llvm-svn: 159481

a reduced testcase, but this fixes pr13209.

llvm-svn: 159479

llvm-svn: 159478

with argument type of size 0. // rdar://11777609
PR13229.

llvm-svn: 159477

llvm-svn: 159476

llvm-svn: 159475

llvm-svn: 159474

to improve wording of a diagnostic message.

llvm-svn: 159473

when an argument type size is 0. // rdar://11777609, PR13229

llvm-svn: 159472

llvm-svn: 159471

Use getUniqueVRegDef.
Replace a loop with existing interfaces: modifiesRegister and readsRegister.
Factor out code into inline functions and simplify the code.

llvm-svn: 159470

In future changes we should:
* use __builtin_trap rather than derefing 'random' volatile pointers.
* avoid dumping temporary files into /tmp when running tests, instead
  preferring a location that is properly cleaned up by lit.

Review by Chandler Carruth.

llvm-svn: 159469

Added documentation for many of our python properties and also made the property help show up by declaring the properties correctly. We previosly declared properties into a local "x" variable, what I didn't realize is that the help will use this as the property name for the help output.

llvm-svn: 159468

llvm-svn: 159467

llvm-svn: 159466

instructions with two register operands.

llvm-svn: 159465

keywords when doing type correction.

llvm-svn: 159464

is selected.  This will allow more flexibility when converting diagnostics to
use template type diffing.

Also updated the internal manual and test cases for correctly keeping the bold
attribute and for tree printing.

llvm-svn: 159463

Jordan Rose corrected me that this actually isn't needed.

llvm-svn: 159462

When a local virtual register is made global, make sure to clear any
existing kill flags.

llvm-svn: 159461

This would previously get reported as the misleading "Virtual register
def doesn't dominate all uses."

llvm-svn: 159460

The solution is a bit inefficient: it creates N checkers, one for each check, and
each check does a dispatch on the function name.  This is redundant, but we can fix
this once we have the proper ability to enable/disable subchecks.

Fixes <rdar://problem/11780180>.

llvm-svn: 159459

initialized with the next to last element to silence the warning.

llvm-svn: 159458

This happens when codegenprepare is invoked via opt.

llvm-svn: 159457

correctly for blocks and function pointer arguments
in the written constructor.  // rdar://11359268

llvm-svn: 159456

llvm-svn: 159455

platforms declare memset in rewritten code instead.

llvm-svn: 159454

This comes in handy during peephole optimization.

llvm-svn: 159453

but apparently I did not.

llvm-svn: 159452

really happening.  No intended functionality change.

llvm-svn: 159451

to produce quotes instead of adding qoute to the test.

llvm-svn: 159450

llvm-svn: 159449

llvm-svn: 159448

llvm-svn: 159447

from Filipe Cabecinhas!

llvm-svn: 159446

entry point which requires nil check before calling objc_msgSend_stret.
// rdar://11359268 - wip.

llvm-svn: 159445

Factor out a check for block commands (that implicitly start a new paragraph) into a separate function.

llvm-svn: 159444

llvm-svn: 159443

Reviewed by Doug Gregor.

llvm-svn: 159442

RefreshCallGraph: ignore 'invoke intrinsic'. IntrinsicInst doesnt not recognize invoke, and shouldnt at this point, since the rest of LLVM codebase doesnt expect invoke of intrinsics

llvm-svn: 159441