Re: Half the world being removed
Paul Gevers writes ("Re: Half the world being removed"):
> On 02-09-2022 07:27, Andrey Rahmatullin wrote:
> > On Thu, Sep 01, 2022 at 11:04:38PM -0500, Steven Robbins wrote:
> >> Suddenly half the packages are marked AUTOREMOVE; many due to gcc-12 and
> >> zlib.
> >> The related two bugs are months-old.
> >>
> >> Why are things suddenly being removed??
> > Both are key packages per
> > https://udd.debian.org/cgi-bin/key_packages.yaml.cgi so it must be some
> > recent problem that causes the things to ignore that.
>
> I'll look into it tonight (UTC+2). I did make one change to udd
> yesterday. I thought it was safe because I just reverted an exception to
> allow scikit-learns removal.
>
> There are probably bugs involved.
Thanks.
I don't want to add to the stressball but: this isn't the first time
that we've had malfunctions which want to remove very large numbers of
packages. Obviously this is a complicated algorithm and I understand
that there will be bugs.
I wonder if it would be possible to detect a sudden large increase in
the number of autoremovals, and stop the autoremoval system instead of
causing blaring klaxons for everyone in the project ?
That might make the failures less disruptive. (And would avoid having
everyone pile-on.)
Ian.
--
Ian JacksonThese opinions are my own.
Pronouns: they/he. If I emailed you from @fyvzl.net or @evade.org.uk,
that is a private address which bypasses my fierce spamfilter.
Re: Comments on proposing NEW queue improvement (Re: Current NEW review process saps developer motivation
Sean Whitton writes ("Re: Comments on proposing NEW queue improvement (Re:
Current NEW review process saps developer motivation"):
> On Sat 27 Aug 2022 at 04:22PM +02, Vincent Bernat wrote:
> > It does not seem to work. Either people don't want to do that, either the
> > FTP
> > team is too picky on the candidates.
>
> Some combination of both, but I don't think I'm suffering from bias if I
> say that it's at least 80% the former. Very few people who say they'd
> like to be trained confirm they'd still like to once they've had a look
> at the docs for trainees, and after that, hardly any do enough trainee
> reviews for the other team members to feel confident they can let them
> at it on their own.
I am in this picture. Some years ago now I volunteered. I was
introduced to the internal ftpmaster documentation and processes. At
the time, these documents were not even published - including,
astonishingly, some elements which read like a manifesto. (I don't
know if these documents are published nowadays.)
What I saw was very far away from what I had expected or hoped and
expected to see - especially in terms of process and culture. In
particular it was far away from Debian's usual norms of transparency,
but there were many other problems.
I deecided I couldn't work with it. I could, of course, have tried to
be the change I wanted to see in the world. But, empirically, I'm not
the best person to be trying to lead organisational change in Debian.
As an institution, ftpmaster has been very successful in establishing
and maintaining its own norms and culture. (It does help of course
that it is a tremendously powerful team, whose cooperation is needed
by almost every developer.)
This has upsides: I'm personally strongly aligned with ftpmaster's
primary goals, and very supportive of most of their controversial
decisions (especially, the ones about what counts as source code).
But: the difficulties we have with ftpmaster are very deep-rooted, and
not simply a lack of volunteers.
Ian.
--
Ian JacksonThese opinions are my own.
Pronouns: they/he. If I emailed you from @fyvzl.net or @evade.org.uk,
that is a private address which bypasses my fierce spamfilter.
Bug#1018955: ITP: ktx -- QuakeWorld mod "Kombat Teams eXtreme"
Package: wnpp Severity: wishlist Owner: Lee Garrett X-Debbugs-Cc: debian-devel@lists.debian.org, deb...@rocketjump.eu * Package name: ktx Version : 1.40 Upstream Author : QW Group * URL : https://github.com/QW-Group/ktx * License : GPLv2+ Programming Lang: C Description : QuakeWorld mod "Kombat Teams eXtreme" KTX (Kombat Teams eXtreme) is a popular QuakeWorld server modification, adding numerous features to the core features of the server. Although it had been developed to be Quakeworld server agnostic, it has over the years been developed very close to MVDSV to which it has become an extent, thus compatibility with other Quakeworld servers might not have been maintained. This package is an addition to mvdsv (#862954). Upstream has a slow turnover of =< 2 releases per year. I intend to team maintain it in the games team. (no need to sponsor, I'm a DD)
Re: Half the world being removed
Hi, On 02-09-2022 13:00, Ian Jackson wrote: I wonder if it would be possible to detect a sudden large increase in the number of autoremovals, and stop the autoremoval system instead of causing blaring klaxons for everyone in the project ? I disabled the cron job that sends out mail yesterday, so the massive klaxons shouldn't go off (or are there klaxons I'm not aware of). And indeed such a check would be worth while. But to be fair, the code is in Perl and I would be afraid that by adding the check I introduce more bugs than I solve. Paul OpenPGP_signature Description: OpenPGP digital signature
Re: Half the world being removed
Paul Gevers writes ("Re: Half the world being removed"):
> On 02-09-2022 13:00, Ian Jackson wrote:
> > I wonder if it would be possible to detect a sudden large increase in
> > the number of autoremovals, and stop the autoremoval system instead of
> > causing blaring klaxons for everyone in the project ?
>
> I disabled the cron job that sends out mail yesterday, so the massive
> klaxons shouldn't go off (or are there klaxons I'm not aware of).
Well, personally I have a grep-excuses cron job that nags me daily
about autoremovals, becaue the official cron job is rather late.
Also it shows up in tracker.d.o for every package, DDPO, etc.
I think to properly silence the alarm it's necessary to prevent the
decision to autoremove, not just the warnings. (There are other
obvious possible difficulties ...)
> And
> indeed such a check would be worth while. But to be fair, the code is in
> Perl and I would be afraid that by adding the check I introduce more
> bugs than I solve.
Do you have someone who will do code review for you ? I hereby
volunteer. (Send me mail off-list, or something.)
Regards,
Ian.
--
Ian JacksonThese opinions are my own.
Pronouns: they/he. If I emailed you from @fyvzl.net or @evade.org.uk,
that is a private address which bypasses my fierce spamfilter.
Bug#1018974: ITP: pangomm2.48 -- C++ wrapper for Pango
Package: wnpp Severity: wishlist Owner: jeremy.bi...@canonical.com Control: block -1 by 1018908 Control: block 1006728 by -1 X-Debbugs-CC: debian-devel@lists.debian.org, debian-gtk-gn...@lists.debian.org Package Name: libpangomm-2.48-1 Version: 2.50.0 Upstream Author: The pangomm Development Team License: LGPL-2.1+ Programming Lang: C++ Description: C++ Wrapper for pango Pangomm is a C++ wrapper for the pango library. Originally part of gtkmm, pangomm provides convenient C++ interfaces for handling both the layout and internationalization of text in graphical applications. Other Info -- This library will be maintained by the Debian GNOME team. Packaging is at https://salsa.debian.org/gnome-team/pangomm2.48 The existing pangomm source package tracks releases from the pangomm 2.46.x series (1.4 ABI) and is intended for use with gtkmm3.0 (which uses GTK 3). The new pangomm2.48 ABI is intended for use with gtkmm4.0 (which uses GTK 4). The pangomm2.48 source package naming has been adopted by other distros: https://repology.org/project/pangomm/versions Thanks, Jeremy Bicha
Re: Comments on proposing NEW queue improvement (Re: Current NEW review process saps developer motivation
Hello,
On Fri 02 Sep 2022 at 12:31PM +01, Ian Jackson wrote:
> Sean Whitton writes ("Re: Comments on proposing NEW queue improvement (Re:
> Current NEW review process saps developer motivation"):
>> On Sat 27 Aug 2022 at 04:22PM +02, Vincent Bernat wrote:
>> > It does not seem to work. Either people don't want to do that, either the
>> > FTP
>> > team is too picky on the candidates.
>>
>> Some combination of both, but I don't think I'm suffering from bias if I
>> say that it's at least 80% the former. Very few people who say they'd
>> like to be trained confirm they'd still like to once they've had a look
>> at the docs for trainees, and after that, hardly any do enough trainee
>> reviews for the other team members to feel confident they can let them
>> at it on their own.
>
> I am in this picture. Some years ago now I volunteered. I was
> introduced to the internal ftpmaster documentation and processes. At
> the time, these documents were not even published - including,
> astonishingly, some elements which read like a manifesto. (I don't
> know if these documents are published nowadays.)
They're all published, unedited
(one of the first things I worked on when I got involved).
--
Sean Whitton
signature.asc
Description: PGP signature
