This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: regression-checker: Add in new passes during regression.
- From: Hans-Peter Nilsson <hp at bitrange dot com>
- To: <gcc-patches at gcc dot gnu dot org>
- Date: Mon, 11 Mar 2002 19:54:54 -0500 (EST)
- Subject: Re: regression-checker: Add in new passes during regression.
On Mon, 11 Mar 2002, Geoff Keating wrote:
> > Date: Mon, 11 Mar 2002 06:23:09 -0500 (EST)
> > From: Hans-Peter Nilsson <hp@bitrange.com>
> > cc: <gcc-patches@gcc.gnu.org>
> >
> > I copied the regression-checking parts to my own regression
> > tester and discovered that during a regress-N state, new passes
> > are not added, so regressions in new tests would then not be
> > discovered. I think it's confusing to not see new regressions
> > during a state of regression, so how about this?
> > Not tested at all, except that "cat - file" works and that I use
> > the corresponding change myself and nothing's broke yet.
>
> This is not done because sometimes a really broken patch can cause
> tests to pass even though they shouldn't. For instance, if GCC did
> 'exit (0)' before actually compiling anything, a broken compile
> testcase would suddenly pass.
Whatever. Anyway, to whomever, don't use the patch as-is if
you like the idea; there's a bug in that patch; you'll never see
regressions.
brgds, H-P