[R-pkg-team] Bug#1125226: Bug#1125226: Bug#1125226: src:r-cran-admisc: fails to migrate to testing for too long

[Charles Plessy]

Le Sun, Jan 11, 2026 at 11:39:33AM +0100, Paul Gevers a écrit :
>
>Noted, you mean all r-cran-* packages on the list right? (I'm doing 
>this semi-automatically). I checked some before filing and the 
>packages were missing RM bugs, otherwise I wouldn't have submitted 
>these bugs. As a Release Manager I still expect you to handle the 
>situation with a bit of urgency, the upload happened more than 30 days 
>ago and I'm not aware of the resolution how you're intending to handle 
>it with ftp-master.

Hi Paul,

thanks for the fast answer,

I am doing it as fast as I can allocate time to it, mostly at work, and
it is not my main mission.  The bottleneck in opening RM bugs is that
either it is one bug by package, leaf by leaf, or I can submit bugs for
whole branches but I need to be sure that it will not cause the removal
of packages not mentionned in the request, and my attempts to explore
the dependency graph are blocked by circular dependencies.

>As noted in the bug text, the bugs are already closed, they merely 
>serve to expose the issue (successfully ;) ), track them [1] and cause 
>autoremoval for non-key packages. If the blockers of the migration get 
>solved and the package migrates, the bug is automatically no longer 
>relevant.

I am sorry that I did not read it.  I am catching up with rusty skills
and I really appreciate your patience.  In that case please go ahead
opening bugs if they are useful to you.

>PS: note that the migration failure of r-cran-av is one of the reasons 
>why the ffmpeg transition isn't done yet.

Please feel free to remove immediately from Testing any r-cran- /
r-bioc- package that we maintain.  For the self-contained part of the
dependency graph that involves exclusively our packages, I fee that we
do not have much of a moral contract with the users of these packages to
guarantee their continuous availability in Testing.  For instance we are
skipping entirely a Bioconductor release and nobody complained about it.

I am really sorry for the trouble r-cran-av causes, but I do not have
the tools to see it happening.  As of now the tracker page of the
package does not mention that it is involved in a transition.

I tried to compute the reverse-dependency graph of r-cran-av but it
seems that there are circular dependencies between the package imports
(represented by Debian Depends) and the package regression tests
(represented by Debian Recommmends because we can not distinguish
between true Recommends and debian/test/control dependencies), is
is 10pm and I will not be able to open a RM bug today…

Frankly speaking, I wish we could reboot the whole set of r-cran-*
packages.  They are causing pain to eveybody and for most of them we do
not remember why we added them and who still needs them.  But this is not
possible; we are not going to send hundreds of packages back to NEW…

In any case, I hope that we will at least have strengthened the
situation by removing non-64-bit non-little-endian arches by ~March.

Cheers,

Charles

-- 
Charles Plessy                         Nagahama, Yomitan, Okinawa, Japan
Debian Med packaging team         http://www.debian.org/devel/debian-med
Tooting from work,               https://fediscience.org/@charles_plessy
Tooting from home,                 https://framapiaf.org/@charles_plessy