtual
function within a class hierarchy (C++ [class.virtual]p2).

We use the final-overrider computation to determine when a particular
class is ill-formed because it has multiple final overriders for a
given virtual function (e.g., because two virtual functions override
the same virtual function in the same virtual base class). Fixes
PR5973.

We also use the final-overrider computation to determine which virtual
member functions are pure when determining whether a class is
abstract or diagnosing the improper use of an abstract class. The
prior approach to determining whether there were any pure virtual
functions in a class didn't cope with virtual base class subobjects
properly, and could not easily be fixed to deal with the oddities of
subobject hiding. Fixes PR6631.

llvm-svn: 99351
2010-03-23 23:47:56 +00:00
Rafael Espindola 8d04f0604e A fixed version of r99174 which also includes a test that we emit vtables when
we see an specialization definition ever if we then see a extern template declaration.

llvm-svn: 99226
2010-03-22 23:12:48 +00:00
Rafael Espindola 6f0a14f9ab revert r99174. It caused PR6677. Will try to debug why tonight.
llvm-svn: 99188
2010-03-22 18:10:25 +00:00
Rafael Espindola 4f22b4b1c7 When handling a TSK_ExplicitInstantiationDefinition after a
TSK_ExplicitInstantiationDeclaration make sure we call
MaybeMarkVirtualMembersReferenced with a method attached to the definition.

Remove the hack that forced vtable emition with declarations.

llvm-svn: 99174
2010-03-22 14:43:41 +00:00
John McCall a0296f7987 Remember the "found declaration" for an overload candidate, which is the
entity (if applicable) which was actually looked up.  If a candidate was found
via a using declaration, this is the UsingShadowDecl;  otherwise, if
the candidate is template specialization, this is the template;  otherwise,
this is the function.

The point of this exercise is that "found declarations" are the entities
we do access control for, not their underlying declarations.  Broadly speaking,
this patch fixes access control for using declarations.

There is a *lot* of redundant code calling into the overload-resolution APIs;
we really ought to clean that up.

llvm-svn: 98945
2010-03-19 07:35:19 +00:00
John McCall a630995f50 Perform access control for the implicit calls to base and member destructors
that occur in constructors (on the unwind path).

llvm-svn: 98681
2010-03-16 21:39:52 +00:00
John McCall ab8c273b4f Access control for implicit calls to copy assignment operators and copy
constructors from implicitly-defined members.

llvm-svn: 98614
2010-03-16 06:11:48 +00:00
John McCall f857e0bbe7 Perform access control even for the implicit destructor calls from implicit
destructor definitions.  Remove some code duplication.

llvm-svn: 98611
2010-03-16 05:36:30 +00:00
John McCall 1064d7ef29 Perform access control for the implicit base and member destructor calls
required when emitting a destructor definition.

llvm-svn: 98609
2010-03-16 05:22:47 +00:00
Rafael Espindola 13327bbe55 Fix PR6562. If a type is dependent, we don't know if it will have implicit
destructors.

llvm-svn: 98440
2010-03-13 18:12:56 +00:00
John McCall f3cd665d5a Remember whether a ParmVarDecl was spelled with a default argument or
whether it inherited one from a previous declaration.

Patch by Enea Zaffanella!

llvm-svn: 98362
2010-03-12 18:31:32 +00:00
John McCall 58f10c3380 Maintain type source information for functions through template
instantiation.  Based on a patch by Enea Zaffanella!  I found a way to
reduce some of the redundancy between TreeTransform's "standard"
FunctionProtoType transformation and TemplateInstantiator's override,
and I killed off the old SubstFunctionType by adding type source info
for the last cases where we were creating FunctionDecls without TSI
(at least that get passed through template instantiation).

llvm-svn: 98252
2010-03-11 09:03:00 +00:00
John McCall e78aac41de Create a new InjectedClassNameType to represent bare-word references to the
injected class name of a class template or class template partial specialization.
This is a non-canonical type;  the canonical type is still a template 
specialization type.  This becomes the TypeForDecl of the pattern declaration,
which cleans up some amount of code (and complicates some other parts, but
whatever).

