Re: Status for RT on `Rules-Requires-Root: no` transition

Paul Gevers:
> Hi Niels,
> 
> On 14-01-2025 10:46, Niels Thykier wrote:
>> Another update on the transition
> 
> One month further. What's the current status?
> 
> Paul
> 

Hi,

Key package-wise, we are down to 5 packages affecting testing:

  * d-i + shim + its helper: All which are not actionable to me as
    stated previously.

  * bash: There is now a NMU in DELAYED/5.

  * lintian: There is now a NMU in DELAYED/5 disabling the 4-5 affected
    tests until the maintainers has time to look into a better fix.

    Note: Someone removed the lintian bugs from the list of usertags
    bugs without notifying me, so I almost missed it was still open.

For non-key packages, only 1 affects testing. It has just been fixed in 
sid via NMU (approved by the maintainer).

Beyond that, to my knowledge there has not been any recent influx of R³ 
related bugs, so I am inclined to consider the transition closed once 
bash and lintian hits testing (since I cannot do anything for the 
remaining key packages).

That concludes my status.

> PS: I happen to track wordpress, which happened to be dropped from 
> trixie because of this transition (it's on its way back). It's exactly 
> those kind of removals that I was hoping you would prevent/track. I know 
> you're a volunteer (thanks a lot for that, you know I appreciate your 
> work :) ) and I know the maintainer had key expiry issues and I know you 
> offered help, but a ping to the bug would have prevented the removal.

Having thought about this paragraph for a while, I feel we should have 
thought about and agreed to use the "disable auto-removal" tags for 
these bugs, since it would have made the "system" fail in the direction 
you wanted (fewer auto-removals). I forgot the tag existed until today 
when I was replying to this email.

That tag makes for a much better way to minimize auto-removals than rely 
on humans to be email sending cronjobs ensuring the correct failure mode 
when the human is unreliable or temporarily out of spoons for Debian in 
a time sensitive period (which was the case here with wordpress for both 
me and apparently also the wordpress maintainer).

> I hope wordpress was the exception.
> 

Several packages have been auto-removed during this transition, which I 
have been quite open about (see previous status emails from me). In that 
sense, `wordpress` was not an exception.

In total, we are talking ~15 packages with about 5 of them being 
"permanent" (suggested removal from sid or removed from testing due to 
    unrelated RC bugs).
   Most of the temporary ones are fixed in sid now. I am leaving one of 
them for a package with an open ITS, because I do not want to cause the 
ITS'er extra work assuming the maintainer does not respond (which they 
do not appear to be) and since my work will not prevent an auto-removal 
at this point (The ITS being #1094601).
   Of the ~10 temporary removals, exactly 4 of them involved the 
maintainer asking me not to NMU or to cancel my NMU. All 4 have now been 
fixed by the respective maintainers in sid.


As mentioned above, if you feel these were "10 too much", then I 
recommend you ask people to apply the "disable auto-removal" usertag to 
the transition bugs and have them "opt-in" to auto-removals for the next 
transition.


Best regards,
Niels

PS: I use approximations for most of the auto-removal numbers since 
exact numbers are hard to come by without going over all ~250 packages 
involved in the transition, which I am not going to do. The problem is 
that once auto-removed from testing and fixed in sid, they disappear 
from the UDD bug page (I assume) because they are no longer "open" in 
any suite. The numbers are based on notes, previous status, memory and 
bugs still listed as open in sid on the UDD bug page (though some of 
them will disappear soon given recent uploads).

-----BEGIN PGP SIGNATURE-----

iQEzBAEBCgAdFiEE9ecZmu9eXGflVYc/dA1oiINl0okFAmeyazQACgkQdA1oiINl
0ok5kgf+J2il9qNrnX/R9DvxU05t48MmpgYAikqkpD5X87cD/ORlwOn8j0VrR94D
yk/jLpvGKTOSoMTn99w+i5bHxMa0BD5Zdz/0GXUYkMcp8r05sEPhSADh6/CQ44ny
eCOKvUypqWP2/Hfp0Uhxs9IyyePnP42fosIcrlk2bnGZQQEzCZDeCY15IboL7uUt
YzfRTeeD/h8vatMu6dAxvK0+uiHunQ9MR1v62khZk9qjSWDOdPYAc8uF7rtkdz1G
rYem+QYn8tgG66RA85RleFz9tGgMc00PdmtVehDSZgc9bgP1/f93Ubgx6CQAKihN
tq2jpVaTQpjPzUMSMoCA4/K+YG6okw==
=cmK6
-----END PGP SIGNATURE-----