[incremental] Patch: FYI: lex all at once

[Tom Tromey]

>>>>> "Chris" == Chris Lattner <clattner@apple.com> writes:

>> This changes the C front end to do all lexing at once, rather than
>> incrementally.  This is infrastructure needed for token checksumming.

Chris> Interesting.  Out of curiosity, does this speed up or slow down the C
Chris> compiler at all?

I haven't tried comparing it in isolation yet.  I made a note to do
that at some point in the future -- but it could be a while.  Do we
have some standard benchmark we use for this?

ISTR that the equivalent change for C++ yielded a speedup, but I
wouldn't assume the same will apply to C.

Tom