scheduling.

llvm-svn: 41556

Finish converting decls over to the new invalid decl API.

Also removed an old FIXME (FIXME: "auto func();" passes through...).

llvm-svn: 41555

This involves the construction of a specialized "dispatch" block that
all basic blocks containing indirect gotos unconditionally transfer
control-flow to.  The successors of the dispatch block has as its successors
all of the blocks containing labels whose address was taken somewhere in
the function.

llvm-svn: 41554

llvm-svn: 41553

Move the "invalid decl" idiom up to Decl (where we have some bits to steal:-)

Converted ParmVarDecl, FileVarDecl, BlockVarDecl, and Sema::ParseIdentifierExpr() to use the idiom.

Updated array-constraint.c to make sure we no longer get "undeclared identifier" errors:-)

llvm-svn: 41552

t2.c:3:18: warning: expression result unused
     x = ++x, ++y, y+2;
     ~~~~~~~~~~~~^ ~~~

llvm-svn: 41551

llvm-svn: 41550

llvm-svn: 41549

directly in it.  Remove TargetInfo::getEnumPolicy, as there is only
one policy that we support right now.

llvm-svn: 41548

arbitrarily complex control-flow in their subexpressions.

llvm-svn: 41547

variables that have a pointer type, or arrays that contain pointers.

This fixes a crash on the following code:

int *h[3];

int **foo(int i)
{
   return &(h[i]);
}

This bug was reported by Keith Bauer (thanks!).

llvm-svn: 41546

llvm-svn: 41545

llvm-svn: 41544

llvm-svn: 41543

llvm-svn: 41542

llvm-svn: 41541

llvm-svn: 41540

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