Hi Salvatore, On Sat, Jan 24, 2026 at 04:21:52PM +0100, Salvatore Bonaccorso wrote: > With the refactorings of the bts command the following was not > possible anymore which I wanted to use for updating the metadata for > the bug #1125797: > > $ bts -n found 1125797 6.17.8-1 , found 1125797 6.18~rc4-1~exp1 , done > 1125797 6.18.5-1 , done 1125797 6.19~rc4-1~exp1 > bts done: Duplicate 'done' command for Bug#1125797! at /usr/bin/bts line 2099. > > Mutltiple closings of bugs would be allowed by BTS and is AFAIU as > well what happends when one closes the bug in the various supported > releases (e.g. first in a unstable upload, followed by a stable update > later). > > I see that the consistency checks explicitly mention: > > Consistency Checks > > • done: checks whether the bug is alread marked 'done', > > but > > Many of the checks involve fetching online information from the BTS. > Use > --offline to disable these. Purely bts internal consistency checks > such > as marking the same bug as done twice cannot be bypassed this way. > > So my request is somehow in conflict with the implemented new > consistency checks.
I consider your use-case is ofc. perfectly valid.
The logic of the consistency check is simply not sensitive to the version
number argument right now. A mistake in the check.
As a workaround please just run the (second) overlapping done command
separately.
Alternatively if you're adventurous s/done/fixed/ as the second `done` does
not affect the bug's completion state anyhow -- hence `fixed` should be
equivalent here:
Eg. replace:
done 1125797 6.18.5-1 , done 1125797 6.19~rc4-1~exp1
with
done 1125797 6.18.5-1 , fixed 1125797 6.19~rc4-1~exp1
BTW: The new "(omit) implicit bug" syntax will save you some typing in
cases like this as you only need to mention the bug you're working on once:
$ bts -n found 1125797 6.17.8-1 , found 6.18~rc4-1~exp1 , done 6.18.5-1 ,
fixed 6.19~rc4-1~exp1
See
https://manpages.debian.org/unstable/devscripts/bts.1.en.html#The_Implicit_Bug
"Alternatively you can also omit the bug number entirely [...]".
I'm working on a fix, but it proved harder than expected since we talked on
IRC. I'm short on Debian time ATM and a shortcut I took in the
implementation needs to be unravelled in order to fix this properly.
Sorry about the annoyance and thanks for the reminder (bug)!
--Daniel
signature.asc
Description: PGP signature
