On 09/28/2013 12:48 PM, Alexandre Moine wrote:
> Le 27/09/2013 22:16, Alec Leamas a écrit :
>> On 09/27/2013 10:02 PM, Richard Shaw wrote:
>>> On Fri, Sep 27, 2013 at 1:23 PM, Alexandre Moine
<nobrakal(a)fedoraproject.org <mailto:nobrakal@fedoraproject.org>> wrote:
>>>
>>> Hi all,
>>>
>>> I'm a new packager in the RpmFusion Project, and in the Fedora
Project
>>> in general.
>>> I submitted a review request of the openmw package [1] [2]. And I've
a
>>> problem. Openmw include 2 bundled package. So, I apply to make two
>>> exceptions. This is the reason:
>>>
>>> For shiny:
>>> Per upstream it is designed to be copied in and is not a stand alone
>>> library.
>>>
https://github.com/scrawl/shiny/blob/master/CMakeLists.txt
>>>
>>> For oics:
>>> Modified from upstream source.
>>>
>>> You can also read our discussion in the openmw forum [3].
>>>
>>> Links:
>>> [1]:
https://bugzilla.rpmfusion.org/show_bug.cgi?id=2921
>>> [2]:
https://openmw.org/en/
>>> [3]:
http://forum.openmw.org/viewtopic.php?f=3&t=1671&start=30
>>>
>>> Please, if you have any question, ask me :)
>>>
>>>
>>> Of course you get my vote, but I'm biased. :)
>>>
>>> Richard
>>
>> To be a little formal, you are required to provide answers to the standard
questions in [1] before there is a decision. If you try to answer those, things will be
much easier to sort out.
>
> Oh sorry, I forgot that:
> Has the library behaviour been modified? No for shiny, it is developped by a member
of the openmw project. Yes for Oics. It is modified to work with sdl.
> Why haven't the changes been pushed to the upstream library? I don't know...
But the team has been heavy modified the code.
> Could we make the forked version the canonical version within Fedora? No, oics is
designed to work with openmw. For shiny, it can't to work alone
> Are the changes useful to consumers other than the bundling application? No, for the
same. Oics is designed for openmw. For shiny, I don't know, but I think not. It's
really designed for openmw.
> What is the attitude of upstream towards bundling? The writer of shiny is very
sympatic (@scrawl). For the writer of OICS, i don't know.
> Overview of the security ramifications of bundling.The sources is sure (it's
remake by developpers of openmw, not dangerous)
> Does the maintainer of the Fedora package of the library being bundled have any
comments about this? The package does not exist.
> Is there a plan for unbundling the library at a later time? For now, no. It's
difficult to do this, and the openmw have other fish to fry! But, with the time, maybe.
> Please include any relevant documentation:
>
> shiny project:
https://github.com/scrawl/shiny
> OICS:
http://sourceforge.net/projects/oics/
>
> The discussion about this on the openmw wiki:
http://forum.openmw.org/viewtopic.php?f=3&t=1671&start=30
>
> Thank you very much Richard ;)
>
> Alexandre
>
>>
>> --alec
>>
>> PS Yes, its a pain... been there, done that. ;) DS
>>
>> [1]
https://fedoraproject.org/wiki/Packaging:No_Bundled_Libraries#Standard_qu...
>>
>
>
> --
> Alexandre, Fedora User and Ambassador
Hm, lets try to split this discussion into two: one for shiny and one for OICS.
For shiny: is there a problem to package this separately? The code is unmodified, so you
could just unbundle a new shiny package, and make sure it works with openmw? There is
something I don't really get here.
It might be that the shiny package just is some source code and not actually linked
against. But that does preclude making a separate package IMHO. However, we need more
input on this - I',m on thin ice here.
Shiny clearly falls under the copylib bundling exception, so I see no problem with
bundling it.
As for oics given the heavy modifications done to it I'm fine with granting a
bundling exception there too.
Regards,
Hans