After this I will set the default back to F_None. The advantage is that
before this patch forgetting to set F_Binary would corrupt a file on windows.
Forgetting to set F_Text produces one that cannot be read in notepad, which
is a better failure mode :-)

llvm-svn: 202052

During the LTO phase LICM will move loop invariant global variables out of loops
(informed by GlobalModRef). This makes more loops countable presenting
opportunity for the loop vectorizer.

Adding the loop vectorizer improves some TSVC benchmarks and twolf/ref dataset
(5%) on x86-64.

radar://15970632

llvm-svn: 202051

llvm-svn: 202050

targets. Just big endian (mips-... and mips64-...)

llvm-svn: 202049

llvm-svn: 202048

We were previously checking at every destructor declaration, but that was a bit
excessive. Since the deleting destructor is emitted with the vtable, do the
check when the vtable is marked used.

Differential Revision: http://llvm-reviews.chandlerc.com/D2851

llvm-svn: 202046

llvm-svn: 202045

Failure is not reproducible on ToT LLDB locally or on the buildbot.

llvm.org/pr18313

llvm-svn: 202043

This will make it easier to switch the default to being binary files.

llvm-svn: 202042

llvm-svn: 202041

llvm-svn: 202040

llvm-svn: 202039

llvm-svn: 202037

Patch by Matthew Gardiner

llvm-svn: 202036

Patch by Marcello Maggioni!

llvm-svn: 202035

llvm-svn: 202034

[asan] simplify the code that compute the shadow offset; get rid of two internal flags that allowed to override it. The tests pass, but still this change might break asan on some platform not covered by tests. If you see this, please submit a fix with a test.

llvm-svn: 202033

No version of clang installed by XCode, ToT or otherwise recognizes the flag
'-no-intel-extensions' and more recent versions error out on it.

Prospectively conditionalize the flag behind "icc" to fix the mainline build
with clang.

llvm-svn: 202030

check-tools.pl was hard-wired to expect the Apple compiler version string which
doesn't make much sense, especially for a LLVM project.

The correctness and value of these compiler version checks still isn't clear
following this build fix but they are retained for now.

llvm-svn: 202029

Differential Revision: http://llvm-reviews.chandlerc.com/D2826

llvm-svn: 202028

gcc/g++ don't exist on modern OS X development environments so the failed
checks for gcc were causing the build to break even with 'make compiler=clang'.

(The rationale for the check refers to a file 'touch-test.c' which no longer
exists in the codebase.)

llvm-svn: 202027

llvm-svn: 202026

llvm-svn: 202025

llvm-svn: 202024

[CMake] Better support for COMPILER_RT_ENABLE_WERROR. Make sure compiler-rt libraries are build by 'make all' command

llvm-svn: 202023

llvm-svn: 202022

llvm-svn: 202021

Before this patch we would only use the fist occurance of a matcher
function in the documentation, for example leaving out
hasType(Matcher<QualType>).

llvm-svn: 202019

llvm-svn: 202018

1. Move internal functions into ASTMatchersInternal.
2. Adapt dump_ast_matchers.py to the new VariadicOperatorMatcherFunc
   signature.
3. Update the actual docs with the updated tool / code.

llvm-svn: 202017

Failing on the bots for an unknown reason.

llvm-svn: 202016

llvm-svn: 202015

llvm-svn: 202014

llvm-svn: 202013

The isPure() CXXMethodDecl matcher matches pure method declaration like "A::x"
in this example:

class A {
  virtual void x() = 0;
}

Patch by Konrad Kleine.

llvm-svn: 202012

AdjustStackSizeLinux() is used in Lsan, Tsan and Msan non-Linux-specific code so it seems it should have more generic name and moved to a common scope.
Renamed to AdjustStackSize.
Patch by Viktor Kutuzov.

llvm-svn: 202011

In 'obsequi' we have a scop in which the current dead code elimination works,
but the generated code is way too complex. To avoid this trouble (and to not
disable the DCE entirely) we add an additional approximative step before
the actual dead code elimination. This should fix one of the two current
nightly-test issues.

Polly could be improved to handle 'obsequi' by teaching it to introduce only a
single parameter for (%1 and zext %1) which halves the number of parameters and
allows polly to derive a simpler representation for the set of live iterations.
However, this needs some time to investigate.

I will commit a test case as soon as we have a reduced one.

llvm-svn: 202010

llvm-svn: 202009

error: address of array 'tctx->name' will always evaluate to 'true'

llvm-svn: 202008

Fix an unintentional stdin read in the darwin-asan-nofortify.c test and replace
it with an explicit test for multiple -E inputs passed to the driver.

Noticed while working on the in-process driver patch.

llvm-svn: 202007