llvm-svn: 40462

This speeds up -E on xalancbmk by 2.4%

llvm-svn: 40461

llvm-svn: 40460

bottleneck for -E computation, because every token that starts a line needs
to determine *which* line it is on (so -E mode can insert the appropriate 
vertical whitespace).  This optimization improves this common case where
it is striding through the line # table.

This speeds up -E on xalancbmk by 3.2%

llvm-svn: 40459

by Neil Booth.  This fixes Preprocessor/output_paste_avoid.c

llvm-svn: 40454

Patch by Neil Booth!

llvm-svn: 40452

llvm-svn: 40450

this fixes test/Sema/switch.c.

llvm-svn: 40438

llvm-svn: 40424

This speeds up -E on 447.dealII by 5.8%

llvm-svn: 40423

trivial to handle and very very common.  This speeds up -E on 
447.dealII by 2.5%

llvm-svn: 40422

and avoid computing the spelling of tokens when not needed.  This speeds
up -E on 447.dealII by 2.2%

llvm-svn: 40421

without having to get the whole token.  This speeds up -E on 
447.dealII by 1.8%

llvm-svn: 40420

llvm-svn: 40419

it did not handle <xyz> headers coming from macro expansions.
This requires special treatment, as the include name is lexed
as multiple tokens, which require reassembly before processing.

llvm-svn: 40418

no functionality changes.

llvm-svn: 40414

that required cleaning.  If the token required cleaning,
don't include the cleaned tokens in the returned length.

llvm-svn: 40410

token was not -> and if the token before it was -, the - and > would avoid pasting.

llvm-svn: 40409

C++ mode, even gnu C++ mode.

llvm-svn: 40408

a llvm::StringMap.  This dramatically reduces the startup time of the preprocessor,
speeding up -Eonly on xalankbmk by 2.2%.

llvm-svn: 40396

memorybuffer instead of a pointer to the memorybuffer itself.  This
reduces coupling and eliminates a pointer dereference on a hot path.
This speeds up -Eonly on 483.xalancbmk by 2.1%

llvm-svn: 40394

compiler inline it.  This speeds up -Eonly on 483.xalancbmk by about 1%

llvm-svn: 40393

preprocessing 483.xalancbmk by about 10%, reducing the number
of file lookup queries from 2139411 to 199466 (over 10x)

llvm-svn: 40390

llvm-svn: 40388

llvm-svn: 40386

llvm-svn: 40385

   // foo
   // bar
   // baz

we'd get two levels of call (bcpl & whitespace) for each line, leading to some
seriously deep stacks in some cases.

llvm-svn: 40384

fileid's and macroid's across files.  Clearing between files keeps the tables
smaller and slightly speeds up compilation.

llvm-svn: 40383

llvm-svn: 40381

redefinition of a macro could cause invalid memory to be deleted.
Found preprocessing 253.perlbmk.

llvm-svn: 40380

int i = /*/ */ 1;

Thanks to Neil for pointing this out.

llvm-svn: 40379

llvm-svn: 40375

keeps the MacroInfo table more compact.

llvm-svn: 40281

llvm-svn: 40264

llvm-svn: 40189

llvm-svn: 40176

llvm-svn: 40173

llvm-svn: 40169

llvm-svn: 40163

llvm-svn: 40162