This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Update to Redhat gcc 2.96
On Tue, Jul 03, 2001 at 09:39:26AM -0700, Joe Buck wrote:
> > Hi,
> >
> > I saw this on Linux Weekly News:
> >
> > "[RHBA-2001:082-05] New version of GCC 2.96-RH available fixing several bugs"
> > http://lwn.net/2001/0628/a/rh-gcc2.96-rh.php3
> >
> > This note should be kept in mind when people file GNATS reports
> > against RH compilers.
>
> People should not be filing GNATS reports against RH compilers using the
> gcc.gnu.org GNATS database. Red Hat has the right to produce their own
People report GNATS bugs against Red Hat compilers all the time.
Have you looked in GNATS lately?
> fork of the compiler, and by doing so they helped to shake out a lot of
> bugs. But RH doesn't have the right to expect the net-wide developer
> community to track bugs for them, and I'm sure they have their own
> internal database in any case.
Red Hat does have a bug reporting system, it's at:
http://bugzilla.redhat.com/bugzilla/
but when the RH compiler ICE's, it still prints out the message to tell people
to look at http://gcc.gnu.org/bugs.html in order to report bugs against
the compiler. That's how RH bugs get recorded in GNATS.
> If someone does file a report against a RH compiler, it makes sense to see
> if the bug can be duplicated with 3.0, and if so it can be turned into a
> 3.0 bug. If not, it should go to "cannot duplicate" status.
I've been doing this. However, it is nice to be able to
give people useful information in order to solve their
problems instead of saying that the bug is fixed in the
latest gcc 3.0 CVS release. Upgrading to gcc 3.0 is not
appropriate for some real-world situations, ie. if you want
to compile the Linux kernel, while upgrading to a newer gcc 2.96
is more appropriate for these cases.
Whether you like it or not, RH bugs will get recorded in GNATS.
If it was up to me, I would immediately close all GNATS reports
filed against RH compilers, and send a response:
"Go to http://bugzilla.redhat.com/bugzilla/ to file a bug report
against the RH version of gcc."
Then, Jakub Jellinek, and the other Red Hat gcc folks can analyze
these reports and file a consolidated report against GNATS in order to
filter things out.
In addition, I would recommend that Red Hat change the
ICE error message in their compiler to tell people to file
bug reports at: http://bugzilla.redhat.com/bugzilla/.
--
Craig Rodrigues
http://www.gis.net/~craigr
rodrigc@mediaone.net