This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: gcc 2.98 release better approach?
- To: howarth at bromo dot med dot uc dot edu, toon at moene dot indiv dot nluug dot nl
- Subject: Re: gcc 2.98 release better approach?
- From: Mike Stump <mrs at windriver dot com>
- Date: Mon, 16 Oct 2000 12:44:47 -0700 (PDT)
- Cc: gcc at gcc dot gnu dot org
> Date: Mon, 16 Oct 2000 20:47:55 +0200
> From: Toon Moene <toon@moene.indiv.nluug.nl>
> We cannot stop Red Hat from issuing a half-way compiler
We should not take pot shots at other peoples compilers, least other
people take pot shots are our compiler. We don't gain by tarnishing
gcc's reputation in public like this.
Could we just agree to disagree and never mention it again?
If one seriously wants to do this, the appropriate way in my book is
to publish the testsuite results, without comment, using automated
analysis tools, if you want them analyzed. For example, cmp_logs does
a rough, by maybe acceptable job.
If gcc is always maintained at zero unexpected failures all the time,
then a snapshot is idential to a release, to the extent the testsuite
is complete. I view it is our job to make that _be_ the case. If it
isn't, we have failed. So, I read your comment as, we blew it and
provided Red Hat with a bad compiler. Fine, that may be true, but if
it is, lets learn from it, never do it again, and move on.