When -ffast-math is in effect (on Linux, at least), clang defines
__FINITE_MATH_ONLY__ > 0 when including <math.h>. This causes the
preprocessor to include <bits/math-finite.h>, which renames the sqrt functions.
For instance, "sqrt" is renamed as "__sqrt_finite". 

This patch adds the 3 new names in such a way that they will be treated
as equivalent to their respective original names.

llvm-svn: 182739

Unify handling of whitespace when breaking protruding tokens with other
whitespace replacements.

As a side effect, the BreakableToken structure changed significantly:
- have a common base class for single-line breakable tokens, as they are
  much more similar
- revamp handling of multi-line comments; we now calculate the
  information about lines in multi-line comments similar to normal
  tokens, and always issue replacements

As a result, we were able to get rid of special casing of trailing
whitespace deletion for comments in the whitespace manager and the
BreakableToken and fixed bugs related to tab handling and escaped
newlines.

llvm-svn: 182738

We were handling only local types in inline External functions before.

llvm-svn: 182737

UseAuto used to replace declarion lists with new expressons where some
variable were not initialized with new.
This fix checks that every DeclStmt has a VarDecl with an initializer and it
also ensures that all declarations have the same type.

Added tests for multiple declarations and for typedefs.

llvm-svn: 182736

This removes a duplicate from Decl.cpp and a followup patch will use
isExternallyVisible.

llvm-svn: 182735

llvm-svn: 182734

Before:
unsigned OriginalStartColumn = SourceMgr.getSpellingColumnNumber(
    Current.FormatTok.getStartOfNonWhitespace()) -
                               1;

After:
unsigned OriginalStartColumn =
    SourceMgr.getSpellingColumnNumber(
        Current.FormatTok.getStartOfNonWhitespace()) -
    1;

llvm-svn: 182733

llvm-svn: 182732

In general, we like to avoid line breaks like:

  ...
  SomeParameter, OtherParameter).DoSomething(
  ...

as they tend to make code really hard to read (how would you even indent the
next line?). Previously we have implemented this in a hacky way, which has now
shown to lead to problems. This fixes a few weird looking formattings, such as:

Before:
aaaaa(
    aaaaa(aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa, aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa)
        .aaaaa(aaaaa),
    aaaaaaaaaaaaaaaaaaaaa);
After:
aaaaa(aaaaa(aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa,
            aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa).aaaaa(aaaaa),
      aaaaaaaaaaaaaaaaaaaaa);

llvm-svn: 182731

In this CL all old tests are removed and one LIT test is added.

llvm-svn: 182730

If -fsanitize=leak is specified, link the program with the
LeakSanitizer runtime. Ignore this option when -fsanitize=address is specified,
because AddressSanitizer has this functionality built in.

llvm-svn: 182729

llvm-svn: 182728

Now compiler-rt tests run correctly if compiler-rt is checked out into
arbitrary directory (not necessarily projects/compiler-rt).
Patch by Greg Fitzgerald!

llvm-svn: 182726

isConsecutiveLS is a slightly more general form of
SelectionDAG::isConsecutiveLoad. Aside from also handling stores, it also does
not assume equality of the chain operands is necessary. In the case of the PPC
backend, this chain condition is checked in a more general way by the
surrounding code.

Mostly, this part of the refactoring in preparation for supporting optimized
unaligned stores.

llvm-svn: 182723

llvm-svn: 182722

llvm-svn: 182721

llvm-svn: 182720

When expanding unaligned Altivec loads, we use the decremented offset trick to
prevent page faults. Unfortunately, if we have a sequence of consecutive
unaligned loads, this leads to suboptimal code generation because the 'extra'
load from the first unaligned load can be combined with the base load from the
second (but only if the decremented offset trick is not used for the first).
Search up and down the chain, through loads and token factors, looking for
consecutive loads, and if one is found, don't use the offset reduction trick.
These duplicate loads are later combined to yield the desired sequence (in the
future, we might want a more-powerful chain search, but that will require some
changes to allow the combiner routines to access the AA object).

This should complete the initial implementation of the optimized unaligned
Altivec load expansion. There is some refactoring that should be done, but
that will happen when the unaligned store expansion is added.

llvm-svn: 182719

llvm-svn: 182718

llvm-svn: 182717

Before:
  @{ NSFontAttributeNameeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee
     : regularFont, };

Now:
  @{ NSFontAttributeNameeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeee :
     regularFont, };

':'s in dictionary literals (and the corresponding {}s) are now marked as
TT_ObjCDictLiteral too, which makes further improvements to dict literal
layout possible.

llvm-svn: 182716

llvm-svn: 182715

llvm-svn: 182714

reject things like: "for (auto Entry : SomeStringMap)".  Previously
this would copy the value but not the tail allocated string data
(the key).

llvm-svn: 182713

stream.

llvm-svn: 182712

John noticed that the fix for pr15930 (r181981) didn't handle indirect
uses of local types. For example, a pointer to local struct, or a
function that returns it.

One way to implement this would be to recursively look for local
types. This would look a lot like the linkage computation itself for
types.

To avoid code duplication and utilize the existing linkage cache, this
patch just makes the computation of "type with no linkage but
externally visible because it is from an inline function"  part of the
linkage computation itself.

llvm-svn: 182711

llvm-svn: 182710

llvm-svn: 182709

The lvsl permutation control instruction is a function only of the alignment of
the pointer operand (relative to the 16-byte natural alignment of Altivec
vectors). As a result, multiple lvsl intrinsics where the operands differ by a
multiple of 16 can be combined.

llvm-svn: 182708

Unit test cases for -pre-RA-sched=source.

llvm-svn: 182706

Remove the old IR ordering mechanism and switch to new one.  Fix unit
test failures.

llvm-svn: 182704

Change SelectionDAG::getXXXNode() interfaces as well as call sites of
these functions to pass in SDLoc instead of DebugLoc.

llvm-svn: 182703

llvm-svn: 182702

Use a field in the SelectionDAGNode object to track its IR ordering.
This adds fields and utility classes without changing existing
interfaces or functionality.

llvm-svn: 182701

llvm-svn: 182700

When x is empty, x ## is suppressed, and when y gets expanded, the fact that it follows ## is not
available in the macro expansion result. The macro definition can be checked instead, the ## will
be available there regardless of what x expands to.

Fixes http://llvm.org/PR12767

Patch by Harald van Dijk!

llvm-svn: 182699

llvm-svn: 182698

llvm-svn: 182697

rather than assuming it lives in the path. Patch by Eitan Adler!

llvm-svn: 182696

Suggested by Jordan.

llvm-svn: 182695