On 12/3/19 2:27 PM, Jonathon Fernyhough wrote:
On 03/12/2019 18:26, notify@aur.archlinux.org wrote:
rober_k [1] filed a deletion request for gtk3-classic [2]:
this is an identical package as gtk3-mushrooms . It pulls from the same repo, only is newer. Proper way would be to mark gtk3-mushrooms out-of-date, not just upload a different package with same content ;)
[1] https://aur.archlinux.org/account/rober_k/ [2] https://aur.archlinux.org/pkgbase/gtk3-classic/
It's not a simply a more up-to-date package. The source repo is the same but gtk3-classic is essentially a forked gtk3-mushrooms which adds experimental/developmental (?) aspects to the PKGBUILD which the original does not (e.g. use of meson and quilt, see e.g. [1]).
"Use quilt and meson" is not a valid reason to upload a new PKGBUILD. The resulting package does not care what technology you use for applying patches, and it *should* not care whether you use autotools or meson to generate a series of compilation commands. Is there something that the resulting .pkg.tar.xz does, which is different from "gtk3-mushrooms"? If it is just about the lib32-* variant, you can just upload a lib32-gtk3-mushrooms package inspired by https://git.archlinux.org/svntogit/community.git/tree/trunk/PKGBUILD?h=packa...
Also, technically the gtk3-mushrooms package is not out-of-date as the upstream is still at 3.24.11.
If all of this is potentially too confusing let me update the PKGBUILD to use a different source rather than having the package deleted.
I don't think changing the url= field makes a difference, if that is what you mean.
-- Eli Schwartz Bug Wrangler and Trusted User