Re: RFS: QA Upload - muine 0.8.8-1 - Simple playlist based music player
On Wed, Feb 20, 2008 at 11:21:10AM -0500, Barry deFreese wrote:
> Frank Lichtenheld wrote:
> >On Tue, Feb 19, 2008 at 11:21:19PM -0500, Barry deFreese wrote:
> >>I have prepared a QA upload for the orphaned package muine which
> >>includes an new upstream that fixes RC bug #440817, and fixes a few
> >>other bugs.
> >>(#415419, #427263, #449835, and probably several of the bugs posted
> >>against the 0.6.x versions). If someone could please review and/or
> >>upload I would appreciate it.
> >
> >from changelog:
> >
> >>* Remove -X revisions on build-depends.
> >>
> >
> >Huh? I can see that removing -1 is most often a good idea, but -3?
> >
> >
> It was my understanding that build-depending on any Debian revision was
> a bad idea? Why would -3 be any better than -1? There shouldn't be any
> API/ABI changes between Debian revisions should there? Obviously, I
> guess it makes sense if there was some specific bug fix but it still
> seems like a bad idea, but what do I know? :)
It is true, that it is always better not to depend on a specific Debian
revision, since that makes it easier for backporters, that e.g. upload
a -1~bpo (which sorts earlier than -1). This is not always correct
though.
The reasoning for the suggestion to remove -1, is that there should
really be no uploads before the -1. But if someone depended explicetly
on a later Debian revision (-3 in this case) I would assume that he had
a good reason to do so (example bugs would be broken shlibs files,
reverting some upstream API/ABI breakage). The chance that the -1 was
really meaningful is usually slim (could happen in case someone uploaded
a -0.1 NMU), the chance that -3 was really meaningful is significantly
higher.
So either you should prove that removing the -3 is harmless or you
should just leave it there. I would be willing to bet on the fact that
removing the -1 is harmless.
Gruesse,
--
Frank Lichtenheld <djpig@debian.org>
www: http://www.djpig.de/
Reply to: