[pkg-wine-party] Anything to help with?

[Ove Kåven]

Den 05. aug. 2011 23:08, skrev Stephen Kitt:
> Would you consider adding me to the wine-pkg-party Alioth project so
> we can both work on wine-gecko in a git repository there?

Perhaps, but I'm not sure how reasonable it will be to set up a git
repository for wine-gecko at this time.

First of all, there's the issue of how to manage a quilt-format package
in git. I haven't determined the best way to convert the other Wine
packages to the quilt format myself. Apparently many debian developers
are still debating the issue.

In my opinion, this is a sign that Git is, in fact, a poor choice for
Debian packaging. No doubt Git is a powerful tool for its uses - but Git
is designed for *upstream* developers, not for packagers. On the other
hand, there's certainly a lot of fanboyism surrounding Git, with people
trying to hack Git into fitting every purpose they can think of, so I'm
sure people can (and have) come up with workflows suitable for Debian
packaging with Git. But that doesn't mean it'll be pretty, except
perhaps to the starry-eyed.

Second, even if it turns out the main Wine packages could be reasonably
managed with Git, I don't think the same could necessarily be said for
wine-gecko. For one, there's the whole DFSG issue. If upstream
distributes files that need to be removed, should they be in Debian's
git? If so, how easy is it to make DFSG-free tarballs from that git
tree? And the wine-gecko sources are massive, but releases aren't
frequent. Is it really appropriate to track this thing in a typical
everything-and-the-kitchen-sink-stored-everywhere Git tree, using
whatever workflow the Git advocates have come up with to bludgeon Git
into doing something worthwhile?

Currently, even with the help of git-buildpackage (which does make Git
more tolerable for packaging), managing the Wine package in Git has been
a pain, I think. I'm really not convinced I need that pain for something
like wine-gecko as well.

When the main Wine packages have been converted to quilt format, it may
be useful to consider how the resulting workflow would work for
wine-gecko. But I don't really want to put wine-gecko into a VCS before
that.