On 28 August 2024 16:32:00 GMT+02:00, Marcell Meszaros
<[email protected]> wrote:
>On 28 August 2024 16:00:47 GMT+02:00, Vitalii Kuzhdin
><[email protected]> wrote:
>>Greetings,
>>
>>I suggest you actually examine the PKGBUILD before submitting deletion
>>requests. To address your point, there isn't a standalone Python version
>>'311' either; it's a reference to Python 3.11, which exists in the AUR
>><https://aur.archlinux.org/packages/python311> and is quite popular. I
>>followed a similar naming convention here, with the proper naming specified
>>in the provides array.
>>
>>The idea of merging all dependencies into a single PKGBUILD, simply because
>>it’s possible, is impractical—and I trust I don’t need to explain why. This
>>is especially true for older source trees that may require future patches.
>>
>>The fact that a package currently has only one dependent, just a few days
>>after being uploaded, doesn’t invalidate its existence. Many packages exist
>>without any dependents and are still valuable. My package serves a purpose,
>>and bundling multiple packages into a single one (like cutter-test) is
>>impractical. Furthermore, removing cutter-test would be unreasonable, given
>>the number of packages that depend on it. I am actively reaching out to
>>other maintainers to integrate testing with cutter-test.
>>
>>Best regards,
>>Vitalii
>>
>>On Wed, Aug 28, 2024 at 4:46 PM <[email protected]> wrote:
>>
>>> MarsSeed [1] filed a deletion request for gstreamer113 [2]:
>>>
>>> There is no gstreamer version 113 (one hundred and thirteen).
>>>
>>> This package was made for use by one package, cutter-test.
>>>
>>> Please don't duplicate repo packages for such a purpose; build your
>>> specifically needed gstreamer version inline in the cutter-test
>>> PKGBUILD, and install it in a location that is specific to cutter-test
>>> only.
>>>
>>> [1] https://aur.archlinux.org/account/MarsSeed/
>>> [2] https://aur.archlinux.org/pkgbase/gstreamer113/
>
>The 'python311' pkgname, but the package itself has legitimate purpose as a
>standalone PKGBUILD. (Even if it has grave packaging mistakes like its
>provides=python entry.)
>
>And, the key difference is, python311 does not conflict with repo's python
>package, whereas your gstreamer113 does. With this, it can potentially break
>hundreds of packages, if they were built against the repo version and not your
>legacy release.
>
>AUR submission guidelines explicitly say not to duplicate the repo builds, not
>with same or different release versions.
>
>If you don't want to merge this build into cutter-test, please make a
>different 'gstreamer-1.13' package that does not conflict with repo's
>gstreamer and does not declare provides=gstreamer.
I wanted to say, the 'python311' pkgname is wrong. (But still, that PKGBUILD is
legitimate, and it does not conflict with repo's python pkg.)
