1. Eliminate the costly live interval "swapping".
2. Change ValueNumberInfo container from SmallVector to std::vector. The former
   performs slowly when the vector size is very large.

llvm-svn: 41536

llvm-svn: 41533

TODO: update the types of the constants and the enum.
llvm-svn: 41532

llvm-svn: 41531

t.c:1:12: warning: ISO C restricts enumerator values to range of 'int' (180388626432 is too large)
enum e {A, B = 42LL << 32, C = -4, D = 12456 };
           ^

llvm-svn: 41530

Changed the stack frame layout, StackGrowsUp fits better to Mips strange stack.
Stack offset calculation bug fixed!

llvm-svn: 41529

llvm-svn: 41528

Added comments about new stack allocation.
Expand SelectCC for i32 results

llvm-svn: 41527

Comments for Mips directives added.

llvm-svn: 41526

llvm-svn: 41525

Note the FIXME: we need some way to mark a decl erroneous :)

llvm-svn: 41524

implicit cast gets linked back into the AST.

llvm-svn: 41523

llvm-svn: 41522

Implement more thoughful error recovery when dealing with bogus declarator types.

For example, the following code was resulting in spurious warnings. This was the result of
Sema::GetTypeForDeclarator() synthesizing a type to hand back to the caller (in this case,
"int []", instead of "struct s[]", which is invalid).

struct s; 
struct s* t (struct s z[]) {   // expected-error {{array has incomplete element type}}
  return z;
}

Strategy: Flag the error in Declarator/DeclaratorChunk. This info is later stored in 
the ParmVarDecl. If the decl is referenced, Sema::ParseIdentifierExpr() will check if 
the type is invalid. If so, it quietly returns "true", without instantiating a DeclRefExpr.

This seems to work nicely. If Chris is happy with the approach, I will generalize this to
all VarDecls.

llvm-svn: 41521

llvm-svn: 41520

llvm-svn: 41517

llvm-svn: 41516

llvm-svn: 41515

llvm-svn: 41514

llvm-svn: 41513

based on an offset within SubExprs.

llvm-svn: 41512

would be invalid when RetValExp == NULL.

llvm-svn: 41511

llvm-svn: 41510

+ Added the creation of an empty Entry block at the end of CFG
  construction if the Entry block in the CFG contains multiple
  predecessors (which can happen with labels and do loops).

+ Fixed bug in the creation of an empty Exit block with functions where not
  all paths end in a return statement (but some do).  Basic blocks with
  return statements were jumping to a (sometimes) non-empty block.

+ FinishBlock no longer checks for labels at the beginning of a basic
  block before reversing the statements in the block.  This is because
  the recursion invariants of the builder methods have been cleaned up,
  and blocks are only passed to FinishBlock at most once.

+ Modified handling of "if", "for", "while", "do", and "switch" to allow
  condition expressions that can span multiple basic blocks.  This allows
  such conditions to contain short-circuit expressions (which span multiple
  blocks in the CFG).

llvm-svn: 41508

Update InvokeInst to work like CallInst

llvm-svn: 41506

llvm-svn: 41503

llvm-svn: 41502

llvm-svn: 41501

llvm-svn: 41500

llvm-svn: 41499

llvm-svn: 41498

latest EH changes, and in any case it is hard to
imagine how the original bug could be reintroduced.

llvm-svn: 41497

llvm-svn: 41496

llvm-svn: 41495

resulting in zero matches, while llvm-gcc-4.1 manages to
remove one pointless selector changing the number of matches.

llvm-svn: 41494

implicit casts from T to T& at the topmost part of the return-value expression.

This checking may be needed within EvalAddr later on.  We'll wait until
test cases show this kind of logic is necessary (as more C++ features are
implemented in clang).

llvm-svn: 41493

llvm-svn: 41492

where both pointers have non-zero offsets.

llvm-svn: 41491

offset needs to be adjusted in this test.

llvm-svn: 41490

to not alias each other, it can be translated as an llvm.memcpy.

llvm-svn: 41489