This is the mail archive of the
mailing list for the GCC project.
Re: 3.1 and precompiled headers
- From: Stan Shebs <shebs at apple dot com>
- To: Jack Howarth <howarth at bromo dot med dot uc dot edu>
- Cc: gcc at gcc dot gnu dot org
- Date: Tue, 21 May 2002 10:52:49 -0700
- Subject: Re: 3.1 and precompiled headers
- Organization: Apple Computer, Inc.
- References: <200205211706.NAA35621@bromo.msbb.uc.edu>
Jack Howarth wrote:
> Hmmm, doesn't Apple have precompiled headers in their gcc objc
> compiler (at least the new 3.1 based one)? Wouldn't the framework
> for precompiled headers be introduced into gcc if they merged in
> their changes?
Yes and yes. At the moment we're in the middle of our Jaguar
release death march :-), which among other things includes testing
the use of PFE (the name for our mechanism) on a large scale. But
our plan is to make PFE available for the community's evaluation
(the sources in Apple's repository only build on Darwin systems now,
but I have patches to make it work on GNU/Linux, not checked in yet
because of the aforementioned death march), and to monitor the PCH
Although any assessments have to be tentative because the PCH
branch isn't done yet, PFE works more at the memory allocation level
so that a single mmap can restore the compiler (not unlike Emacs'
unexec), while PCH looks more like a conventional tree save/restore.
The PCH branch has some good infrastructure that may make it easier
to maintain long-term.
In any case, I'm looking forward to being able to get some comparative
performance numbers; we're getting 6x speedup building some of the big
hairy C++-based subsystems in Mac OS X, but I'd like to be able to do
apples-to-apples on both Darwin and Linux once the PCH branch is