Fixes PR6326 and probably a few others, primarily by re-establishing a few
invariants about TypeLoc sizes.     

llvm-svn: 98134
2010-03-10 03:28:59 +00:00
Rafael Espindola e7113ca907 Delay codegen of vtables when handling implicit instantiations.
This fixes PR6474.

llvm-svn: 98123
2010-03-10 02:19:29 +00:00
Douglas Gregor 104ee00181 Downgrade errors when trying to catch a pointer or reference to
incomplete type to warnings; GCC (and EDG in GCC compatibility mode)
permit such handles. Fixes PR6527.

(For real this time)

llvm-svn: 97927
2010-03-08 01:47:36 +00:00
Douglas Gregor 281c486e1b Robustify callers that rebuild typename type nodes again NULL return
types. Fixes PR6463.

llvm-svn: 97924
2010-03-07 23:26:22 +00:00
Douglas Gregor 1ce52caf0d Reference binding via user-defined conversion can compute a binding
that is not reference-related (because it requires another implicit
conversion to which we can find). Fixes PR6483.

llvm-svn: 97922
2010-03-07 23:17:44 +00:00
Douglas Gregor 6de584c2d2 When declaring a catch block in C++, make sure that the type being
caught can be copy-initialized and destructed. Fixes PR6518.

llvm-svn: 97853
2010-03-05 23:38:39 +00:00
Rafael Espindola 70e040d552 During codegen assert that any copy assignment, destructor or constructor that
we need to synthesize has been marked as used by Sema.

Change Sema to avoid these asserts.

llvm-svn: 97589
2010-03-02 21:28:26 +00:00
Douglas Gregor beab56e923 Skip dependent virtual base classes; fixes PR6413.
llvm-svn: 97291
2010-02-27 00:25:28 +00:00
Douglas Gregor 422f155ca3 Don't try to finalize an ill-formed variable or one whose class type is ill-formed. Fixes PR6421
llvm-svn: 97152
2010-02-25 18:11:54 +00:00
John McCall 65eb879d22 Catch more uses of uninitialized implicit conversion sequences.
When diagnosing bad conversions, skip the conversion for ignored object
arguments.  Fixes PR 6398.

llvm-svn: 97090
2010-02-25 01:37:24 +00:00
Douglas Gregor 36c569fb33 Eliminate the default arguments to ASTContext::getFunctionType(),
fixing up a few callers that thought they were propagating NoReturn
information but were in fact saying something about exception
specifications.

llvm-svn: 96766
2010-02-21 22:15:06 +00:00
John McCall d8d0d43fa3 Support local namespace aliases and permit them to be instantiated.
llvm-svn: 96335
2010-02-16 06:53:13 +00:00
Chandler Carruth 284bb2ec72 Defer covariance checks for dependent types. Add test cases that also ensure
they are re-checked on instantiation.

llvm-svn: 96217
2010-02-15 11:53:20 +00:00
Douglas Gregor f40863caff Work around an annoying, non-standard optimization in the glibc
headers, where malloc (and many other libc functions) are declared
with empty throw specifications, e.g.,  

  extern void *malloc (__SIZE_TYPE__ __size) throw () __attribute__
  ((__malloc__)) ;

The C++ standard doesn't seem to allow this, and redeclaring malloc as
the standard permits (as follows) resulted in Clang (rightfully!)
complaining about mis-matched exception specifications.

  void *malloc(size_t size);

We work around this by silently propagating an empty throw
specification "throw()" from a function with C linkage declared in a
system header to a redeclaration that has no throw specifier.

Ick.

llvm-svn: 95969
2010-02-12 07:32:17 +00:00
Douglas Gregor 50dc219e8b When we have a dependent direct initializer but not a dependent
variable type, we can (and should) still check for completeness of the
variable's type. Do so, to work around an assertion that shows up in
Boost's shared_ptr.

llvm-svn: 95934
2010-02-11 22:55:30 +00:00
Douglas Gregor 4a62bdfb2a Remove another redundant ASTContext parameter
llvm-svn: 95843
2010-02-11 01:30:34 +00:00
Douglas Gregor d505812422 Eliminate a bunch of unnecessary ASTContexts from members functions of
Decl subclasses. No functionality change.

llvm-svn: 95841
2010-02-11 01:19:42 +00:00
Douglas Gregor 0a5a2216e2 Eliminate the ASTContext parameter from RecordDecl::getDefinition()
and CXXRecordDecl::getDefinition(); it's totally unnecessary. No
functionality change.

