Re: more binNMUs for rdeps of OpenSceneGraph? -- was: Re: Bug#768199: marked as done (nmu: simgear_3.0.0-6)
2014-11-07 11:03 GMT+00:00 Jonathan Wiltshire <jmw@debian.org>:
> On 2014-11-06 21:40, Manuel A. Fernandez Montecelo wrote:
>>>>
>>>> (You might want to do the other reverse dependencies as well: while
>>>> simgear
>>>> is the only one that we know is affected, it's possible that others are.
>>>
>>>
>>> Can somebody who knows the packages better (i.e. not me) please check and
>>> file additional bugs if they are?
>>>
>>> According to dak the reverse build-dependencies are:
>>>
>>> choreonoid: libopenscenegraph-dev
>>> flightgear: libopenscenegraph-dev (> 3.0.0)
>>> libcitygml: libopenscenegraph-dev (>= 3.0.1-4~)
>>> osgearth: libopenscenegraph-dev (>= 3.0.1) openscenegraph (>= 3.0.1)
>>> ossim: libopenscenegraph-dev
>>> qgis: libopenscenegraph-dev
>>> simgear: libopenscenegraph-dev (> 3.0.0)
>>
>>
>>
>> I haven't looked in enough detail [yet] to see if those rdepends need
>> binNMUing, and I don't know if there's a clear-cut way to determine
>> that (especially if the issue is because of subtle changes of code
>> which rdepends embed, which could be also the case if they use C++
>> templates compiled statically into the rdepend binary, and similar
>> artifacts).
>>
>> Maybe it's better to schedule binNMUs for all, to be better safe than
>> sorry?
>>
>> Some packages for Jessie, like choreonoid and osgearth were uploaded
>> in February and April, so they were not even compiled with GCC 4.9 and
>> the latest libstcd++ for example. They could also be affected by
>> other cases of subtle ABI changes in OSG or other dependencies.
>
>
> Not quite true: chorenoid was binNMUd 20 days ago and 43 days ago, so they
> will not be in the same state as they were at the time of source upload.
Hmmm, sorry for the confusion. I looked at the PTS and changelog
linked from PTS page, I didn't know that the binNMUs do not appear
there. Even with aptitude and apt, changelog doesn't show the NMUs.
But most/all of the rdepends have some +b1 or +b2 appended in the
version available, so yes, they should be somewhat recent.
> Looking at the relevant dates though, nothing will have picked up the fix
> for #765855; scheduling. Don't worry about filing bugs.
OK, thanks.
--
Manuel A. Fernandez Montecelo <manuel.montezelo@gmail.com>
Reply to: