to Eli for pointing out that these forms don't ignore the high bits of
their index operands, and as such are not immediately suitable for use
by isel.

llvm-svn: 62194

- Use canonical FileID when using getSpelling() caching.  This
  addresses some cache misses we were seeing with -fsyntax-only on
  Cocoa.h
- Added Preprocessor::getPhysicalCharacterAt() utility method for
  clients to grab the first character at a specified sourcelocation.
  This uses the PTH spelling cache.
- Modified Sema::ActOnNumericConstant() to use
  Preprocessor::getPhysicalCharacterAt() instead of
  SourceManager::getCharacterData() (to get PTH hits).

These changes cause -fsyntax-only to not page in any sources from
Cocoa.h.  We see a speedup of 27%.

llvm-svn: 62193

llvm-svn: 62192

Keep "has debug info" big in MachineModuleInfo to avoid circular dependency between AsmPrinter and CodeGen.

llvm-svn: 62191

llvm-svn: 62190

llvm-svn: 62189

llvm-svn: 62187

PTH: Fix remaining cases where the spelling cache in the PTH file was being missed when it shouldn't.  This shaves another 7% off PTH time for -Eonly on Cocoa.h

llvm-svn: 62186

llvm-svn: 62185

llvm-svn: 62184

llvm-svn: 62183

llvm-svn: 62182

for the Darwin tool chain.
 - Ideally we would localize these to tool specific argument
   processing but for now this matches gcc closely.

llvm-svn: 62181

llvm-svn: 62180

llvm-svn: 62179

llvm-svn: 62178

scheduling dependencies. Add assertion checks to help catch
this.

It appears the Mips target defaults to list-td, and it has a
regression test that uses a physreg dependence. Such code was
liable to be miscompiled, and now evokes an assertion failure.

llvm-svn: 62177

least their preferred alignment.

llvm-svn: 62176

Turn off some Destroy calls that are currenly causing double-destruction of ScopedDecls. We will re-enable this later, when we have time to fully solve the ownership issue.

llvm-svn: 62175

llvm-svn: 62174

been modified, to avoid trouble in the (unlikely) scenario that
D is a reference to an element in one of those arrays.

llvm-svn: 62173

 - Pulled -Xarch processing into this.

 - Get rid of manual creation of forwarding arg array.

 - Use Darwin/CC1 instead of generic GCC cc1 on X86.

llvm-svn: 62172

llvm-svn: 62170

via two paths, process it once not twice, d'oh!
Analysis, testcase and original patch thanks to
Mon Ping Wang.

llvm-svn: 62169

llvm-svn: 62168

was not being cleaned by ExpungeNode.

llvm-svn: 62167

llvm-svn: 62166

llvm-svn: 62165

DeclContext::KindTrait was not meant to be used outside of DeclContext::CastTo (causes compilation error on MSVC).
Add DeclContext::getDeclKind() and use that instead of DeclContext::KindTrait.

llvm-svn: 62164

PR3296 and PR3302.

llvm-svn: 62160

llvm-svn: 62158

llvm-svn: 62157

support -###.

llvm-svn: 62156

sees attributes it doesn't know.

llvm-svn: 62155

llvm-svn: 62154

llvm-svn: 62153

 - --gstabs only goes to Darwin/Assembler when dealing with an
     assembly file from the command line.

 - Relative placement of -o option for cc1 moves depending on
   -fsyntax-only/-S, how quaint.

llvm-svn: 62152

llvm-svn: 62151

llvm-svn: 62150

 - Simple mechanism for group together sets of options so the driver
   can efficiently deal with them as a group (i.e., for forwarding -i*
   to cc1).

 - Use to finish off the major missing pieces of Darwin/CC1 support.

llvm-svn: 62149