Bug#1055269: udd: bugs.cgi does not show bugs for source packages in non-free-firmware
Hi,
On 03/11/23 at 13:52 +0100, Cyril Brulebois wrote:
> Andreas Beckmann <anbe@debian.org> (2023-11-03):
> > The list of RC bugs in sid
> > https://udd.debian.org/bugs.cgi?release=sid&merged=ign&fnewerval=7&flastmodval=7&rc=1&sortby=id&sorto=desc&format=html#results
> > does not contain e.g. #1055136 filed against src:nvidia-graphics-drivers
> > which is in non-free-firmware, but it lists the clones of this bug that
> > are assigned to various old driver series
> > src:nvidia-graphics-drivers-{tesla,legacy}-* that are still in non-free
> > (not in non-free-firmware).
> >
> > Interestingly the RC bug list for bullseye
> > https://udd.debian.org/bugs.cgi?release=bullseye&merged=ign&fnewerval=7&flastmodval=7&rc=1&sortby=id&sorto=desc&format=html#results
> > does list the bug. src:nvidia-graphics-drivers/bullseye is in non-free.
>
> Indeed, udd doesn't seem to know about non-free-firmware at all, e.g.:
>
> CONSTRAINT check_component CHECK ((component = ANY (ARRAY['main'::text, 'contrib'::text, 'non-free'::text])))
>
> Things like udd/ftpnew_gatherer.py could almost work by accident since
> that's using .startswith() (but then assigning "non-free" as value, so
> that probably doesn't work anyway).
>
> I'm afraid I'm not learning udd's codebase and configuration today.
UDD uses several independant "importers". The constraint you quoted is
in the "blends" importer (maintained by Andreas Tille, cced).
To debug this, you can use the SQL query at the bottom of the page:
select id, bugs.package, bugs.source, severity, title, last_modified,
status, affects_stable, affects_testing, affects_unstable,
affects_experimental
from bugs
where id in (select id from bugs_rt_affects_unstable)
and not (id in (select id from bugs_merged_with where id > merged_with))
AND (severity >= 'serious')
order by id desc
#1055136 is not listed because it's not in "select id from
bugs_rt_affects_unstable".
That view uses the bugs_rt_affects_dist function, that basically does
"check if the BTS thinks the bug affects a distribuation, and then
override if release team tags (e.g. xx-ignore) are used".
The reason why UDD thinks that #1055136 does not affect unstable, is
because the BTS thinks it does not affect unstable. If you look at the
version graph for the bug, you see that the BTS only knows about the
version in oldstable, not about the versions in stable/testing/unstable.
The same happens for other packages in non-free-firmware (see #1038610
for example).
Lucas
Reply to: