[aur-general] Unexpected aur package removal.

[Yichao Yu]

On Wed, Mar 13, 2013 at 11:01 PM, Daniel Wallace
<danielwallace at gtmanfred.com> wrote:
> On Thu, Mar 14, 2013 at 12:05:04AM +0000, Xyne wrote:
>> Ionut Biru wrote:

Well, technically, it is just ~23h from my last email but hopefully
Brock.Zheng will agree to call this "one day". :P

>>
>> >mate, you did something wrong, stop trying to find out excuses for your
>> >actions.
>> >
>> >please do not delete builds like that anymore, leave comments about any
>> >issues with the build and give them time to accommodate with the changes
>>
>> This.
>>
>> Obliterating votes and comments and forcing the user to recreate the package
>> does not help anyone, and one minute is not enough to read a comment, address
>> the issue and re-upload the package even if the maintainer is glued to the
>> chair slamming F5. It discourages new users from continuing to contribute and
>> it leaves a very bad impression of the TUs and Arch by extension.
>
> Did you actually read what I explained what I did? because your first
> sentence would cause me to believe you didn't.  I specifically said I
> did not delete any packages that had comments on them or any substantial
> amount of votes.  This one packages that is being complained about had 3
> votes, that is 2 more than any other packages, and all of the other ones
> were only voted on by the maintainer.
>
>>
>>
>> Bartłomiej Piotrowski wrote:
>>
>> >What's the point of your drama? It was perfectly correct to remove it
>> >(especially if it was new package). You reuploaded and fixed it.
>> >Everyone is happy, hurray!
>>
>> Everyone is clearly not happy. If you honestly think that it's just "drama"
>> when a user has a valid complaint about an overzealous TU then I think you need
>> to reconsider your viewpoint. TUs are not beyond scrutiny. If anything we
>> should be held to a higher standard than other users.

I was happy before this last reply since I think everyone agrees there
shouldn't be immediate package removal just because the PKGBUILD is
not completely following the standard as well as starts to document
the case for future references, but not now since clearly someone
still think he should delete the package immediately just because he
doesn't want to spend less than one more minute recording it and
coming back on it even one week later.

I don't really think any PKGBUILD should be removed unless it is not
useful anymore (replaced by other packages, drop by upstream) since
AUR is a place to share useful PKGBUILD and let all users to improve
them (instead of examine normal users for their packaging skills as
well as reaction time). I have seen many examples on aur-general (well
2-3 at least) of packages being requested to be removed either because
the package maintainer want to drop them or even don't build but all
end up staying orphan in AUR because they can easily be fixed and may
still be helpful to others.

Another problem that may be caused by this kind of removal is the
PKGBUILD may just be lost. I personally don't keep a copy of some of
the PKGBUILD I am maintaining locally before I started to do that for
all my PKGBUILD's one month ago. I don't think that's a requirement
either. I may be wrong (maybe it is hidden in somewhere in the wiki I
haven't found), but I don't think not keeping a local copy will cause
any problem unless someone decide to pointlessly remove them from AUR.

>>
> The users complaint is valid. Fine, he asked for it to be merged to the
> new package name, and I missed it on aur-general.
>
> How about this, I would like to lodge a complaint against the TU that
> merged the package in the first place without checking for it's validity
> and following of packaging guidelines.

Well, everyone make mistakes, but that should NOT be a excuse to
discourage anyone to upload new PKGBUILD's. A TU may check for errors
on PKGBUILDs before performing any actions but I don't think it is
possible to check for all possible errors especially when there is no
tools to do that automatically (or before pacman start to complain
about it for the package() case). And there is also no reason to
reject a merge if the new one is better than the old one(s). If you
would REALLY like to complaint about this, may I just remind you that
you have at least 5 packages on AUR with missing quote on srcdir, or
is that following your own PKGBUILD standard. I am not saying they
should be removed immediately but somehow I don't think that supports
what you have just said very well.

>
>>
>> As for my own anecdote, you will find many comments from me linking to updated
>> PKGBUILDs that I kindly *recommend* to the maintainer. I wait a few weeks for a
>> reply and then either post again or email the maintainer. I think it is the
>> most effective way encourage users to adhere to packaging standards and adopt
>> better PKGBUILD styles.
>
> YES ABSOLUTELY! I do this, all the time! Including on one of yours
> today! Right after you complained about environment variables not being
> quoted because they might include spaces.

Well, see above and check yourself as well, please....

>
>>
>>
>>
>> Alexander Rødseth wrote:
>>
>> >tl;dr Ioni is always right
>>
>> What alternate reality is this? :P
> I have no comment

Am I supposed to say sth about this too?...

>>
> --
> Daniel Wallace
> Archlinux Trusted User (gtmanfred)
> Georgia Institute of Technology