Is there any issue accepting several versions at once (Re: [ubuntu/precise] gnome-control-center 1:3.3.90-0ubuntu3 (Rejected))

[Micah Gersten]

On 02/27/2012 10:19 AM, Scott Kitterman wrote:
>
> Sebastien Bacher <seb128 at ubuntu.com> wrote:
>
>> Le 25/02/2012 17:15, Ubuntu Installer a écrit :
>>> Rejected:
>>> Rejected by archive administrator.
>>>
>>> gnome-control-center (1:3.3.90-0ubuntu3) precise; urgency=low
>>>
>>>    * debian/patches/62_update_translations_template.patch:
>>>      - update POTFILES.in with the ubuntu specific sources for
>> translation
>> Hey,
>>
>> Dunno who rejected that upload so comment on the list, is there an
>> issue 
>> to accept several versions at once? What happened there is that
>> somebody 
>> else uploaded a new revision while mine was in the queue, so I guess 
>> whoever dealt with those accepted the new one and rejected the previous
>>
>> one, which is fine but:
>>
>> - the change I did was never advertised on -changes (which means people
>>
>> who use the list to track what is changing get confused)
>> - if there is any bug that should have been closed with that upload if 
>> wouldn't have been closed
>> - the uploader gets a rejection mail with explanation which is
>> confusing
>>
>> We could say that uploaders should check the unapproved queue before 
>> uploading a new revision of something in the queue and use -v, but 
>> wouldn't it be simpler to just accept both versions and let launchpad busy and
>> deal with the builds?
> I think that was me.
>
> IIRC the buildd's were busy and it seemed wasteful. The lower revision would probably end up failed to upload. 
>
> While I don't know of specific bugs, I think inviting Soyuz to get in a race with itself is generally not prudent.
>
> Scott K
>
If it's the same source package, if the later one is accepted first, the
earlier one will fail to build for superseded source (i.e. never
start).  Still, I think in this case either both should be allowed in
descending order, or both rejected so that someone can do a proper
upload showing all the changes.  There's also the worry that changes can
be missed in cases like this as like what happened with apt-cacher-ng
[1][2][3]  It seems like rejecting both with a message telling the
uploaders to talk to each other seems most prudent unless it's a
critical fix.

Thanks,
Micah

[1]
https://lists.ubuntu.com/archives/precise-changes/2012-February/010766.html
[2]
https://lists.ubuntu.com/archives/precise-changes/2012-February/010772.html
[3]
https://lists.ubuntu.com/archives/precise-changes/2012-February/010805.html