[aur-general] Tarball Guidelines

Loui Chang louipc.ist at gmail.com
Thu Dec 9 20:10:22 EST 2010


On Tue 07 Dec 2010 16:49 +0100, Xyne wrote:
> keenerd wrote:
> 
> > On Mon, Dec 6, 2010 at 4:59 AM, Nicky726 <nicky726 at gmail.com> wrote:
> > > been told by the bot, that selinux-flex has a binary
> > > (selinux-flex/flex- arch.patch.gz), which is a gziped patch. Guess
> > > I can ungzip it, though as this package is just a copy of a [core]
> > > package from some time ago, I guess the original maintainers new,
> > > what they were doing, if they included it this way.  So should I
> > > do it to not include evil gziped patches?
> > 
> > Evil is such a strong word.  It is just without benefit.  Disturbs the
> > transparency of things.  Technically against the rules.
> > 
> > Zipped patches was an edge case.  Here, I chose to take a strict
> > interpritation of the edge cases.  It is only a comment after all,
> > very little of consequence.  Besides, Arch tries hard to not patch
> > things  :-)
> > 
> > But thank you for taking the time to read and respond.  So many
> > maintainers ignore comments.
> 
> If the patch is large then what's the problem with compressing it?

I would argue that we should not have large patches applied to Arch, or
AUR packages at all. If there is enough patching, that constitues a
fork, and we shouldn't be hosting project files for defacto forks on the
AUR.  They should find some other place to host their project. That
large patch, and any other source tarballs should be downloadable from
the project's webspace, not the AUR.



More information about the aur-general mailing list