[Kde-scm-interest] Splitting kdebindings when moving to git

Matt Williams lists at milliams.com
Thu Aug 19 18:00:32 CEST 2010


On 19 August 2010 16:55, Richard Dale <richard.dale at telefonica.net> wrote:
> On Thursday, August 19, 2010 04:42:42 pm Thomas Zander wrote:
>> On Thursday 19. August 2010 17.02.59 Arno Rehn wrote:
>> > With KDE bindings we have the problem that many people only want the Qt-
>> > specific part of the bindings (like QtRuby or Qyoto) - and then only for
>> > one  language. Having all of the different bindings in a KDE module makes
>> > it difficult for them to install only the Qt portion and even prevents
>> > some people from contributing (there's already a fork of QtRuby on
>> > github, because people don't like checking out the complete kdebindings
>> > and then going through the hassle of figuring out how to compile only
>> > QtRuby).
>>
>> Is this really *the* problem?
>> With a total checkout size of 116Mb I'm thinking the git repo is not really
>> that big. Maybe you can give us an exact number of the unsplitted size.
> That isn't the issue.
>
>> Figuring out which one to compile sounds even stranger; the detection is
>> pretty advanced and it won't even try to compile the parts you don't
>> touch.
> Well qt-only guys like the MeeGo people don't wan't the compilation of the qt-
> only and tools parts of kdebindings depending on advanced and complex cmake
> macros.

I would say that in this situation, they're a good candidate for
splitting out into a separate module called kdebindings-support or
even moving into kdesupport itself. This is, however, completely
orthogonal to the Git migration.

>> The reason for keeping it together is because the KDE team intends to
>> release the bindings as a 'kdebindings' for each KDE release. One git repo
>> then equals one tarball that is released. I think Thiago gave a nice
>> overview of the reasons behind that; which I won't repeat here as this is
>> a discussion that should not be reopened again without really good reasons
>> :-)
> I wasn't involved in this discussion.
>
>> As Aaron wrote some time ago; "Its a discussion we had and closed, its
>> decided"
> ?
>
>> The reasons you cite are not entirely convincing to me, at this late stage,
>> to warrent a change in strategy.
> A change in strategy in general, or a change in strategy for kdebindings?
>
> -- Richard

-- 
Matt Williams
http://milliams.com


More information about the Kde-scm-interest mailing list