Re: [cjwatson@debian.org: Accepted grub2 2.02~beta2-7 (source i386)]
On 11/03/14 10:50, Colin Watson wrote:
> "Distribution: unstable". Whoops. It is far too easy to make
> this mistake with sbuild if you're not quite paying absolutely
> perfect attention. My apologies ...
On the Lintian side of things, I attached a patch to
<https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=542747> in 2010 and
am still waiting for comments.
My patch just looks for Distribution=unstable, Changes=experimental
because that's the most annoying case ("I uploaded a version that
wasn't ready and now I have to use an epoch to roll it back").
On the other bug, Thorsten said:
> a) Distribution=unstable, Changes=experimental b)
> Distribution=experimental, Changes=unstable c) Distribution=*,
> Changes=UNRELEASED
>
> All other mismatch cases should probably be a W, but for these
> three, I think not just an E but an autoreject could, maybe, be
> justified.
which I think is also reasonable.
Russ said
> I think we only want to do this check if the first line of the
> Changes file says UNRELEASED, since there are valid use cases for a
> mismatch otherwise.
but I don't know what those valid use-cases are. BinNMUs in testing
for a package uploaded to experimental, possibly? Anything else?
On the sbuild side:
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=529281
tl;dr: extensive discussion of why this mistake is easy to make and
why sbuild can't trivially fix it; Raphael Hertzog suggests making
"sbuild foo_source.changes" do the right thing; Thorsten Glaser
suggests lintian autorejects for certain particularly bad and likely
combinations; Roger Leigh points out that just "sbuild" with no .dsc
file will now do the right thing.
Regards,
S
Reply to: