This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [3.0 critical] Make fixproto deal with assert.h
- To: zackw at stanford dot edu
- Subject: Re: [3.0 critical] Make fixproto deal with assert.h
- From: Mark Mitchell <mark at codesourcery dot com>
- Date: Thu, 10 May 2001 15:12:36 -0700
- Cc: gcc-patches at gcc dot gnu dot org
- Organization: CodeSourcery, LLC
- References: <20010510135819.A5658@saga9.stanford.edu>
>>>>> "Zack" == Zack Weinberg <zackw@stanford.edu> writes:
Zack> <assert.h> is used in several of the non-C front ends and
Zack> language runtimes. I could certainly go through and change
Zack> them all to if/abort pairs. It'd be a lot more invasive,
Zack> but it would allow us to get rid of __eprintf, so it might
Zack> be worth it.
Nah, too much work. If we really had to, we could provide an
`assert.h' that just calls `abort' (rather than __eprintf) and use it
only when building GCC, but not install it.
Zack> Or I could get rid of assert.h and not bother changing its
Zack> users. Autoconf uses assert.h unconditionally in
Zack> AC_CHECK_FUNC and no one's complained -- suggests that it's
Zack> near-universal on systems people actually care about.
Zack> What would you prefer?
I like that solution, and I think its usage in AC_CHECK_FUNC is very
good evidence for its universality.
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com