llvm-svn: 95836
2010-02-11 01:04:33 +00:00
John McCall 5b0829a321 Improve access control diagnostics. Perform access control on member-pointer
conversions.  Fix an access-control bug where privileges were not considered
at intermediate points along the inheritance path.  Prepare for friends.

llvm-svn: 95775
2010-02-10 09:31:12 +00:00
John McCall bc077cf589 Thread a source location into the template-argument deduction routines. There
may be some other places that could take advantage of this new information,
but I haven't really looked yet.

llvm-svn: 95600
2010-02-08 23:07:23 +00:00
Anders Carlsson a7bcade229 Add attributes to namespace decls.
llvm-svn: 95510
2010-02-07 01:09:23 +00:00
Douglas Gregor 1aa3edbb99 A function declarator with a non-identifier name in an anonymous class
is a constructor for that class, right? Fixes PR6238. 

llvm-svn: 95367
2010-02-05 06:12:42 +00:00
John McCall 67da35c832 Extract a common structure for holding information about the definition
of a C++ record.  Exposed a lot of problems where various routines were
silently doing The Wrong Thing (or The Acceptable Thing in The Wrong Order)
when presented with a non-definition.  Also cuts down on memory usage.

llvm-svn: 95330
2010-02-04 22:26:26 +00:00
John McCall 03c4848bf4 Mark dtors for parameter variables and eliminate some redundant type munging.
llvm-svn: 95079
2010-02-02 09:10:11 +00:00
John McCall 6781b05a92 Access control for implicit destructor calls. Diagnostic could be orders of
magnitude clearer.

llvm-svn: 95078
2010-02-02 08:45:54 +00:00
Sebastian Redl 5ca7984bb4 In C++, an initializer on a variable doesn't necessarily mean it's the definition. With that in mind, rename getDefinition to getAnyInitializer (to distinguish it from getInit) and reimplement it in terms of isThisDeclarationADefinition. Update all code to use this new function.
llvm-svn: 94999
2010-02-01 20:16:42 +00:00
Eli Friedman a682427e42 Switch expressions like T() and T(1,2) over to new-style initialization. I'm
not quite sure what we want to do about the AST representation; comments
welcome.

llvm-svn: 94967
2010-01-31 20:58:15 +00:00
Douglas Gregor 7ae2d7758f Rework base and member initialization in constructors, with several
(necessarily simultaneous) changes:

  - CXXBaseOrMemberInitializer now contains only a single initializer
    rather than a set of initialiation arguments + a constructor. The
    single initializer covers all aspects of initialization, including
    constructor calls as necessary but also cleanup of temporaries
    created by the initializer (which we never handled
    before!).

  - Rework + simplify code generation for CXXBaseOrMemberInitializers,
    since we can now just emit the initializer as an initializer.

  - Switched base and member initialization over to the new
    initialization code (InitializationSequence), so that it

  - Improved diagnostics for the new initialization code when
    initializing bases and members, to match the diagnostics produced
    by the previous (special-purpose) code.

  - Simplify the representation of type-checked constructor initializers in
    templates; instead of keeping the fully-type-checked AST, which is
    rather hard to undo at template instantiation time, throw away the
    type-checked AST and store the raw expressions in the AST. This
    simplifies instantiation, but loses a little but of information in
    the AST.

  - When type-checking implicit base or member initializers within a
    dependent context, don't add the generated initializers into the
    AST, because they'll look like they were explicit.

  - Record in CXXConstructExpr when the constructor call is to
  initialize a base class, so that CodeGen does not have to infer it
  from context. This ensures that we call the right kind of
  constructor.

There are also a few "opportunity" fixes here that were needed to not
regress, for example:

  - Diagnose default-initialization of a const-qualified class that
    does not have a user-declared default constructor. We had this
    diagnostic specifically for bases and members, but missed it for
    variables. That's fixed now.

