llvm-svn: 120131

llvm-svn: 120130

llvm-svn: 120129

names.

llvm-svn: 120091

llvm-svn: 120089

of the ASTs.  Only available in assertions builds.  No stability guarantee.

This is intended solely as a debugging tool.  I'm not sure if the goals
are sufficiently aligned with the XML printer to allow a common
implementation.

Currently just falls back on the StmtDumper to display statements,
which means it doesn't produce valid XML in those cases.

llvm-svn: 120088

llvm-svn: 120084

 - Default argument expressions pick up the value kind of the incoming
   expression, not the value kind of the parameter it initializes.
 - When building a template argument for substitution, A::x is an rvalue
   if x is an instance method.
 - Anonymous struct/union paths pick up value kind the same way that
   normal member accesses do;  extract out a common code path for this.

Enable the value-kind assertion, now that it passes self-host.

llvm-svn: 120055

FileSystemOpts through a ton of apis, simplifying a lot of code.
This also fixes a latent bug in ASTUnit where it would invoke
methods on FileManager without creating one in some code paths
in cindextext.

llvm-svn: 120010

- Negative ChainingSize doesn't make sense, make it unsigned.

llvm-svn: 119943

A new AST node is introduced:
   def IndirectField : DDecl<Value>;
IndirectFields are injected into the anonymous's parent scope and chain back to
the original field. Name lookup for anonymous entities now result in an
IndirectFieldDecl instead of a FieldDecl.
There is no functionality change, the code generated should be the same.

llvm-svn: 119919

and use a better and more general approach, where NullStmt has a flag to indicate whether it was preceded by an empty macro.

Thanks to Abramo Bagnara for the hint!

llvm-svn: 119887

  if (condition)
    CALL(0); // empty macro but don't warn for empty body.

Fixes rdar://8436021.

llvm-svn: 119838

llvm-svn: 119817

-Move the stuff of Diagnostic related to creating/querying diagnostic IDs into a new DiagnosticIDs class.
-DiagnosticIDs can be shared among multiple Diagnostics for multiple translation units.
-The rest of the state in Diagnostic object is considered related and tied to one translation unit.
-Have Diagnostic point to the SourceManager that is related with. Diagnostic can now accept just a
   SourceLocation instead of a FullSourceLoc.
-Reflect the changes to various interfaces.

llvm-svn: 119730

out because there are still bugs left.

llvm-svn: 119722

http://llvm.org/bugs/show_bug.cgi?id=8558).  This patch fixes it.  Thanks to
rjmccall for all the coaching!

Approved by rjmccall

llvm-svn: 119697

store it on the expression node.  Also store an "object kind",
which distinguishes ordinary "addressed" l-values (like
variable references and pointer dereferences) and bitfield,
@property, and vector-component l-values.

Currently we're not using these for much, but I aim to switch
pretty much everything calculating l-valueness over to them.
For now they shouldn't necessarily be trusted.

llvm-svn: 119685

no longer depends on Preprocessor, so we can move it out of Sema into
a nice new StringLiteral::getLocationOfByte method that can be used by
any AST client.

llvm-svn: 119481

  @synthesize foo = _foo;

keep track of the location of the ivar ("_foo"). Teach libclang to
visit the ivar as a member reference.

llvm-svn: 119447

about a __block cxx object.

llvm-svn: 119411

assignment to volatiles in C.  This in effect reverts some of mjs's
work in and around r72572.  Basically, the C++ standard is quite
clear, except that it lies about volatile behavior approximating
C's, whereas the C standard is almost actively misleading.

llvm-svn: 119344

in more situations.  In particular, for code like

   template<class T> void Fn() { T* x; delete x; }

getDestroyedType() will now return T rather than T*, as it would
before this change.  On the other hand, for code like this:

   template<class T> void Fn() { T x; delete x; }

getDestroyedType() will return an empty QualType(), since it doesn't
know what the actual destroyed type would be.  Previously, it would
return T.

OKed by rjmccall

llvm-svn: 119334

llvm-svn: 119331

llvm-svn: 119302

This is needed for Neon types when it is most natural to define them in terms
of a typedef.  For example, Neon poly8_t is a typedef for "signed char", and
we want to define polynomial vectors as vectors of that typedef.  Without this
change, the result will be a generic GCC-style vector.  I think this is safe
for other vector types as well, but I would appreciate a review of this.

llvm-svn: 119300

certain internal type-checking procedures as well as for representing
certain implicitly-generated operations.  Uses to follow.

llvm-svn: 119289

llvm-svn: 119181

llvm-svn: 119138

implicit conversions;  the last batch was specific to promotions.
I think this is the full set we need.  I do think dividing the cast
kinds into floating and integral is probably a good idea.

Annotate a *lot* more C casts with useful cast kinds.

llvm-svn: 119036

between complex types.

llvm-svn: 118994

operands appropriately.  There are a lot of missing complex-related
cast kinds.

llvm-svn: 118993

llvm-svn: 118966

llvm-svn: 118901

SourceManager::getPresumedLoc(), so that we don't try to make use of
an invalid presumed location. Doing so can cause crashes.

llvm-svn: 118885

NEON vector types need to be mangled in a special way to comply with ARM's ABI,
similar to some of the AltiVec-specific vector types.  This patch is mostly
just renaming a bunch of "AltiVecSpecific" things, since they will no longer
be specific to AltiVec.  Besides that, it just adds the new "NeonVector" enum.

llvm-svn: 118724

Fixes rdar://8649963.

llvm-svn: 118674

own subcategory, -Wconstant-conversion, which is on by default.

Tweak the constant folder to give better results in the invalid
case of a negative shift amount.

Implements rdar://problem/6792488

llvm-svn: 118636

// rdar://8632525

llvm-svn: 118634

mangler. Now member functions and pointers thereof have their calling
convention mangled as __thiscall if they have the default CC (even though,
they technically still have the __cdecl CC).

llvm-svn: 118598