[aur-general] Handling coincidental name collisions

alad alad at archlinux.org
Sat Feb 9 13:36:16 UTC 2019


Am 09.02.2019 um 14:34 schrieb Xyne:
> Hi everyone,
>
> This is in regard to this thread on the forum:
> https://bbs.archlinux.org/viewtopic.php?id=244051
>
> The packaged contained this project:
> https://github.com/Aniket-Pradhan/lsd
>
> To summarize the thread, an AUR package that had existed for a while was
> deleted when an unrelated package of the same name was moved to [community].
> The reason given was that the AUR package was "not useful enough", either
> because it only had 2 votes or because the acting TU saw no personal use for it.
>
> For trivial packages, it would be good to at least clarify the reason for
> deletion in a little more detail. There are plenty of AUR packages that persist
> for years with 0 votes so a maintainer with 2 votes may be understandably
> confused by the terse statement "not useful enough". A little clarification can
> easily disperse that confusion and better guide the user through future
> contributions.
>
> However, "trivial" here usually means that someone uploaded a bash script to do
> something like open arandr and click on it with xdotool to save half a second,
> or baked some convoluted ls-cat-cat-grep-cat-sed-cat pipe into a 3-line script.
> The project involved here is not in the same category. It may not be practically
> "useful" for many users, but it does do something that is not trivial to
> replicated in a few lines of shell code. It's "usefulness" is subjective.

The "original" lsf looks like a joke/troll package to me, rather than
"trivial". I'd have deleted it even without community duplicate.

Alad

>
> In this case, the TU should have proposed renaming the package, given the
> maintainer some time to pick a new name and re-upload the package, and then
> merged the old one. Even a single vote from another user can be encouraging so
> the merge is worthwhile unless the maintainer states otherwise.
>
> Regards,
> Xyne


More information about the aur-general mailing list