KDE/kdevplatform [POSSIBLY UNSAFE]

[Matt Rogers]

On Friday 27 June 2008 09:56:00 Andreas Pakulat wrote:
> On 27.06.08 08:38:44, Matt Rogers wrote:
> > On Friday 27 June 2008 02:03:14 Andreas Pakulat wrote:
> > > On 26.06.08 18:42:54, Matt Rogers wrote:
> > > > On Thu, Jun 26, 2008 at 10:00:00AM +0000, Manuel Breugelmans wrote:
> > > > > SVN commit 824598 by mbreugel:
> > > > >
> > > > > Move xtest to kdevplatform & rename.
> > > >
> > > > I don't think that this makes sense. How is a unit testing library
> > > > a general plugin to have for more than just kdevelop? Why can't we
> > > > move this to the platform when another application actually needs
> > > > to use it instead of just assuming there will be another app that
> > > > will actually use it?
> > >
> > > You're misunderstanding. What is now in kdevplatform (or rather what
> > > should be there, I didn't have time yet to look into it closely) is a
> > > support library to write plugins for kdevelop that support running
> > > testcases for a specific unit testing framework. It aims in the same
> > > direction as eclipse TPTP (just not so extensive currently). Its AFAIK
> > > currently mostly the duplicated code from the 3 unit-test-framework
> > > implementations we have in kdevelop.
> >
> > Thanks for clarifying that for me. I was indeed misunderstanding.
> >
> > Even though I was misunderstanding, are we ready for this in kdevplatform
> > now? If we released tomorrow, would we have issues with BC, for example?
> > I know that we're not releasing tomorrow, but that's the mindset we
> > should attempt to take with things like this, IMHO.
>
> IMHO you're wrong there:
>
> a) we only need to care about that once we start the release cycle -
> i.e. probably with the first Beta no BC changes. There are BiC changes
> going on all over the place in platform these days.
>
> b) we agreed that we don't promise BC of platform between KDevelop4.0
> and 4.1, i.e. we allow the platform libs to mature within our first
> minor release.

I didn't remember this part, so that does indeed make my argument invalid.

Thanks
--
Matt