convert this to filecheck, hopefully it will fix PR4888.  If nothing
else it will make tests run faster and make 4888 easier to diagnose.

llvm-svn: 81286

llvm-svn: 81284

Fix PR4922, where Sema would complete tentative definitions in nondeterminstic
order because it was doing so while iterating over a densemap.

There are still similar problems in other places, for example
WeakUndeclaredIdentifiers is still written to the PCH file in a nondeterminstic
order, and we emit warnings about #pragma weak in nondeterminstic order.

llvm-svn: 81280

Unbreak this test.

llvm-svn: 81271

When remat'ing and destination virtual register has a sub-register index. Make sure the sub-register class matches the register class of the remat'ed instruction definition register class.

llvm-svn: 81270

fix PR4767, a crash because fp stackifier visited blocks in
depth first order, so it wouldn't process unreachable blocks.
When compiling at -O0, late dead block elimination isn't done
and the bad instructions got to isel.

llvm-svn: 81268

llvm-svn: 81253

Packed unions should be packed. Fixes an assert Daniel reported.

llvm-svn: 81036

If the alignment of the chosen field in a union is greater than the alignment of the union, we need to use a packed LLVM struct. Fixes <rdar://problem/7184250>.

llvm-svn: 81035

Don't use '%ll', this doesn't workin on Win32.

llvm-svn: 81034

Improve llvm::getHostTriple for some cases where the LLVM_HOSTTRIPLE is not
reliable.

llvm-svn: 81033

Fixed a test that ensures the LocalRewriter does not attempt to
avoid reloads by reusing clobbered registers.

This was causing issues in 256.bzip2 when compiled with PIC for
a while (starting at r78217), though the problem has since been masked. 

llvm-svn: 81032

Fix the syntax of add/sub/mul nsw/nuw and sdiv exact.

llvm-svn: 81031

Opaque types didn't work if llvm_is_multithreaded().
AlwaysOpaqueTy is always NULL at this point, and it causes an assertion failure.
Fix it by using the just constructed tmp instead.

llvm-svn: 81030

fix PR4815: some cases where DeleteDeadInstruction can delete
the instruction BBI points to.

llvm-svn: 81029

fix PR4848 an infinite loop when indexing down through a recursive gep
and we get the original pointer type.  This doesn't mean that we're
at the first pointer being indexed.  Correct the predicate.

llvm-svn: 81028

fix PR4837, some bugs folding vector compares.  These
return a vector of i1, not i1 itself.

llvm-svn: 81027

Fix PR4845: r77946 completely broke x86_64 Darwin (or any situation where the
desired triplet is a sub-target, e.g. thumbv7 vs. arm host). Reverting the
patch isn't quite right either since the previous behavior does not allow the
triplet to be overridden with -march.

llvm-svn: 81023

For now, only run MC tests if X86 is configured.

llvm-svn: 80565

Force triple for these tests.

llvm-svn: 80564

Fix off-by-one in llvm::Format::print.
 - This also shortens the Format.h implementation, and uses the print buffer
  fully (it was wasting a character).

 - This manifested as llvm-test failures, because one side effect was that
  raw_ostream would write garbage '\x00' values into the output stream if it
  happened that the string was at the end of the buffer. This meant that grep
  would report 'Binary file matches', which meant the silly pattern matching
  llvm-test eventually does would fail. Cute. :)

llvm-svn: 80563

This is passing for PPC on Mac OS X.

llvm-svn: 80561

Make sure we specify no arguments for context functions.

llvm-svn: 80560

Fix the start source location for type-specs like long, short, etc.

llvm-svn: 80559

Fix TypeLoc::operator bool().

llvm-svn: 80558

convert to filecheck and force a triple so that this passes on big
endian hosts.  Fixes PR4792

llvm-svn: 80556

This should use isIndenticalToWhenDefined.

llvm-svn: 80554

Rename Instruction::isIdenticalTo to Instruction::isIdenticalToWhenDefined,
and introduce a new Instruction::isIdenticalTo which tests for full
identity, including the SubclassOptionalData flags. Also, fix the
Instruction::clone implementations to preserve the SubclassOptionalData
flags. Finally, teach several optimizations how to handle
SubclassOptionalData correctly, given these changes.

This fixes the counterintuitive behavior of isIdenticalTo not comparing
the full value, and clone not returning an identical clone, as well as
some subtle bugs that could be caused by these.

Thanks to Nick Lewycky for reporting this, and for an initial patch!

llvm-svn: 80553

When undoing a reuse in ReuseInfo::GetRegForReload, check if it was only a
sub-register being used. The MachineOperand::getSubReg() method is only valid
for virtual registers, so we have to recover the sub-register index manually.

llvm-svn: 80552

Special-case static allocas in IndVarSimplify's loop invariant
sinking code, since they are special. If the loop preheader happens
to be the entry block of a function, don't sink static allocas
out of it. This fixes PR4775.

llvm-svn: 80551

Fix PR4772 ARM JIT.GlobalInFuction unittest by explicitly initialize MMI 
to 0 during JITEmitter constructor.

llvm-svn: 80550

Don't install FileCheck man page

llvm-svn: 80549

Don't install FileCheck and FileUpdate

llvm-svn: 80548

Don't install Clang libraries

llvm-svn: 80547

Fix a few tests to be -Asserts agnostic.
- Ugh.
fix a GCC type punning warning.

llvm-svn: 80546

Temporarily switch clang back to not using clang-cc by default for C++, and only
using clang-cc on i386 and x86_64.

llvm-svn: 80545

Try to fix MSVC build after r79846.

llvm-svn: 80544

Fix non-determinism in DAGISel emitter.
 - This manifested as non-determinism in the .inc output in rare cases (when two
   distinct patterns ended up being equivalent, which is rather rare). That
   meant the pattern matching was non-deterministic, which could eventually mean
   the code generator selected different instructions based on the arch.

 - It's probably worth making the DAGISel ensure a total ordering (or force the
   user to), but the simple fix here is to totally order the Record* maps based
   on a unique ID.

 - PR4672, PR4711.

llvm-svn: 80543

llvm-mc: Hopefully fix Mach-O tests on big-endian platforms, force values to be
converted to Python ints if possible.

llvm-svn: 79810

Type of a ?: expression whose either expression is a built-in 'id'
type is 'id' type.

llvm-svn: 79788