There are no unnecessary action calls period (courtesy of the annotation scheme) and too many 'this means'..

llvm-svn: 72263

Emit debug information for globals (which include automatic variables as well because on PIC16 they are emitted as globals by the frontend).

llvm-svn: 72262

Targets like PIC16 generate Static decls for automatic variables, emit the appropriate debug descriptor as well in that case.

llvm-svn: 72261

Handle correctly a very ugly part of the C++ syntax. We cannot disambiguate between a parenthesized type-id and
a paren expression without considering the context past the parentheses.

Behold:
  (T())x;  - type-id
  (T())*x; - type-id
  (T())/x; - expression
  (T());   - expression

llvm-svn: 72260

Factor the compound literal parsing out from ParseParenExpression and into a new ParseCompoundLiteralExpression.
No functionality change.

llvm-svn: 72259

Now it parses the cast expression unless 'stopIfCastExpr' is true.

llvm-svn: 72258

Embed its functionality into it's only user, ParseCXXCasts.
CXXCasts now get the "actual" expression directly, they no longer always receive a ParenExpr. This is better since the
parentheses are always part of the C++ casts syntax.

llvm-svn: 72257

'ParseExprAfterTypeofSizeofAlignof' method.

llvm-svn: 72256

-Makes typeof consistent with sizeof/alignof
-Fixes a bug when '>' is in a typeof expression, inside a template type param:
  A<typeof(x>1)> a;

llvm-svn: 72255

llvm-svn: 72254

llvm-svn: 72253

llvm-svn: 72252

This attempts to fix PR4239.

llvm-svn: 72251

use in expanding SCEVAddExprs with GEPs. The operands of a
SCEVMulExpr need to be multiplied together, not added.

llvm-svn: 72250

indirect call into direct call, thus the verifier would reject something it
previously accepted.

llvm-svn: 72249

This only rejects mismatches between target specific calling convention
and C/LLVM specific calling convention.
There are too many fastcc/C, coldcc/cc42 mismatches in the testsuite, these are
not reject by the verifier.

llvm-svn: 72248

PR4238.

llvm-svn: 72244

 - This commit has some messy stuff in it to extend string lifetimes, but that
   will go away once we switch to using the enum'd Triple interfaces.

llvm-svn: 72243

DragonFly OS type.

llvm-svn: 72242

llvm-svn: 72241

llvm-svn: 72240

llvm-svn: 72239

llvm-svn: 72238

llvm-svn: 72237

llvm-svn: 72236

llvm-svn: 72235

llvm-svn: 72234

can. Also, delay semantic analysis of initialization for
value-dependent as well as type-dependent expressions, since we can't
always properly type-check a value-dependent expression.

llvm-svn: 72233

operand is the last in a pattern.  There is no
reason this should be true (although apparently
it always is right now).

llvm-svn: 72232

llvm-svn: 72231

llvm-svn: 72229

llvm-svn: 72228

llvm-svn: 72227

evaluated first.  This can also improve codegen just a bit as we might
have another register to play with for the evaluation of the rhs.

llvm-svn: 72226

setting of NSObject attribute.

llvm-svn: 72225

llvm-svn: 72224

PR4238.

llvm-svn: 72223

llvm-svn: 72222

If this causes any new assertion failures that I didn't catch in
testing, the fix is usually to change "&v[0]" to "v.data()" for some
SmallVector v.

llvm-svn: 72221

llvm-svn: 72220