  - When defining the implicit constructors, destructor, and
    copy-assignment operator, set the CurContext to that constructor
    when we're defining the body.

llvm-svn: 94952
2010-01-31 09:12:51 +00:00
Douglas Gregor d1e08648c6 Fix reference-binding when we have a reference to const volatile type;
previously, we were allowing this to bind to a temporary. Now, we
don't; add test-cases and improve diagnostics.

llvm-svn: 94831
2010-01-29 19:39:15 +00:00
John McCall 2cb941642e Access control for surrogate function calls. Required a moderately gross hack
to get the access bits set properly in conversion sets.

llvm-svn: 94744
2010-01-28 07:38:46 +00:00
Douglas Gregor 3edc4d5ec3 Fix a major oversight in the comparison of standard conversion
sequences, where we would occasionally determine (incorrectly) that
one standard conversion sequence was a proper subset of another when,
in fact, they contained completely incomparable conversions. 

This change records the types in each step within a standard
conversion sequence, so that we can check the specific comparison
types to determine when one sequence is a proper subset of the
other. Fixes this testcase (thanks, Anders!), which was distilled from
PR6095 (also thanks to Anders).

llvm-svn: 94660
2010-01-27 03:51:04 +00:00
John McCall b89836b6db Pass access specifiers around in overload resolution.
llvm-svn: 94485
2010-01-26 01:37:31 +00:00
Anders Carlsson 7caa4cb87a No need to canonicalize the type and use dyn_cast. Also, correctly diagnose trying to override a function returning an lvalue reference with a function overriding an rvalue reference.
llvm-svn: 94183
2010-01-22 17:37:20 +00:00
Chandler Carruth df7fd5f8d5 Fix an obvious goof that caused us to only see the top level of return types
when checking for covariance. Added some fun test cases, fixes PR6110.

This felt obvious enough to just commit. ;] Let me know if anything needs
tweaking.

llvm-svn: 94173
2010-01-22 13:07:41 +00:00
John McCall ad371258aa Give UnresolvedSet the ability to store access specifiers for each declaration.
Change LookupResult to use UnresolvedSet.  Also extract UnresolvedSet into its
own header and make it templated over an inline capacity.

llvm-svn: 93959
2010-01-20 00:46:10 +00:00
Douglas Gregor a3b624aa55 In a mem-initializer, a nested-name-specifier followed by an
identifier always names a type. In the case of a dependent
nested-name-specifier, build a TypenameType to describe the dependent
base type. I'd like to move more of this behavior up into the parser,
but this fixes PR6062.

llvm-svn: 93871
2010-01-19 06:46:48 +00:00
Douglas Gregor 9de54ea41b Reimplement constructor declarator parsing to cope with template-ids
that name constructors, the endless joys of out-of-line constructor
definitions, and various other corner cases that the previous hack
never imagined. Fixes PR5688 and tightens up semantic analysis for
constructor names.

Additionally, fixed a problem where we wouldn't properly enter the
declarator scope of a parenthesized declarator. We were entering the
scope, then leaving it when we saw the ")"; now, we re-enter the
declarator scope before parsing the parameter list.

Note that we are forced to perform some tentative parsing within a
class (call it C) to tell the difference between

  C(int); // constructor

and

  C (f)(int); // member function

which is rather unfortunate. And, although it isn't necessary for
correctness, we use the same tentative-parsing mechanism for
out-of-line constructors to improve diagnostics in icky cases like:

  C::C C::f(int); // error: C::C refers to the constructor name, but
                  // we complain nicely and recover by treating it as
                  // a type.

llvm-svn: 93322
2010-01-13 17:31:36 +00:00
John McCall 6a61b5203d Record some basic information about bad conversion sequences. Use that
information to feed diagnostics instead of regenerating it.  Much room for
improvement here, but fixes some unfortunate problems reporting on method calls.

llvm-svn: 93316
2010-01-13 09:16:55 +00:00
Alexis Hunt c88db06565 Implement semantic checking for C++ literal operators.
This now rejects literal operators that don't meet the requirements.
Templates are not yet checked for.

llvm-svn: 93315
2010-01-13 09:01:02 +00:00
Douglas Gregor 08dc584c3f Remove broken fix-it when a default function argument has been
redefined. There's a FIXME with an apology about why we don't try to
do better here. Fixes <rdar://problem/7513023>.

llvm-svn: 93274
2010-01-13 00:12:48 +00:00
John McCall ad907777e8 So I was sitting around, trying vainly to think of something to commit, and then
I said to myself, self, why don't you go add a couple of parameters to a method
and then fail to use them, and I thought that sounded like a pretty good idea,
so I did it.

