llvm-svn: 108663

llvm-svn: 108662

portable enough.
 - Downside is we now double dispatch through a stub function, but this isn't
   performance critical.

llvm-svn: 108661

StringPointers vector will also be used in the normal case to handle @file
arguments.

llvm-svn: 108660

llvm-svn: 108659

llvm-svn: 108658

fixed return types.

llvm-svn: 108657

linkage specification.  Not sure if this is the ideal fix, but I'm reasonably
sure it's correct vs. gcc.

llvm-svn: 108656

sequences, not just strings.

llvm-svn: 108655

llvm-svn: 108654

llvm-svn: 108653

llvm-svn: 108652

is a very reasonable position on life!

llvm-svn: 108651

is a very reasonable position on life!

llvm-svn: 108650

support macros.

llvm-svn: 108649

llvm-svn: 108648

llvm-svn: 108647

llvm-svn: 108646

llvm-svn: 108645

llvm-svn: 108644

The rationale is that we are copying the entire definition including
parameter names which may differ between the declaration and the
definition.

This is particularly important if any parameters are unnamed in the
declaration, as a DeclRef to an unnamed ParmVarDecl would cause the
pretty printer to produce invalid output.

llvm-svn: 108643

llvm-svn: 108642

llvm-svn: 108641

llvm-svn: 108640

Add a DAGCombine xform to fold away redundant float->double->float conversions around sqrt instructions.
I am assured by people more knowledgeable than me that there are no rounding issues in eliminating this.

This fixed <rdar://problem/8197504>.

llvm-svn: 108639

their call expressions synthetically have the "deduced" types based on their
first argument. We only insert conversions in the AST for arguments whose
values require conversion to match the value type expected. This keeps PR7600
closed by maintaining the return type, but avoids assertions due to unexpected
implicit casts making the type unsigned (test case added from Daniel).

The magic is moved into the codegen for the atomic builtin which inserts the
casts as needed at the IR level to raise the type to an integer suitable for
the LLVM intrinsic. This shouldn't cause any real change in functionality, but
now we can make the builtin be more truly polymorphic.

llvm-svn: 108638

LoopSplitter be run prior to register allocation.

Entirely for testing purposes at the moment.

llvm-svn: 108634

llvm-svn: 108633

try to evaluate the initializer as a constant.

llvm-svn: 108632

llvm-svn: 108631

incomplete union (PR5692) and incomplete enum, and fixes obscure
accepts-invalid on cast to incomplete struct.

llvm-svn: 108630

llvm-svn: 108629

llvm-svn: 108628

llvm-svn: 108627

llvm-svn: 108626

llvm-svn: 108625

allow invalid token pastes (when in -fms-extensions mode)
with -Wno-invalid-token-paste

llvm-svn: 108624

llvm-svn: 108623

llvm-svn: 108622

llvm-svn: 108621