This is the mail archive of the
mailing list for the GCC project.
Re: Problem with PFE approach [Was: Faster compilation speed]
- From: "Timothy J. Wood" <tjw at omnigroup dot com>
- To: Devang Patel <dpatel at apple dot com>
- Cc: Mike Stump <mrs at apple dot com>, gcc at gcc dot gnu dot org
- Date: Fri, 16 Aug 2002 14:45:45 -0700
- Subject: Re: Problem with PFE approach [Was: Faster compilation speed]
On Friday, August 16, 2002, at 01:54 PM, Devang Patel wrote:
I wanted to stop reading after "a) Toss some or all of your PFE code
in the bin (yikes!)"
Yeah, I'm sure that a ton of work has gone in to PFE. I certainly
don't mean to belittle PFE -- my current project builds HUGELY faster
with PFE enabled (you guys rule!). I simply hope to help drive you
guys on to even more stunning achievements :)
In any case, you want best of both worlds -- cpp-precomp (Apple's yet
another precomp implementation) and PFE.
Really what I want is to spend zero time thinking about these issues.
Sure, I'm a very experienced developer and I can think about these
issues, understand them and make it all work.
But, as a third party Apple developer, it is not my job to babysit
the IDE and compiler so that I get decent performance. The less time I
have to spend on this, the more time I can be writing Mac OS X
As such, what I ideally want is a single switch, defaulting to on,
which controls whether "precompiled headers" are used. When that
switch is on, my build should go fast. When it is off, my build will
be slower. The only reason for the switch to exist is the (hopefully
unlikely) case where header precompilation is busted and I need the
default preprocessor behavior.
Anyway, I've flogged this horse enough and I'm sure you all
understand my concerns :)