for __unknown_anytype resolution to destructively modify the AST.  So that's
what it does now, which significantly simplifies some of the implementation.
Normal member calls work pretty cleanly now, and I added support for
propagating unknown-ness through &.

llvm-svn: 129331

If there's an invoke destination, we should use invoke instead of call when calling the __cxa_bad_typeid function. Fixes PR7400.

llvm-svn: 129273

llvm-svn: 129269

llvm-svn: 129265

CodeGenFunction::EmitDynamicCast always return null or throw a bad_cast 
exception.

llvm-svn: 129264

llvm-svn: 129262

llvm-svn: 129261

represents a dynamic cast where we know that the result is always null.

For example:

struct A {
  virtual ~A();
};
struct B final : A { };
struct C { };

bool f(B* b) {
  return dynamic_cast<C*>(b);
}

llvm-svn: 129256

Strip off parens and no-op casts when deciding if an expr can be devirtualized. Fixes the second half of PR9660.

llvm-svn: 129253

llvm-svn: 129252

llvm-svn: 129250

While I'm here, FileCheck-ize the ext-vector test, so we actually check
what it is generating.

llvm-svn: 129241

check is triggered appropriately.  Reported on cfe-dev.

llvm-svn: 129231

pageexec@freemail.hu, tweaks by me.

llvm-svn: 129206

llvm-svn: 129202

Patch by Dave Zarzycki!

llvm-svn: 129189

to eliminate a divide-by-8. No change in functionality intended.

llvm-svn: 129180

the base offset. No change in functionality intended.

llvm-svn: 129179

llvm-svn: 129176

Do not use zero as an upper bound for unbounded array because upper bound zero also indicates one element array. 

llvm-svn: 129157

Change the return type of CodeGenVTables::getVirtualBaseOffsetOffset() to
CharUnits. No change in functionality intended.

llvm-svn: 129072

The idea is that you can create a VarDecl with an unknown type, or a
FunctionDecl with an unknown return type, and it will still be valid to
access that object as long as you explicitly cast it at every use.  I'm
still going back and forth about how I want to test this effectively, but
I wanted to go ahead and provide a skeletal implementation for the LLDB
folks' benefit and because it also improves some diagnostic goodness for
placeholder expressions.

llvm-svn: 129065

with debug info.]

Use CharUnits for the offsets in the VirtualBaseClassOffsetOffsetsMapTy. No
change in functionality intended.

llvm-svn: 129048

info.]

Use CharUnits for the offset type in the ClassNamesAndOffsets map in
dumpLayout(). No change in functionality intended.

llvm-svn: 129046

Use CharUnits for the offsets in the VBaseOffsetOffsetsMapTy types. No
change in functionality intended.

llvm-svn: 129043

pass a previously failing clang test.
// rdar://8808439

llvm-svn: 129004

llvm-svn: 129000

llvm-svn: 128957

As a result, I had to remove a c++ version of a clang
test which requires more scrutiny on my part.

llvm-svn: 128950

llvm-svn: 128948

targets) when load/store results in multiple instructions.
// rdar://8808439

llvm-svn: 128937

llvm-svn: 128928

llvm-svn: 128921

llvm-svn: 128915

developers can see if their driver changed any cl::Option's. The
current implementation isn't perfect but handles most kinds of
options. This is nice to have when decomposing the stages of
compilation and moving between different drivers. It's also a good
sanity check when comparing results produced by different command line
invocations that are expected to produce the comparable results.

Note: This is not an attempt to prolong the life of cl::Option. On the
contrary, it's a placeholder for a feature that must exist when
cl::Option is replaced by a more appropriate framework. A new
framework needs: a central option registry, dynamic name lookup,
non-global containers of option values (e.g. per-module,
per-function), *and* the ability to print options values and their defaults at
any point during compilation.

llvm-svn: 128911

llvm-svn: 128908

llvm-svn: 128893

Does anyone want to buy me a new brain?

llvm-svn: 128890

llvm-svn: 128866

llvm-svn: 128857