Bug#837622: Use and abuse of the unreproducible tag

[Markus Koschany]

On 13.09.2016 18:25, Adam D. Barratt wrote:
> On 2016-09-13 12:55, Sebastiaan Couwenberg wrote:
>> On 09/13/2016 01:49 PM, Santiago Vila wrote:
>>> You can't reproduce it, or you don't want to reproduce it?
>>
>> I added the tag because I couldn't reproduce the issue in unstable where
>> we build our packages. It's great that it's reproducible in testing, but
>> we don't upload to testing.
> 
> Sometimes we do...
> 
> In any case, that's not the point. Testing is what will become stable.
> Packages in stable need to be buildable on stable (for security updates
> and point releases) and "does the package currently build in testing?"
> is the best approximation we can have to "will the package build when
> it's in stable?".
> 
>> I consider the tag appropriate, if there is
>> consensus in the project that it's not feel free to remove the tags
>> again.
> 
> Regardless of whether there's consensus that you agree with, it's an RC
> bug to not build within the same release, and has been for several
> releases now.

Several people are missing the point here in my opinion. The issue here
is that the packages, python-geopandas (#837614) and python-stetl
(#837622) have the exact same version in testing and unstable. The
maintainer rightfully pointed out that this issue might be related to
the current transition of gdal.

There are usually two common reasons why a package FTBFS in testing and
not in unstable. The first one is a missing versioned dependency or
build-dependency, the other one is unrelated to the package itself. It
is affected by issues in other packages.

So instead of playing severity and tag ping-pong, the time would be
better spent into investigating if the gdal transition is causing the
FTBFS in Stretch. Help from the bug submitter would be surely welcome.
Then the bug report should be reassigned to the related transition bug
and both python packages marked as "affected". There is nothing else the
maintainer could have done in this case.

What I find disturbing is that we really think that shaming the
maintainer would be a better approach. Who cares if he uses the
unreproducible tag, if he wants to get feedback about the issue from
others? We should care more about the human cost of this thread.

Regards,

Markus


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 949 bytes
Desc: OpenPGP digital signature
URL: <http://lists.alioth.debian.org/pipermail/pkg-grass-devel/attachments/20160913/89e412a3/attachment-0001.sig>