Patch by Sanjoy Das!

llvm-svn: 133910

of a single if block. This is really annoying to track down and test.
Silly changes to the test case caused it to stop showing up. I wish
there were a more concrete way of asserting that a note attaches to the
intended diagnostic.

This fixes PR10195.

llvm-svn: 133907

arithmetic into a couple of common routines. Use these to make the
messages more consistent in the various contexts, especially in terms of
consistently diagnosing binary operators with invalid types on both the
left- and right-hand side. Also, improve the grammar and wording of the
messages some, handling both two pointers and two (different) types.

The wording of function pointer arithmetic diagnostics still strikes me
as poorly phrased, and I worry this makes them slightly more awkward if
more consistent. I'm hoping to fix that with a follow-on patch and test
case that will also make them more helpful when a typedef or template
type parameter makes the type completely opaque.

Suggestions on better wording are very welcome, thanks to Richard Smith
for some initial help on that front.

llvm-svn: 133906

alloca that only holds a copy of a global and we're going to replace the users
of the alloca with that global, just nuke the lifetime intrinsics. Part of
PR10121.

llvm-svn: 133905

passes as well.

llvm-svn: 133904

This allows for more live scratch registers which is needed to handle
live ST registers before return and inline asm instructions.

llvm-svn: 133903

Both become <earlyclobber> defs on the INLINEASM MachineInstr, but we
now use two different asm operand kinds.

The new Kind_Clobber is treated identically to the old
Kind_RegDefEarlyClobber for now, but x87 floating point stack inline
assembly does care about the difference.

This will pop a register off the stack:

  asm("fstp %st" : : "t"(x) : "st");

While this will pop the input and push an output:

  asm("fst %st" : "=&t"(r) : "t"(x));

We need to know if ST0 was a clobber or an output operand, and we can't
depend on <dead> flags for that.

llvm-svn: 133902

The INLINEASM MachineInstrs have an immediate operand describing each
original inline asm operand. Decode the bits in MachineInstr::print() so
it is easier to read:

  INLINEASM <es:rorq $1,$0>, $0:[regdef], %vreg0<def>, %vreg1<def>, $1:[imm], 1, $2:[reguse] [tiedto:$0], %vreg2, %vreg3, $3:[regdef-ec], %EFLAGS<earlyclobber,imp-def>

llvm-svn: 133901

llvm-svn: 133900

remove the analysis group.

llvm-svn: 133899

Fix PR10187: when diagnosing a two-phase-lookup-related failure, don't assert that any names we find are valid candidates for the call.

llvm-svn: 133898

llvm-svn: 133897

llvm-svn: 133896

llvm-svn: 133895

llvm-svn: 133894

llvm-svn: 133893

llvm-svn: 133892

llvm-svn: 133891

llvm-svn: 133890

llvm-svn: 133889

llvm-svn: 133886

Move stl dir to now reside under lang/cpp.

llvm-svn: 133884

llvm-svn: 133883

llvm-svn: 133882

llvm-svn: 133881

llvm-svn: 133880

llvm-svn: 133879

llvm-svn: 133878

llvm-svn: 133877

llvm-svn: 133876

llvm-svn: 133875

llvm-svn: 133874

The .b8 operations in PTX are far more limiting than I first thought. The mov operation isn't even supported, so there's no way of converting a .pred value into a .b8 without going via .b16, which is
not sensible. An improved implementation needs to use the fact that loads and stores automatically extend and truncate to implement support for EXTLOAD and TRUNCSTORE in order to correctly support
boolean values.

llvm-svn: 133873

llvm-svn: 133872

llvm-svn: 133871

llvm-svn: 133870

llvm-svn: 133869

llvm-svn: 133868

pointing to the range [first character, last character] instead of just not
after the last character. Patch by Yan Ivnitskiy!

llvm-svn: 133867

llvm-svn: 133865