- Disables the freeing of the ASTContext and the TranslationUnit
   after parsing & sema.
 - Primarily for timing the impact on -fsyntax-only timings.

llvm-svn: 57643

and
http://llvm.org/bugs/show_bug.cgi?id=2893

llvm-svn: 57640

is to encode the state of the #pragma pack stack as an attribute when
the structure is declared. 

 - Extend PackedAttr to take an alignment (in bits), and reuse for
   both __attribute__((packed)) (which takes no argument, instead
   packing tightly (to "minimize the memory required") and for #pragma
   pack (which allows specification of the maximum alignment in
   bytes). __attribute__((packed)) is just encoded as Alignment=1.

   This conflates two related but different mechanisms, but it didn't
   seem worth another attribute.

 - I have attempted to follow the MSVC semantics as opposed to the gcc
   ones, since if I understand correctly #pragma pack originated with
   MSVC. The semantics are generally equivalent except when the stack
   is altered during the definition of a structure; its not clear if
   anyone does this in practice. See testcase if curious.

llvm-svn: 57623

typedef int f();
struct S {
   f *x; // incorrectly assuming this is function decl, leading to failed assertions.
};

llvm-svn: 57598

Note: One day, we should consider moving the actual diags to ObjCQualifiedIdTypesAreCompatible(), since it has more information on the actual problem. GCC currently emits slightly more instructive errors for some cases involving protocols. I added a FIXME to the code.

llvm-svn: 57529

llvm-svn: 57514

llvm-svn: 57511

 - Follows the MSVC (original) implementation, including support of
   pack(show) (useful for testing).
 - Implements support for named pack records which gcc seems to
   ignore (or implements incorrectly).
 - Not currently wired to anything, only functionality change is the
   type checking of the pragma.

llvm-svn: 57476

llvm-svn: 57391

Final phase of converting BlockDecls over to DeclContext. This is unfortunately a largish/complex diff, however it was necessry to pass all the current block tests.

llvm-svn: 57337

class C {
  static const int number = 50;
  static int arr[number];
};

Here's how it worked:
-GetTypeForDeclarator was called from both Sema::ActOnCXXMemberDeclarator and Sema::ActOnDeclarator.
-VariableArrayTypes are not uniqued so two VariableArrayTypes were created with the same DeclRefExpr.
-On exit they both tried to destroy that one DeclRefExpr.

The fix is not to use GetTypeForDeclarator from the Sema::ActOnCXXMemberDeclarator.

llvm-svn: 57313

This required changes to attach the compound statement later on (like we do for functions).

llvm-svn: 57304

- Modify BlockExpr to reference the BlockDecl.

This is "cleanup" necessary to improve our lookup semantics for blocks (to fix <rdar://problem/6272905> clang block rewriter: parameter to function not imported into block?).

Still some follow-up work to finish this (forthcoming).

llvm-svn: 57298

This also removes the ugly hack needed in CFG.cpp for subclassing DeclStmt to create a DeclStmt with one Decl*.

llvm-svn: 57275

llvm-svn: 57220

Simplify handling of direct initializers by letting Sema::AddInitializerToDecl handle conversions, instead of using Sema::ActOnCXXTypeConstructExpr.
Additional benefit is that diagnostics are the same for both direct-initialization and copy-initialization.

In the case of "int x( expression );":
-The Init expression of VarDecl 'x' will be the expression inside the parentheses.
-VarDecl::hasCXXDirectInitializer for VarDecl 'x' will return true to let clients distinguish from "int x = expression ;".

llvm-svn: 57219

When processing Objective-C foreach statements, first check to see if the statement has a DeclStmt with a single Decl.  Afterwards, use DeclStmt::getSolitaryDecl() to access that Decl (thus avoiding an assertion being triggered).  These changes remove an unneeded use of ScopedDecl::getNextDeclarator() and DeclStmt::getDecl().

llvm-svn: 57207

Use "unsigned" instead of "int" for i to remove a "comparison between unsigned and signed" warning (potential integer overflow).

llvm-svn: 57201

Use the DeclStmt::decl_iterator to get the first decl in a DeclStmt instead of using DeclStmt::getDecl().

llvm-svn: 57196

Add a FIXME until class constructors are supported.

llvm-svn: 57188

This is how this kind of initializers appear in the AST:
-The Init expression of the VarDecl is a functional type construction (of the VarDecl's type).
-The new VarDecl::hasCXXDirectInitializer() returns true.

e.g, for "int x(1);":
-VarDecl 'x' has Init with expression "int(1)" (CXXFunctionalCastExpr).
-hasCXXDirectInitializer() of VarDecl 'x' returns true.

A major benefit is that clients that don't particularly care about which exactly form was the initializer can handle both cases without special case code.
Note that codegening works now for "int x(1);" without any changes to CodeGen.

llvm-svn: 57178

constant lvalue.  Implement this in codegen by moving the code out of CGBuiltin
into EmitConstantExpr.

llvm-svn: 57163

which is checking for something that can be inconsistent with
what we can constant fold.

llvm-svn: 57159

condition as a constant even if the unevaluated side is a not a constant.
We don't do this when extensions are off, and we emit a warning when this 
happens:

t.c:22:11: warning: expression is not a constant, but is accepted as one by GNU extensions
short t = __builtin_constant_p(5353) ? 42 : somefunc();
          ^                                 ~~~~~~~~~~

suggestions for improvement are welcome.  This is obviously horrible, but
is required for real-world code.

llvm-svn: 57153

llvm-svn: 57142

llvm-svn: 57121

llvm-svn: 57019

llvm-svn: 56992

This is a temporary solution to help with the block rewriter (though it certainly has general utility).
Once DeclGroup's are implemented, this SourceLocation should be stored with it (since it applies to all the decls).

llvm-svn: 56985

 - Enabled for builtins which are always constant expressions
   (__builtin_huge_val*, __builtin_inf*, __builtin_constant_p,
   __builtin_classify_type, __builtin___CFStringMakeConstantString).

Added Builtin::Context::isConstantExpr.
 - Currently overly simply interface which only works for builtins
   whose constantexprness does not depend on their arguments.

CallExpr::isBuiltinConstantExpr now takes an ASTContext argument.

llvm-svn: 56983

versions of sprintf and friends.
 - Added FIXME that this mechanism should be generalized.

llvm-svn: 56962

This commit also includes some name changes in the blocks rewriter (no functionality change).

llvm-svn: 56955

llvm-svn: 56880

llvm-svn: 56879

llvm-svn: 56859

Fix <rdar://problem/6191148> [sema] Objective-C method lookup (at global scope) fails to handle overloaded selectors properly.

Long standing bug in Sema::ActOnInstanceMessage(). We now warn when messaging an "id" with multiple method signatures in scope. The diags are a little verbose, however they can be streamlined if necessary. 

llvm-svn: 56843

Needed to make sure the relaxed type checking only applies to "readonly" properties.

llvm-svn: 56838

llvm-svn: 56821

llvm-svn: 56800

Teach Sema::CheckAssignmentConstraints() to allow assignments between id and block pointer types (^{}).

llvm-svn: 56793