llvm-svn: 93233
2010-01-12 07:18:19 +00:00
John McCall 0d1da2298a Introduce a specific representation for the ambiguous implicit conversion
sequence.  Lots of small relevant changes.  Fixes some serious problems with
ambiguous conversions;  also possibly improves associated diagnostics.

llvm-svn: 93214
2010-01-12 00:44:57 +00:00
Douglas Gregor 411e5acce1 Eliminate an embarrassing performance regression in C/ObjC, where we
were performing name lookup for template names in C/ObjC and always
finding nothing. Turn off such lookup unless we're in C++ mode, along
with the check that determines whether the given identifier is a
"current class name", and assert that we don't make this mistake
again.

llvm-svn: 93207
2010-01-11 23:29:10 +00:00
Douglas Gregor 133bc74d7a Add support for out-of-line definitions of conversion function
templates. Previously, a little thinko in the code that replaced a
conversion function template with its redeclaration was causing some
very weird lookup behavior.

llvm-svn: 93166
2010-01-11 18:53:25 +00:00
Douglas Gregor ea0a0a9b3f Implement name lookup for conversion function template specializations
(C++ [temp.mem]p5-6), which involves template argument deduction based
on the type named, e.g., given

  struct X { template<typename T> operator T*(); } x;

when we call

  x.operator int*();

we perform template argument deduction to determine that T=int. This
template argument deduction is needed for template specialization and
explicit instantiation, e.g.,

  template<> X::operator float*() { /* ... */ }

and when calling or otherwise naming a conversion function (as in the
first example). 

This fixes PR5742 and PR5762, although there's some remaining ugliness
that's causing out-of-line definitions of conversion function
templates to fail. I'll look into that separately.

llvm-svn: 93162
2010-01-11 18:40:55 +00:00
John McCall 12f97bc48a Change the printing of OR_Deleted overload results to print all the candidates,
not just the viable ones.  This is reasonable because the most common use of
deleted functions is to exclude some implicit conversion during calls;  users
therefore will want to figure out why some other options were excluded.

Started sorting overload results.  Right now it just sorts by location in the
translation unit (after putting viable functions first), but we can do better than
that.

Changed bool OnlyViable parameter to PrintOverloadCandidates to an enum for better
self-documentation.

llvm-svn: 92990
2010-01-08 04:41:39 +00:00
Douglas Gregor 43a0857631 When we typo-correct a base class initializer, point to the base class
specifier that we corrected to.

llvm-svn: 92878
2010-01-07 00:26:25 +00:00
Douglas Gregor 6da83624e4 Whenever we emit a typo-correction diagnostic, also emit a note
pointing to the declaration that we found that has that name (if it is
unique).

llvm-svn: 92877
2010-01-07 00:17:44 +00:00
Douglas Gregor ccecc1bb43 Fix marking of virtual members for nested classes whose first non-pure virtual function has a body inlined in the class
llvm-svn: 92855
2010-01-06 20:27:16 +00:00
John McCall fd0b2f8fe4 Improve the diagnostics used to report implicitly-generated class members
as parts of overload sets.  Also, refer to constructors as 'constructors'
rather than functions.

Adjust a lot of tests.

llvm-svn: 92832
2010-01-06 09:43:14 +00:00
Douglas Gregor 0a0f04dcb9 Make our marking of virtual members functions in a class be
deterministic and work properly with templates. Once a class that
needs a vtable has been defined, we now do one if two things:

  - If the class has no key function, we place the class on a list of
    classes whose virtual functions will need to be "marked" at the
    end of the translation unit. The delay until the end of the
    translation unit is needed because we might see template
    specializations of these virtual functions.
  - If the class has a key function, we do nothing; when the key
    function is defined, the class will be placed on the
    aforementioned list.

At the end of the translation unit, we "mark" all of the virtual
functions of the classes on the list as used, possibly causing
template instantiation and other classes to be added to the
list. This gets LLVM's lib/Support/CommandLine.cpp compiling again.

llvm-svn: 92821
2010-01-06 04:44:19 +00:00
Douglas Gregor a318efd1f2 Improve key-function computation for templates. In particular:
- All classes can have a key function; templates don't change that.
  non-template classes when computing the key function.
  - We always mark all of the virtual member functions of class
  template instantiations. 
  - The vtable for an instantiation of a class template has weak
  linkage. 

