[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#838014: RFS: imagemagick



Le samedi 17 septembre 2016 08:50:56 CEST, vous avez écrit :
> control: owner -1 !
> control: tag -1 moreinfo
> 
> On Fri, Sep 16, 2016 at 02:29:45PM +0200, Bastien ROUCARIES wrote:
> >   I am looking for a sponsor for my package "imagemagick"
> 
> great!
> 
> >     dget -x
> >     https://mentors.debian.net/debian/pool/main/i/imagemagick/imagemagick
> >     _6.9.5.9+dfsg-1.dsc
> I've eventually managed to domate gitpkg, and I succeeded at building a
> .dsc out of it.  dsc that matches yours in all the things, except for
> the timestamps in debian.tar.xz.  Am I right that you built your package
> in jessie?  (Or anyway, with a dpkg << 1.18.10, which is the one that
> started to build reproducible source packages).

I build in jessie but using pbuilder....
> 
> This also means that with me you don't really need (anymore) to provide
> me a .dsc to dget, git pushing changes is going to be enough.
> 
> Another question about that git repository: I see there are very few
> tags.  Shall I suppose tags means nothing in this repository format?
>
Means nothing. They are only branch

> 
> I recall you were saying something about this aiming to experimental,
> because some SONAME changed (I didn't really checkout), which must also
> be the reason you're asking for sponsorship (=> the package will go
> through NEW, so you can't upload yourself).  But your upload is
> targetting unstable.

Corrected

> 
> Things I want to see fixed:
> 
> 1/ you didn't incorporate any NMU, nor mine (which upsets me as a NMUer
> trying to do QA work), nor pochu's (which is fixing a CVE, that I
> suppose has been fixed upstream already so that's not so relevant.
> Not including my NMU means your package gains another RC bug, and just
> because you're lucky (src:kodi has yet to be fixed…) your package would
> migrate to testing, but in normal circumstances it wouldn't.
> The changelog is also lacking the unstable uploads for -6 and -7.  I can
> understand -7 where you only backported some patches that are probably
> already present, but judging by the changelog I suppose -6 is also some
> packaging changes; have them be included?

Change included and NMU aknowledge. This version go back before the NMU

> 2/ the Vcs-* field, Vcs-Git is plain broken.  If I were you I'd set both
> Vcs fields to
>     https://anonscm.debian.org/git/collab-maint/imagemagick.git
> That url works both as a cgit frontend.
> It's something that I'm pointing out and "forcing" to be fixed due to so
> seldom uploads of imagemagick...  and it's fairly annoying not being
> able to use `debcheckout`.
> 

Done

> Other things that you may consider changes in future, not needed now:
> I've only peaked randomly at d/rules, nothing more, so this is probably
> incomplete.
> 
> 1/
> STATIC_DEB_SOURCE_PACKAGE could be take by
>     dpkg-parsechangelog -SSource
> Also STATIC_DEB_VERSION you could use
>     dpkg-parsechangelog -SVersion
> instead of doing grep/cut.  Or you could include
> /usr/share/dpkg/pkg-info.mk and use the variables provided by it (not
> exported though).

Done. Need to be exported so I do not think will use pkg-info

> 2/
> Actually I'm really not sure why you export all the variables...  I'm
> sure you don't need all of them exported.

Will review for next version

> 3/
> you have a weird if in the % target.  The bug cited there is long fixed,
> so probably that if is not needed anymore?

Bug number was false. Corrected the bug number. This bug is still opened and 
in automake

> 4/ you got a 'fpu_is present' target, totally unused.  The most funny
> bit is the presence of the space in the target name.  I figure you are
> better by removing the target.  And if you don't want to remove it, then
> you might avoid a small DRY violation by making override_dh_autoreconf
> depend on that target (after the name is fixed).

Removed

> 5/ all the patches have 'Forwarded: Yes'.  That's cool, but according
> to DEP-3, Forwarded should contain the place where they have been
> forwarded, not just "Yes".  i.e., a patch reviewing place, an upstream
> bug, a mailing list post, ...

Sending by mail to upstream. Will add indication in next version.

Now building will remove package from mentors and uploading a new one

Bastien

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply to: