This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Regression testing...
- To: pfeifer at dbai dot tuwien dot ac dot at, scottb at netwinder dot org
- Subject: Re: Regression testing...
- From: Mike Stump <mrs at windriver dot com>
- Date: Tue, 18 Jan 2000 11:07:11 -0800 (PST)
- Cc: gcc at gcc dot gnu dot org
> Date: Tue, 18 Jan 2000 19:38:59 +0100 (MET)
> From: Gerald Pfeifer <pfeifer@dbai.tuwien.ac.at>
> We had quite a bit of documentation, but removed it before the GCC
> 2.95 release which did not contain the testsuite due to legal
> reasons and concerns by the FSF.
I would like to see advancement made in terms of the testsuite. In
particular, I would like to add back the infrastructure that was
appropriately donated to the FSF (testsuite/lib and old-deja.exp), and
then work on a list of known clean testcases.
I would like the see written down the guidelines given to the gcc team
by the FSF for what can be considered ours, so that I may use these
guidelines as the FSF intended to sort though my part of the testsuite
for C++ (I know it best), and add to the file that contains the list
clean testcases the files I think are clean.
I think others may be willing to step forward also, and do a similar
amount of work for the other parts. In time, I think all the nice and
easy and short cases can be added back.
Then, I would like to see the release and snapshot process include the
clean version of the testsuite. The unclean portions we can migrate
to unclean directories.
I haven't heard much in the way of planning concerning the testsuite,
and I don't know if this fits in with how things need to be done or
not, but I thought I might try and get the ball rolling with a
suggestion.
Thoughts from people in the know?