We could probably use available_externally linkage for vtables of
classes instantiated by explicit instantiation declarations (extern
templates), but GCC doesn't do this and I'm not 100% that the ABI
permits it.

llvm-svn: 92753
2010-01-05 19:06:31 +00:00
Anders Carlsson 0c9dd8476f It's OK for a return type to be incomplete if it's being defined.
llvm-svn: 92367
2009-12-31 18:54:35 +00:00
Anders Carlsson e60365b7b5 Make sure that an overriding return type is complete before checking if it's covariant. Fixes PR5920.
llvm-svn: 92365
2009-12-31 18:34:24 +00:00
Douglas Gregor 15e77a2fd3 Typo correction for C++ base and member initializers, e.g.,
test/FixIt/typo.cpp:41:15: error: initializer 'base' does not name a non-static
      data member or base class; did you mean the base class 'Base'?
  Derived() : base(),
              ^~~~
              Base
test/FixIt/typo.cpp:42:15: error: initializer 'ember' does not name a non-static
      data member or base class; did you mean the member 'member'?
              ember() { }
              ^~~~~
              member

llvm-svn: 92355
2009-12-31 09:10:24 +00:00
Chris Lattner b0283c06b2 improve diagnostics for case when a field type is unknown by
not emitting a follow-on error about 'int', which the user
never wrote.  PR5924.

llvm-svn: 92339
2009-12-31 03:10:55 +00:00
Chandler Carruth 607f38e05f Correctly refer to element CVR qualifications when determining if a type is
more or less cv-qualified than another during implicit conversion and overload
resolution ([basic.type.qualifier] p5). Factors the logic out of template
deduction and into the ASTContext so it can be shared.

This fixes several aspects of PR5542, but not all of them.

llvm-svn: 92248
2009-12-29 07:16:59 +00:00
Eli Friedman 11c7b15148 Attempted fix for PR5884; this code will be dead soon, but this fix should
help for the moment.

llvm-svn: 92165
2009-12-25 23:59:21 +00:00
Benjamin Kramer d20ef75b91 Remove some dead variables clang-analyzer found.
llvm-svn: 92162
2009-12-25 15:43:36 +00:00
Eli Friedman fddc26cc64 Make copy constructor elimination work in more cases; the case in question
here affects clang-on-clang.

llvm-svn: 92151
2009-12-24 23:33:34 +00:00
Douglas Gregor 33636e66c7 Egregious, disgusting workaround for PR5866. We need to rework how we
keep track of friends within templates, which will provide a real for
PR5866. For now, this makes sure we don't do something entirely stupid
with friends of specializations.

llvm-svn: 92143
2009-12-24 20:56:24 +00:00
Daniel Dunbar 2db411f5a7 Fix -Asserts warning.
llvm-svn: 92137
2009-12-24 19:19:26 +00:00
Douglas Gregor 684d7bdc43 Allow the first parameter of operator new to be a cv-qualified
size_t. Also, fix an issue with initialization of parameters in calls,
where we weren't removing the cv-qualifiers on the parameter type
itself. Fixes PR5823.

llvm-svn: 91941
2009-12-22 23:42:49 +00:00
Douglas Gregor b6ea60872d Switch Sema::AddCXXDirectInitializerToDecl over to InitializationSequence
llvm-svn: 91927
2009-12-22 22:17:25 +00:00
Eli Friedman 5f101b95c6 Switch default arguments over to InitializationSequence.
llvm-svn: 91883
2009-12-22 02:46:13 +00:00
John McCall 463e10cbd5 Stop diagnosing the use of inner classes as friends. ddunbar asked whether
this was useful, and on review Doug and I decided it was probably on the level
of a bug in the standard and therefore not worth a warning even in -pedantic.
If someone disagrees and urgently wants clang++ to warn about this in strict
c++98 mode, we can talk about it.

llvm-svn: 91868
2009-12-22 00:59:39 +00:00
Douglas Gregor 507eb87f05 Eliminate the ASTContext argument to CXXConstructorDecl::isCopyConstructor, since the context is available in the Decl
llvm-svn: 91862
2009-12-22 00:34:07 +00:00
John McCall b5a0d316dc Reorganize the base-lookup bits of ActOnMemInitializer in order to better
support diagnostics and error recovery.

