llvm-svn: 84436

form that LLVM code generators can turn into efficient code.  For example,
int4 a, b, c;
a = (int4)(b.yzw, a.x)

llvm-svn: 84434

llvm-svn: 84423

llvm-svn: 84418

llvm-svn: 84329

anymore.  In C++ it is bool.

llvm-svn: 84308

FIXME in CGCXX.cpp that I would like Anders to
take a look at.

llvm-svn: 84265

llvm-svn: 84255

llvm-svn: 84237

TypeLoc records for declarations;  it should not be necessary to represent it
directly in the type system.

Please complain if you were using these classes and feel you can't replicate
previous functionality using the TypeLoc API.

llvm-svn: 84222

non-virtual part of the return result adjustments for covariant
thunks.  WIP.

llvm-svn: 84178

Track path information completely to ensure we get all the overrides.  WIP.

llvm-svn: 84166

struct A { };
struct B : A { };

void f() {
  const A& a = B();
}

correctly. (This now does the offset conversion if necessary and calls the destructor when a goes out of scope).

llvm-svn: 84162

llvm-svn: 84142

class templates the proper linkage. 

Daniel, please look over the CodeGenModule bits.

llvm-svn: 84140

the virtual base so they can be reused properly.  Don't reuse vcall
slots across a virtual boundary.  WIP.  I have a testcase, but there
are still things that need to be fixed before the testcase can go in.

llvm-svn: 84120

llvm-svn: 84114

Fix the clang build.

llvm-svn: 84107

from Var itself rather than the decl for DebugInfo metadata.

llvm-svn: 84102

1) -fwritable-string does affect the non-utf16 version of cfstrings
   just not the utf16 ones.
2) utf16 strings should always be marked constant, as the __TEXT segment
   is readonly.
3) The name of the global doesn't matter, remove it from TargetInfo.
4) Trust the asmprinter to drop cstrings into the right section, like llvmgcc does now.

This fixes rdar://7115750

llvm-svn: 84077

llvm-svn: 84075

llvm-svn: 84050

covariant thunks.  WIP.

llvm-svn: 84046

llvm-svn: 84039

llvm-svn: 84031

llvm-svn: 83986

llvm-svn: 83981

llvm-svn: 83979

Let replaceAllUsesWith() adjust VHs even though there are no uses.

llvm-svn: 83978

offsets for covariant thunks.

llvm-svn: 83965

llvm-svn: 83964

which is a common idiom to improve PIC'ness of code using the addr of
label extension.  This implementation is a gross hack, but the only other
alternative would be to teach evalutate about this horrid combination.
While GCC allows things like "&&foo - &&bar + 1", people don't use this
in practice.  This implements PR5131.

llvm-svn: 83957

1. CGF now has fewer bytes of state (one pointer instead of a vector).
2. The generated code is determinstic, instead of getting labels in
   'map order' based on pointer addresses.
3. Clang now emits one 'indirect goto switch' for each function, instead
   of one for each indirect goto.  This fixes an M*N = N^2 IR size issue
   when there are lots of address-taken labels and lots of indirect gotos.
4. This also makes the default cause do something useful, reducing the
   size of the jump table needed (by one).

llvm-svn: 83952

for a null pointer.  In other words, "&&foo != NULL" will always work out to
true.

llvm-svn: 83948

more places in clang codegen now.

llvm-svn: 83947

llvm-svn: 83928

llvm-svn: 83913

llvm-svn: 83912

llvm-svn: 83900

llvm-svn: 83886