- <rdar://problem/6551577> [ccc] require -x with -

llvm-svn: 63901

Refactor my previous change to maintain the distinction between AliasAnalysis and BasicAliasAnalysis.  This involves some wider changes because it
folds away some never-used methods.

llvm-svn: 63900

canonicalize by template parameter depth, index, and name, and the
unnamed version of a template parameter serves as the canonical.

TemplateTypeParmDecl no longer needs to inherit from
TemplateParmPosition, since depth and index information is present
within the type.

llvm-svn: 63899

Ignore dbg intrinsics while propagating conditional expression info. Take 2.

llvm-svn: 63898

llvm-svn: 63897

llvm-svn: 63896

llvm-svn: 63895

errors to 'fatal' error severity.

llvm-svn: 63894

Add 'AppendValue' to the list of magic CF function names that cause a tracked object to escape.  Fixes <rdar://problem/6560661>.

llvm-svn: 63891

isAsCheapAsMove instructions can have register src operands. Check if they are really re-materializable.
This fixes sse.expandfft and sse.stepfft.

llvm-svn: 63890

llvm-svn: 63889

Revert rev. 63876. It is causing llvm-gcc bootstrap failure.

llvm-svn: 63888

llvm-svn: 63887

llvm-svn: 63886

llvm-svn: 63883

compilers that don't support those flags.  This hopefully will help
gcc 3.X compile this code.  http://llvm.org/PR3487

llvm-svn: 63882

llvm-svn: 63881

llvm-svn: 63880

llvm-svn: 63879

llvm-svn: 63878

Also, put Objective-C protocols into their own identifier
namespace. Otherwise, we find protocols when we don't want to in C++
(but not in C).

llvm-svn: 63877

llvm-svn: 63876

llvm-svn: 63875

llvm-svn: 63874

llvm-svn: 63872

the documentation to the contrary).

llvm-svn: 63871

Patch by Daniel Shelton!

llvm-svn: 63870

llvm-svn: 63869

However, the cause still remains: the Decl is linked into the chain of its DeclContext and remains there despite being deleted.

llvm-svn: 63868

This shrinks OwningResult by one pointer. Since it is no longer larger than OwningPtr, merge the two.
This leads to simpler client code and speeds up my benchmark by 2.7%.
For some reason, this exposes a previously hidden bug, causing a regression in SemaCXX/condition.cpp.

llvm-svn: 63867

llvm-svn: 63866

scalar/complex/aggregate cases.
 - Currently disabled for x86_64, triggering a misoptimization
   (PR3489).

llvm-svn: 63864

 - No intended functionality change, this is essentially enabling
   direct passing of complex and aggregate values, which no ABI is
   using.

llvm-svn: 63863

really solve the problem.  Also it changed the
name from libLLVMgold to LLVMgold (correcting
the name resulted in the original problem
suddenly reappearing).

llvm-svn: 63861

dependencies between libraries due to __dso_handle
when doing "make install".  I don't know what the
LINK_COMPONENTS stuff is about, so I left it alone.

llvm-svn: 63860

value.
 - No functionality change.

llvm-svn: 63859

 - No intended functionality change, all current ABI implementations
   were only using indirect for complex/aggregate types, which were
   being passed indirectly with the Direct ABIInfo kind.

llvm-svn: 63858

Machine LICM increases register pressure and it almost always increase code size. For now, disable it for optimizing for size.

llvm-svn: 63856

llvm-svn: 63855

llvm-svn: 63854