llvm-svn: 187918

llvm-svn: 187917

-fdelayed-template-parsing mode. Patch by Will Wilson!

llvm-svn: 187916

llvm-svn: 187915

The problem was that an enum without closing semicolon could be associated as a forward enum
in an erroneous declaration, leading to the identifier being associated with the enum decl but
without a declaration actually referencing it.
This resulted in not having it serialized before serializing the identifier that is associated with.

Also prevent the ASTUnit from querying the serialized DeclID for an invalid top-level decl; it may not
have been serialized.

rdar://14539667

llvm-svn: 187914

.tbss section for zerofill thread locals. Make sure we do this
before emitting the zerofills.

Fixes PR15972.

llvm-svn: 187913

llvm-svn: 187912

llvm-svn: 187911

llvm-svn: 187910

llvm-svn: 187909

Correct logic bug in find optimization for vector<bool>.  This fixes http://llvm.org/bugs/show_bug.cgi?id=16816

llvm-svn: 187908

This allows llvm-tblgen to link successfully when compiling with clang.

Both MSBuild and CMake will automatically add advapi32 as part of a set
of other dlls comprising the win32 API to the link line, but CMake
doesn't do that when compiling with clang.  Until someone adds that info
to cmake upstream, this seems like a reasonable work around.

llvm-svn: 187907

llvm-svn: 187906

llvm-svn: 187905

llvm-svn: 187904

llvm-svn: 187903

llvm-svn: 187902

In particular, left braces after an enum declaration now occur on their
own line.  Further, when short ifs/whiles are allowed these no longer
cause the left brace to be on the same line as the if/while when a
brace is included.

Patch by Thomas Gibson-Robinson.

llvm-svn: 187901

llvm-svn: 187900

llvm-svn: 187899

llvm-svn: 187898

Before this patch we would align

long long int big[1024];

to 4 bytes on 32 bit systems. The problem is that we were only looking
at the element type when getLargeArrayMinWidth returned non zero.

llvm-svn: 187897

llvm-svn: 187896

llvm-svn: 187895

// rdar://14258334

llvm-svn: 187893

Before:
  #pragma omp reduction( | : var)
After:
  #pragma omp reduction(| : var)

llvm-svn: 187892

This field is just IsDefaulted && !IsDeleted; in all places it's used,
a simple check for isDefaulted() is superior anyway, and we were forgetting
to set it in a few cases.

Also eliminate CXXDestructorDecl::IsImplicitlyDefined, for the same reasons.

No intended functionality change.

llvm-svn: 187891

llvm-svn: 187890

llvm-svn: 187889

Patch by David Wiberg.

llvm-svn: 187888

llvm-svn: 187887

For now, builders in bb.pgr.jp are using it.

llvm-svn: 187886

If there's a race between a memory access and a free() call in the client program,
it can be reported as a use-after-free (if the access occurs after the free()) or an ordinary race
(if free() occurs after the access).
We've decided to use a single "race:" prefix for both cases instead of introducing a "use-after-free:" one,
because in many cases this allows us to keep a single suppression for both the use-after-free and free-after-use.

This may be misleading if the use-after-free occurs in a non-racy way (e.g. in a single-threaded program).
But normally such bugs shall not be suppressed.

llvm-svn: 187885

with lowering logic and a test.

llvm-svn: 187884

This follows the same lines as the integer code.  In the end it seemed
easier to have a second 4-bit mask in TSFlags to specify the compare-like
CC values.  That eats one more TSFlags bit than adding a CCHasUnordered
would have done, but it feels more concise.

llvm-svn: 187883

These instructions can also be used as comparisons with zero.

llvm-svn: 187882

llvm-svn: 187881

JumpThreading: Turn a select instruction into branching if it allows to thread one half of the select.

This is a common pattern coming out of simplifycfg generating gross code.

a:                                       ; preds = %entry
  %sel = select i1 %cmp1, double %add, double 0.000000e+00
  br label %b

b:
  %cond5 = phi double [ %sel, %a ], [ %sub, %entry ]
  %cmp6 = fcmp oeq double %cond5, 0.000000e+00
  br i1 %cmp6, label %if.then, label %if.end

becomes

a:
  br i1 %cmp1, label %b, label %if.then

b:
  %cond5 = phi double [ %sub, %entry ], [ %add, %a ]
  %cmp6 = fcmp oeq double %cond5, 0.000000e+00
  br i1 %cmp6, label %if.then, label %if.end

Skipping block b completely if possible.

llvm-svn: 187880

r187874 seems to have been missed by the build bot infrastructure, and
the subsequent commits to compiler-rt don't seem to be queuing up new
build requsets. Hopefully this will.

As it happens, having the space here is the more common formatting. =]

llvm-svn: 187879

llvm-svn: 187878