Added getSubRegIndex(A,B) that returns subreg index of A to B. Use it to replace broken code in VirtRegRewriter.

llvm-svn: 88753

definitions just because the type happens to be an array type.

llvm-svn: 88752

cast) that is converting to a class type, enumerate its constructors
as in any other direct initialization. This ensures that we get the
proper conversion sequence.

llvm-svn: 88751

 - Not tested, but -verify with multiple inputs should work now.

llvm-svn: 88750

llvm-svn: 88749

 - This reimplements -verify as just another DiagnosticClient, which buffers the diagnostics and checks them when the source file is complete. There are some hacks to make this work, but they are all internal, and this exposes a better external interface.

 - This also tweaks a few things:
   o Errors are now just regular diagnostics.
   o Frontend diagnostics are now caught (for example, errors in command line arguments), although there isn't yet a way to specify that they are expected. That would be nice though.

 - Not yet used.

llvm-svn: 88748

llvm-svn: 88747

- If destination is a physical register and it has a subreg index, use the
  sub-register instead.
This fixes PR5423.

llvm-svn: 88745

llvm-svn: 88744

llvm-svn: 88743

llvm-svn: 88742

in a category implementation (objc 32bit api related).

llvm-svn: 88741

llvm-svn: 88739

llvm-svn: 88738

llvm-svn: 88737

llvm-svn: 88735

llvm-svn: 88734

like a copy constructor to the overload set, just ignore it. This
ensures that we don't try to use such a constructor as a copy
constructor *without* triggering diagnostics at the point of
declaration.

Note that we *do* diagnose such copy constructors when explicitly
written by the user (e.g., as an explicit specialization).

llvm-svn: 88733

llvm-svn: 88732

llvm-svn: 88731

llvm-svn: 88729

When splitting an edge after a machine basic block with fall-through, we
forgot to insert a jump instruction. Fix this by calling updateTerminator() on
the fall-through block when relevant.

Also be more precise in PHIElimination::isLiveIn.

llvm-svn: 88728

llvm-svn: 88727

llvm-svn: 88726

PreAllocSplitting is now using this API to insert code.

llvm-svn: 88725

Revert r88718, which does NOT solve the constructor-template-as-copy-constructor issue. Big thanks to John for finding this

llvm-svn: 88724

llvm-svn: 88723

template argument.

llvm-svn: 88722

llvm-svn: 88719

constructor. Make sure that such declarations can never be formed.

llvm-svn: 88718

llvm-svn: 88717

llvm-svn: 88716

got ghost linkage. It's better than aborting.

llvm-svn: 88715

(still part of pr5472).

llvm-svn: 88712

so that isa<Instructon> doesn't return true for FixedStackPseudoSourceValue
values. This fixes a variety of problems, including crashes with -debug
and -print-machineinstrs. Also, add a comment to warn about this.

llvm-svn: 88711

llvm-svn: 88710

emitFunctionStubAtAddr.

llvm-svn: 88708

The BasicBlock associated with a MachineBasicBlock does not necessarily
correspond to the code in the MBB.

Don't insert a new IR BasicBlock when splitting critical edges. We are not
supposed to modify the IR during codegen, and we should be able to do just
fine with a NULL BB.

llvm-svn: 88707

llvm-svn: 88706

llvm-svn: 88705