On Wed, Jul 26, 2023 at 01:46:19AM +0200, Sandro wrote:
On 25-07-2023 18:23, Kevin Fenzi wrote:
> On Tue, Jul 25, 2023 at 02:23:00AM +0200, Sandro wrote:
> > On 24-07-2023 20:30, Samyak Jain wrote:
> > > 21426 builds have been tagged into f39, there are currently 1017 failed
builds
> > > that need to be addressed by the package maintainers. FTBFS bugs will be
> > > filed shortly.
> >
> > Will all the Python packages that failed during the Python3.12 mass rebuild
> > and haven't been fixed yet, receive another FTBFS/FTI bug? Or will those
be
> > filtered out?
>
> If they already have a FTBFS bug (attached to the tracker) they will not
> get a new bug. If they do not, they will. :)
Well, the first batch of bugs for the Python3.12 mass rebuild were mostly
FTI bugs, even though some of them turned out to be in fact FTBFS.
Since FTI and FTBFS use separate tracker bugs, I/we now have two different
bugs pointing at the same issue:
FTI:
https://bugzilla.redhat.com/show_bug.cgi?id=2220504
FTBFS:
https://bugzilla.redhat.com/show_bug.cgi?id=2226331
Yeah, the ftbfs script doesn't know about the fti bugs.
We could extend it, but they are different (but very related) things.
I also noticed that the FTBFS bugs filed after the f39 mass rebuild
have log
files attached. While not harmful, wouldn't it be better to just link to the
failed build in Koji, that has all the logs for all the archs stored anyway?
Especially since large log files are truncated.
After a while, the logs will get garbage collected and no longer will be
available in koji. Of course at this point you can do a scratch (or
real) build to get current logs, but the bugs have some logs so you can
see what the failure was at the time it happened. ;)
We don't attach full logs because some packages have logs that are...
crazy large.
kevin