-E -P mode by 10% (i.e. takes away almost all i/o time).

llvm-svn: 38669

llvm-svn: 38668

llvm-svn: 38667

llvm-svn: 38666

This allows the SourceManager to not compute line number information for each buffer.

llvm-svn: 38665

llvm-svn: 38663

up clang -E iostream by 30%.

llvm-svn: 38660

llvm-svn: 38659

llvm-svn: 38658

llvm-svn: 38644

llvm-svn: 38643

test/Preprocessor/_Pragma-dependency.c

llvm-svn: 38642

llvm-svn: 38631

llvm-svn: 38623

llvm-svn: 38612

llvm-svn: 38579

llvm-svn: 38571

llvm-svn: 38570

llvm-svn: 38569

llvm-svn: 38565

information to print the 3/4 flags correctly on #line directives emitted
in -E mode.

llvm-svn: 38562

still pretty hacky because it doesn't compute the 3/4 markers correctly.

llvm-svn: 38561

right number of newlines between tokens when needed.  This reduces the
delta of the gcc -E output from 12198 differences to 6764.  It still needs
to emit filenames on #line directives, track filename switches, and track
entry/exit of include files.

llvm-svn: 38559

llvm-svn: 38553

Now, instead of keeping a pointer to the start of the token in memory, we keep the
start of the token as a SourceLocation node.  This means that each LexerToken knows
the full include stack it was created with, and means that the LexerToken isn't
reliant on a "CurLexer" member to be around (lexer tokens would previously go out of
scope when their lexers were deallocated).

This simplifies several things, and forces good cleanup elsewhere.  Now the
Preprocessor is the one that knows how to dump tokens/macros and is the one that
knows how to get the spelling of a token (it has all the context).

llvm-svn: 38551

relying on TheLexer.

llvm-svn: 38549

Start/End pointer.

llvm-svn: 38548

llvm-svn: 38540

llvm-svn: 38539