I couldnt work out where to ask this: Are emails about packages being
removed from testing generated using stale data on what bugs are open?
(are im not complaining about the process just trying to understand it)
According to
https://udd.debian.org/udd-status.cgi data on bugs is
updated a lot less frequently than data on testing-autoremovals, is
there a reason more frequent updates of the bts data can't be done?
I think this led to the following:
Wed, 18 Sep 2024 08:27:03 UTC
- A (completely valid) rc bug against chkrootkit 0.58b-1 was opened
(
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1082087)
- At some point a (valid) warning is produced saying that the package
(0.58b-1) would be removed from testing due to this bug -- this is all
as expected.
(Thu, 26 Sep 2024 06:34:02 +0000
- 0.58b-2 was uploaded with a changelog closing the bug, and other
changes which meant some autopkgtests failed. This never made it to
testing -- all fine, 0.58b-3 was prepared to solve that)
Mon, 30 Sep 2024 14:34:02 +0000
- 0.58b-3 was uploaded which fixed everything
(
https://tracker.debian.org/news/1570784/accepted-chkrootkit-058b-3-source-into-unstable/)
- by coincidence, this version was just after a new glibc was
uploaded, so migration to testing had to wait until that migrated
-- all fine and 'expected'.
So far, this all seems to be working as expected. The warning about
removal of 0.58b-1 from testing was still 'valid', because testing was
still affected by an rc bug.
Thu, 10 Oct 2024 04:39:07 +0000
- 0.58b-3 makes it into testing.
https://tracker.debian.org/news/1574012/chkrootkit-058b-3-migrated-to-testing/
At this point, the version in testing is no longer affected by any rc
bugs. The bts seems to know this. However:
Fri, 11 Oct 2024 04:39:06 +0000
- the autoremoval script sends an email saying
"chkrootkit 0.58b-3 is marked for autoremoval from testing on 2024-11-08"
due to it "being affected by" 1082087
I assume this is a false positive, but it's quite confusing as it seems
to have got the latest version in testing correct, but is wrong about
what bugs affect -- i'd assume there is always a race condition here,
but 0.58b-3 was known not to be effected by that bug for well over a
week.
the email pointed me to
https://salsa.debian.org/qa/udd/-/raw/master/udd/testing_autoremovals_gatherer.pl
but it doesnt say how old the data is. I found
https://udd.debian.org/udd-status.cgi which suggets info on bugs are
updated less frequently than most things, but not sure that is the issue
here?
I assume the actual removal process would use up-to-date data?
--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)