Re: PCH, and more generally C++ parser performance

[Mark Mitchell <mark at codesourcery dot com>]

>>>>> "Zack" == Zack Weinberg <zack@wolery.cumb.org> writes:

    Zack> I believe you.  I am extrapolating badly from the code I
    Zack> have in front of me, which is indeed quite ugly - but then
    Zack> it tries to dump out and reload the entire data segment
    Zack> (including libc state and other dragons).

Ugh.  I probably cannot say much more about the code I have in front
of me.  But suffice it to say that it does something much more like
dumping out the GC pool.  It takes some care (more care than presently
taken in GCC!) to make that work (in fact, that compiler uses
something like obstacks, only in a much, much cleaner way than the way
they were used in GCC), but with care you can get away with doing
vastly less than dumping your entire core state.

--
Mark Mitchell                   mark@codesourcery.com
CodeSourcery, LLC               http://www.codesourcery.com