Make tail duplication of indirect branches much more aggressive (for targets
that indicate that it is profitable), based on further experience with
this transformation.  I compiled 3 large applications with and without
this more aggressive tail duplication and measured minimal changes in code
size.  ("size" on Darwin seems to round the text size up to the nearest
page boundary, so I can only say that any code size increase was less than
one 4k page.) Radar 7421267.

llvm-svn: 89814

This violates the ABI (that area is "reserved"), and
while it is safe if all code is generated with current
compilers, there is some very old code around that uses
that slot for something else, and breaks if it is stored
into.  Adjust testcases looking for current behavior.
I've verified that the stack frame size is right in all
testcases, whether it changed or not.  7311323.

llvm-svn: 89811

Convert test case to FileCheck to test the behavior of the nil-receiver checker when the code is targetted for either Tiger or Leopard.

llvm-svn: 89810

For the nil-receiver checker, take into account the behavioral changes that got introduced in Mac OS X 10.5 and later, notably return values of double, float, etc., will not be garbage.  Fixes <rdar://problem/6829160>.

llvm-svn: 89809

initial transition of the nil-receiver checker to the Checker
interface as done in r89745.  Some important changes include:

1) We consolidate the BugType object used for nil receiver bug
reports, and don't include the type of the returned value in the
BugType (which would be wrong if a nil receiver bug was reported more
than once)

2) Added a new (temporary) flag to CheckerContext: DoneEvauating.
This is used by GRExprEngine when evaluating message expressions to
not continue evaluating the message expression if this flag is set.
This flag is currently set by the nil receiver checker.  This is an
intermediate solution to allow the nil-receiver checker to properly
work as a plug-in outside of GRExprEngine.  Basically, this flag
indicates that the entire message expression has been evaluated, not
just a precondition (which is what the nil-receiver checker does).
This flag *should not* be repurposed for general use, but just to pull
more things out of GRExprEngine that already in there as we devise a
better interface in the Checker class.

3) Cleaned up the logic in the nil-receiver checker, making the
control-flow a lot easier to read.

llvm-svn: 89804

llvm-svn: 89803

the common routine.

llvm-svn: 89802

llvm-svn: 89801

cleanups for while loops: 

1) Make sure that we destroy the condition variable of a while statement each time through the loop for, e.g.,

   while (shared_ptr<WorkInt> p = getWorkItem()) {
         // ...
         }

2) Make sure that we always enter a new cleanup scope for the body of the while loop, even when there is no compound expression, e.g.,

   while (blah)
     RAIIObject raii(blah+1);

llvm-svn: 89800

With this change, the clang-on-clang test result is now

Expected Passes    : 224
Unexpected Failures: 37

Which means that we can compile over 80% of clang with clang! :)

llvm-svn: 89799

member-reference contexts.  Fixes some clang-on-clang asserts.

llvm-svn: 89796

llvm-svn: 89793

llvm-svn: 89791

llvm-svn: 89790

llvm-svn: 89789

llvm-svn: 89787

llvm-svn: 89786

DependentScopeDeclRefExpr support storing templateids.  Unite the common   
code paths between ActOnDeclarationNameExpr and ActOnTemplateIdExpr.

This gets us to a point where we don't need to store function templates in
the AST using TemplateNames, which is critical to ripping out OverloadedFunction.

Also resolves a few FIXMEs.

llvm-svn: 89785

constructor does anyway.

llvm-svn: 89784

Handle cases where we're constructing an array of objects and the constructor has default arguments.

llvm-svn: 89783

llvm-svn: 89782

Add support for variadic collection functions. More to do
here.

llvm-svn: 89781

llvm-svn: 89778

llvm-svn: 89777

make sure that this variable is destroyed when we exit the switch
statement.

llvm-svn: 89776

We always need to emit the base expression of a member expression, even when the member decl refers to an enum. Thanks to Eli for pointing this out!

llvm-svn: 89775

llvm-svn: 89774

  - Outside the "if", to ensure that we destroy the condition variable
    at the end of the "if" statement rather than at the end of the
    block containing the "if" statement.
  - Inside the "then" and "else" branches, so that we emit then- or
    else-local cleanups at the end of the corresponding block when the
    block is not a compound statement.

To make adding these new cleanup scopes easier (and since
switch/do/while will all need the same treatment), added the
CleanupScope RAII object to introduce a new cleanup scope and make
sure it gets cleaned up.

llvm-svn: 89773

llvm-svn: 89772

llvm-svn: 89771

llvm-svn: 89770

llvm-svn: 89769

llvm-svn: 89768

llvm-svn: 89767

llvm-svn: 89766

llvm-svn: 89765

llvm-svn: 89764

llvm-svn: 89760

llvm-svn: 89758

llvm-svn: 89751