This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Building egcs 2.95 on UnixWare 2.1.3
- To: egcs at egcs dot cygnus dot com, tangent at mail dot com
- Subject: Re: Building egcs 2.95 on UnixWare 2.1.3
- From: mrs at wrs dot com (Mike Stump)
- Date: Tue, 6 Jul 1999 10:55:28 -0700
> Date: Sun, 04 Jul 1999 05:03:36 -0600
> From: Warren Young <tangent@mail.com>
> To: egcs Mailing List <egcs@egcs.cygnus.com>
> But I have a more serious problem. EGCS 1.1.x was always slow about
> compiling a particular module of ours that was template-heavy. (About
> 40 seconds on a PPro-180, for a module less than 2000 LOC, including
> headers.) 2.95 fails on this file with a signal 11.
:-) I wonder if -fno-inline helps any (not for the sig 11, you already
said it won't, but for the speed problem). If so, then try
-finline-limit-@var{n}, if you compiler supports it, if not, try a
newer compiler.
> I think it'd be tough to distill this down to a simple example, because
> of the problem's very nature, but I can describe what's going on fairly
> concisely:
The English description is almost totally useless. The testcase is
trivial to generate. g++ -E .... >testcase.ii
Yes, it would be nice to have you cut it down, but that isn't strictly
required.