Control: tag -1 + pending patch
Hi Samuel,
Samuel Thibault wrote:
> Seeing that qa.debian.org is using version 2.114.123 of lintian, I
> upgraded my lintian from its git tree, only to find that it seems I
> cannot update my suppression rule according to the new output:
>
> W: libbrlapi-dev: bad-whatis-entry [usr/share/man/man3/brlapi_authClientPacket_t.3.gz]
> W: libbrlapi-dev: mismatched-override bad-whatis-entry [usr/share/man/man3/brlapi_*] [usr/share/lintian/overrides/libbrlapi-dev:2]
>
> I did have updated this rule to include brackets:
>
> bad-whatis-entry [usr/share/man/man3/brlapi_*]
>
> but that doesn't seem to be working. I also tried to use escaping (which
> had fixed things for the ${} case), but to no avail:
>
> bad-whatis-entry \[usr/share/man/man3/brlapi_*\]
Using * or ? in place of the brackets probably would have worked.
Cause was the switch of was this commit from 6 months ago:
commit 139009d5a54225ebff4509ec37b979cb898c17fe
Author: Felix Lechner <felix.lechner@lease-up.com>
Date: Wed Dec 1 21:46:24 2021 -0800
Use Text::Glob to match hint contexts with override patterns.
Replaces a trusted homegrown routine.
Gbp-Dch: ignore
I neither understand why you would replace something "trusted" and
additionally not even plan to mention such a invasive commit in the
changelog.
And because Text::Glob also interprets brackets (and curly braces) and
you can't disable that, such issues were just waiting to happen.
And together with the brackets using "pointed hints" concept
introduced at around the same time, this caused really havoc.
> Lintian has recently been annoying enough that I'm unsure I'd continue
> monitoring its output any more.
*nod* JFYI, in case you didn't notice it (it happend months after your
bug report), but there was an "O: lintian" WNPP bug report in the
meantime: https://bugs.debian.org/1012289
So I'm now trying to clean up the shattered remains which the previous
(recent) Lintian maintainer(s) left behind.
For now I'm trying to get a current lintian version back into Testing,
i.e. I'm concentrating on RC and important bugs as well as bugs which
are annoying and easy to fix, e.g. those with a patch or MR.
At least the lintian internal test suite is no more failing. Although
it likely has some false negatives as it seems that the test suite is
no more able to run more than one check in one test. And running a
test against multiple checks at the same time seems critical to write
tests for issues like this one. Meh.
Anyway, my feature branch for this bug report is
https://salsa.debian.org/lintian/lintian/-/commits/brackets-in-overrides-rc-bug-1003353
but I will likely force-push that branch some more times, so the
commit message in
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1003353#50 might
become stale soon while I refine the commit before that (which
contains a test for this bug), but the patch will likely stay valid.
You also might get further such mails when I force-push new histories
in that feature branch. I hope you don't mind, but I wanted to publish
what I already have.
Regards, Axel
--
,''`. | Axel Beckert <abe@debian.org>, https://people.debian.org/~abe/
: :' : | Debian Developer, ftp.ch.debian.org Admin
`. `' | 4096R: 2517 B724 C5F6 CA99 5329 6E61 2FF9 CD59 6126 16B5
`- | 1024D: F067 EA27 26B9 C3FC 1486 202E C09E 1D89 9593 0EDE
Attachment:
signature.asc
Description: PGP signature