llvm-svn: 91825
2009-12-21 10:41:20 +00:00
John McCall 6df5fef637 Refactor to remove more dependencies on PreDeclaratorDC. I seem to have made
the redeclaration problems in the [temp.explicit]p3 testcase worse, but I can
live with that;  they'll need to be fixed more holistically anyhow.

llvm-svn: 91771
2009-12-19 10:49:29 +00:00
John McCall 1f4ee7bd2f Just push a new scope when parsing an out-of-line variable definition.
Magically fixes all the terrible lookup problems associated with not pushing
a new scope.  Resolves an ancient xfail and an LLVM misparse.

llvm-svn: 91769
2009-12-19 09:28:58 +00:00
Douglas Gregor e1314a64b8 Switch the initialization required by return statements over to the
new InitializationSequence. This fixes some bugs (e.g., PR5808),
changed some diagnostics, and caused more churn than expected. What's
new:

  - InitializationSequence now has a "C conversion sequence" category
    and step kind, which falls back to
  - Changed the diagnostics for returns to always have the result type
    of the function first and the type of the expression second.
    CheckSingleAssignmentConstraints to peform checking in C. 
  - Improved ASTs for initialization of return values. The ASTs now
    capture all of the temporaries we need to create, but
    intentionally do not bind the tempoary that is actually returned,
    so that it won't get destroyed twice.
  - Make sure to perform an (elidable!) copy of the class object that
    is returned from a class.
  - Fix copy elision in CodeGen to properly see through the
    subexpressions that occur with elidable copies.
  - Give "new" its own entity kind; as with return values and thrown
    objects, we don't bind the expression so we don't call a
    destructor for it.

Note that, with this patch, I've broken returning move-only types in
C++0x. We'll fix it later, when we tackle NRVO.

llvm-svn: 91669
2009-12-18 05:02:21 +00:00
Eli Friedman 5dd02a0f65 Correctly calcluate abstract-ness in the case where an implicitly declared
method overrides a pure virtual method.

llvm-svn: 91558
2009-12-16 20:00:27 +00:00
Douglas Gregor 4f4b186215 When value-initializing a class with no user-defined constructors but
with a non-trivial default constructor, zero-initialize the storage
and then call the default constructor. Fixes PR5800.

llvm-svn: 91548
2009-12-16 18:50:27 +00:00
Douglas Gregor 7c3bbdfd28 Fix semantic diagnostics that embed English works, from Nicola Gigante!
llvm-svn: 91503
2009-12-16 03:45:30 +00:00
John McCall 0db42252f7 Successive anonymous namespaces name the same scope. I misinterpreted the
standard the last time.  Fixes PR5766.

llvm-svn: 91493
2009-12-16 02:06:49 +00:00
Douglas Gregor 85dabae6ad Switch the C++ new expression over to InitializationSequence, rather
than using its own partial implementation of initialization. 

Switched CheckInitializerTypes over to
InitializedEntity/InitializationKind, to help move us closer to
InitializationSequence.

Added InitializedEntity::getName() to retrieve the name of the entity,
for diagnostics that care about such things.

Implemented support for default initialization in
InitializationSequence.

Clean up the determination of the "source expressions" for an
initialization sequence in InitializationSequence::Perform.

Taught CXXConstructExpr to store more location information.

llvm-svn: 91492
2009-12-16 01:38:02 +00:00
Anders Carlsson 6e997b2993 ShouldDestroyTemporaries? I don't think so.
llvm-svn: 91450
2009-12-15 20:51:39 +00:00
Douglas Gregor 906db8a546 Fix some diagnostic-related FIXMEs, from Nicola Gigante
llvm-svn: 91433
2009-12-15 16:44:32 +00:00
Anders Carlsson 7e0b207e54 More improvements to checking allocation and deallocation functions.
llvm-svn: 91244
2009-12-13 17:53:43 +00:00
Anders Carlsson 60659a8249 Fix tyop.
llvm-svn: 91204
2009-12-12 02:43:16 +00:00
Anders Carlsson e363c8e1df Correctly diagnose [basic.stc.dynamic.allocation]p1
llvm-svn: 91190
2009-12-12 00:32:00 +00:00