llvm-svn: 122394

llvm-svn: 122393

loads properly.  We miscompiled the testcase into:

_test:                                  ## @test
	movl	$128, (%rdi)
	movzbl	1(%rdi), %eax
	ret

Now we get a proper:

_test:                                  ## @test
	movl	$128, (%rdi)
	movsbl	(%rdi), %eax
	movzbl	%ah, %eax
	ret

This fixes PR8757.

llvm-svn: 122392

unhanded cases faster and simplify code.

llvm-svn: 122391

llvm-svn: 122390

llvm-svn: 122389

This change is necessary when the variable is a const reference and we need
the l-value of the construct expr.  After that, when binding the variable,
recover the lazy compound value when the variable is not a reference.

In Environment, use the value of a no-op cast expression when it has one.
Otherwise, blast-through it.

llvm-svn: 122388

Previously designated anonymous fields were found via name lookup. This redesign uses the fact that an IndirectFieldDecl declaration will always follow an anonymous implicit field to remove the special case of name lookup. 

llvm-svn: 122387

0 mid-stack, stop backtracing.

SectionLoadList.cpp (ResolveLoadAddress): Don't assert on an
out-of-range address, just return an invalid Address object.
The unwinder may be passing in invalid addresses on the final
stack frame and the assert is a problem.

llvm-svn: 122386

llvm-svn: 122385

llvm-svn: 122384

llvm-svn: 122382

llvm-svn: 122381

and waiting for two expected state changed events to arrive: "running" followed by
"stopped".

llvm-svn: 122380

llvm-svn: 122379

I still think that LVI should be handling this, but that capability is some ways off in the future,
and this matters for some significant benchmarks.

llvm-svn: 122378

types, but they're just getting converted to unsigned anyway, so cast first
(and ask questions later).

llvm-svn: 122377

llvm-svn: 122371

the same physical register. Simplifies the fix from the previous
checkin r122211.

llvm-svn: 122370

to cope with parameter packs. This is a band-aid I will be
revisiting this section when I implement declaration matching
semantics for variadic templates.

llvm-svn: 122369

llvm-svn: 122368

llvm-svn: 122367

the shift type was needed one place, the shift count
type another.  The transform in 123555 had the same
problem.

llvm-svn: 122366

llvm-svn: 122365

(add Y, (sete  X, 0)) -> cmp X, 1; adc  0, Y
(add Y, (setne X, 0)) -> cmp X, 1; sbb -1, Y
(sub (sete  X, 0), Y) -> cmp X, 1; sbb  0, Y
(sub (setne X, 0), Y) -> cmp X, 1; adc -1, Y

for
  unsigned foo(unsigned a, unsigned b) {
    if (a == 0) b++;
    return b;
  }
we now get:
  foo:
    cmpl  $1, %edi
    movl  %esi, %eax
    adcl  $0, %eax
    ret
instead of:
  foo:
    testl %edi, %edi
    sete  %al
    movzbl  %al, %eax
    addl  %esi, %eax
    ret

llvm-svn: 122364

conversions, make sure that the (possibly) derived type is complete
before looking for base classes.

Finishes the fix for PR8801.

llvm-svn: 122363

llvm-svn: 122362

declaration, also look for an instantiation of its previous
declarations. Fixes PR8801.

llvm-svn: 122361

llvm-svn: 122360

the folding it can.

llvm-svn: 122359

are not actually relaxed. For example, a section with only alignments will never
needs relaxation.

llvm-svn: 122356

llvm-svn: 122355

count operand.  These should be the same but apparently are
not always, and this is cleaner anyway.  This improves the
code in an existing test.

llvm-svn: 122354

llvm-svn: 122353

llvm-svn: 122352

inconsistent with the type that the builtin *should* have, forget
about the builtin altogether: we don't want subsequence analyses,
CodeGen, etc., to think that we have a proper builtin function.

C is protected from errors here because it allows one to use a
library builtin without having a declaration, and detects inconsistent
(re-)declarations of builtins during declaration merging. C++ was
unprotected, and therefore would crash.

Fixes PR8839.

llvm-svn: 122351

take into account the region of interest. Otherwise, we may fail to
traverse some important preprocessed entity cursors. 
Fixes <rdar://problem/8554072>.

llvm-svn: 122350

llvm-svn: 122349

llvm-svn: 122348

classes, categories, protocols, and class extensions, where the
methods and properties of these entities would be inserted into the
DeclContext in an ordering that doesn't necessarily reflect source
order. The culprits were Sema::ActOnMethodDeclaration(), which did not
perform the insertion of the just-created method declaration into
the DeclContext for these Objective-C entities, and
Sema::ActOnAtEnd(), which inserted all method declarations at the
*end* of the DeclContext. 

With this fix in hand, clean up the code-completion actions for
property setters/getters that worked around this brokenness in the AST.

Fixes <rdar://problem/8062781>, where this problem manifested as poor
token-annotation information, but this would have struck again in many
other places.

llvm-svn: 122347