This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Snapshots
- To: "Joseph S. Myers" <jsm28 at cam dot ac dot uk>
- Subject: Re: Snapshots
- From: Gerald Pfeifer <pfeifer at dbai dot tuwien dot ac dot at>
- Date: Sat, 10 Mar 2001 00:32:23 +0100 (CET)
- cc: Jeffrey A Law <law at redhat dot com>, <gcc at gcc dot gnu dot org>
On Fri, 9 Mar 2001, Joseph S. Myers wrote:
> Where does mail to gccadmin (generated by cron jobs, etc.) end up?
It's sent to gcc@gcc.gnu.org and then bounces, as gcc@gcc.gnu.org is
not in the To:/Cc: headers. I have informed overseers.
> Does gccadmin have write access to the ftp site?
Yes.
> Is it appropriate for cron jobs on gcc.gnu.org to check out and configure
> (but not build) a source tree? (For some purposes - regenerating files
> such as gcc.pot which should be regenerated every so often but don't need
> to be with every commit - it seems better to me to use the existing
> Makefile rules rather than duplicating them in scripts - which requires
> configuring a source tree. The same applies to the inclusion of info and
> Bison output files in release tarballs.)
I think there's no sensible way around that, is it?
> What's the appropriate (local to gcc.gnu.org) time of day for cron
> jobs to run at?
I'm not aware of any restrictions, probably sometime during the early
morning, if the jobs are more heavy. To be sure, I'd ask overseers.
> Should requests for software to be installed for such cron jobs (e.g. CVS
> gettext for regenerating gcc.pot, podlators for regenerating manpages) go
> to overseers, or should I install them under gccadmin's home directory?
The former, I'd think. Much of this kind of software should also be useful
for other projects.
> Should I just go ahead and test and set up cron jobs for this sort of
> thing, then send them to gcc-patches and commit them to a
> maintainer-scripts directory, or should scripts run from cron be
> approved in advance?
Peer-reviews for this kind of stuff are a good thing, usually, so I'd
post non-trivial patches, wait a bit, and then commit.
I think we want all such scripts under CVS control, and I believe Jeff
or Mark already suggested a directory name directly in the main directory
of the gcc CVS module.
Gerald
--
Gerald "Jerry" pfeifer@dbai.tuwien.ac.at http://www.dbai.tuwien.ac.at/~pfeifer/