KDE/kdelibs

Volker Krause vkrause at kde.org
Fri Nov 27 09:58:55 CET 2009


On Thursday 26 November 2009 18:01:31 Alexander Neundorf wrote:
> On Thursday 26 November 2009, Volker Krause wrote:
> > On Thursday 26 November 2009 00:47:33 Alexander Neundorf wrote:
> > > SVN commit 1054396 by neundorf:
> > >
> > > Revert the new cmake module committed without any review or
> > > announcement and which was in a broken state.
> > >
> > > This has the effect that the condition for building nepomuk is now
> > > never true (instead of sometimes),
> >
> > great, you effectively disabled half of KDE PIM now, as well as
> > everything else depending on Nepomuk. I'm sure everyone working on
> > Nepomuk integration appreciates that as much as I do :-(
>
> The build of several KDE modules was broken since the nepomuk query was
> added to kdelibs last week. Nepomuk was considered optional in kdelibs, but
> other modules relied on it without testing. And it took me days of getting
> every small piece of information of what is necessary when and what is
> optional. Yesterday I finally got that part (raptor & redland) fixed.
>
> Now there was the next commit again breaking the build, and not respecting
> the policies for kdelibs.
>
> I mean, I don't really have the most exciting "job" in KDE. I'm mostly
> watching that stuff doesn't get broken. To help me doing this we have a
> commit policy for kdelibs/cmake/modules/.
> It was announced several times, and this time I just want that it is
> followed. Which means to post the file for review to kde-buildsystem (or
> k-c-d) and after that it can be added.

I'm not disagreeing with the policies, or suggesting we should ignore them, 
quite the contrary. Also, I really appreciate the work you are doing on the 
buildsystem.

> > Finding and installing the new dependency took me two minutes, noticing
> > that the cmake check doesn't actually search for it, despite the message
> > shown, easily five times as long...
> >
> > Of course the cmake check could have used a review (it at least misses
> > Windows support) and I'm certainly not happy about that, but making
> > things worse is not helping either.
>
> It made the breakage which was there obvious, and increases the pressure to
> fix it.

Right. The only problem I have with this is that it hit a number of innocent 
people (not including me here, I was one of those asking for a solution for 
the ontology duplication for quite some time). For example this caused quite 
some support load on IRC. Not the most exiting job either, as you can 
imagine, so we should try to avoid this kind of needlessly wasting everybodys 
time.

regards
Volker
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: This is a digitally signed message part.
Url : http://mail.kde.org/pipermail/kde-buildsystem/attachments/20091127/63273a97/attachment.sig 


More information about the Kde-buildsystem mailing list