SDIsel processes llvm.dbg.declare by recording the variable debug information descriptor and its corresponding stack frame index in MachineModuleInfo. This only works if the local variable is "homed" in the stack frame. It does not work for byval parameter, etc.
Added ISD::DECLARE node type to represent llvm.dbg.declare intrinsic. Now the intrinsic calls are lowered into a SDNode and lives on through out the codegen passes.
For now, since all the debugging information recording is done at isel time, when a ISD::DECLARE node is selected, it has the side effect of also recording the variable. This is a short term solution that should be fixed in time.

llvm-svn: 46659

llvm-svn: 46658

llvm-svn: 46657

than double through the C bindings. Thanks to Tomas Lindquist
Olsen for reporting it.

llvm-svn: 46656

llvm-svn: 46655

Cleanup previous patch (based on feedback from Ted).

Since this behavior is useful for most classes, we might consider adding a simple 3 method class that implements the behavior. Ted said that Boost has such a class.

llvm-svn: 46654

Make sure SourceManager/HeaderSearch don't support default copy constructors (since they result in bad runtime behavior).

I'm sure there are other classes that might need this "guard", however I was bitten by these 2 recently (so I thought I'd fix them).

llvm-svn: 46653

llvm-svn: 46652

llvm-svn: 46651

Rename diagnostic to reflect it's role...

llvm-svn: 46650

Minor API cleanup...

llvm-svn: 46649

Remove temporary workaround (thanks to Ted for fixing ObjCAtCatchStmt so quickly:-).

llvm-svn: 46648

and put the the next ObjcAtCatchStmt* as part of SubExprs.  This fixes a bug with
iterating over the children of ObjcAtCatch, where the next @catch was not
properly being iterated over as a child.

Altered serialization of ObjCAtCatchStmt to reflect this new layout of
its subexpressions, and fixed an ownership issue with the next @catch not
being serialized as an owned pointer.

llvm-svn: 46647

llvm-svn: 46646

llvm-svn: 46645

Fix two rewriter bugs with @catch.

- Support @catch(...), rather than crash:-)
- Make sure all catch bodies get rewritten. This "fix" is really a workaround until the iterator for the "try" AST is fixed. Will fix this in a separate commit.

llvm-svn: 46644

llvm-svn: 46640

llvm-svn: 46639

llvm-svn: 46638

llvm-svn: 46636

Remove the nasty LABEL hack with a much less evil one. Now llvm.dbg.func.start implies a stoppoint is set. SelectionDAGISel records a new source line but does not create a ISD::LABEL node for this special stoppoint. Asm printer will magically print this label. This ensures nothing is emitted before.

llvm-svn: 46635

It is allowed to get the address of an array subscript, even if the array has the register qualifier, if the array is really a pointer.

llvm-svn: 46634

Fix Sema::ActOnInstanceMessage to correctly do things in terms of canonical 
types, fixing bogus errors like:

NSDistantObject.m:10383:120: error: bad receiver type 'typeof((id<NSMutableCopying>)self)'
    id mess = ({ id __inv__ = ((void *)0); id __mb__ = _NSMessageBuilder((id <NSMutableCopying>)self, &__inv__); (void)[(__typeof__((id <NSMutableCopying>)self))__mb__ mutableCopyWithZone:((void *)0)]; if (!objc_collecting_enabled()) object_dispose(__mb__); __inv__; });

llvm-svn: 46633

here I fix just one.  The loop that rips through pointers should use
getAsPointerType() not static_cast<PointerType*> to get the pointee.  This
fixes a crash on a large testcase.

llvm-svn: 46632

llvm-svn: 46631

Implemented '==' and '!=' for ConcreteIntLValue.

llvm-svn: 46630

llvm-svn: 46629

llvm-svn: 46628

llvm-svn: 46627

llvm-svn: 46626

llvm-svn: 46625

llvm-svn: 46624

llvm-svn: 46623

rewriter more robust.

llvm-svn: 46622

fixes crashes where the insertion point was in a macro.  Instead of crashing, we want
to emit a warning.

llvm-svn: 46621

llvm-svn: 46620

Add a RewriteTest::ReplaceStmt method to factor the 'checking for
rewrite failed + emitting diagnostic if so' code.

llvm-svn: 46619

Moved ValueKey/ValueMap declaration to ValueState.h.

llvm-svn: 46618

exposed a bug in APFloat's long double->double conversion of
NaNs.  Broke several things in the ieee part of gcc testsuite.

llvm-svn: 46617

Fix http://llvm.org/bugs/show_bug.cgi?id=1967.

llvm-svn: 46616