when KDE SC 4.6.3 migrates to testing, I plan to switch away from monolithic 
kdebindings to the split source packages. It is no secret that kdebindings has 
always been a huge pain for many people (esp. qt/kde team and release team). 
It links many languages complicating and tying various transitions together. 
In the past, it stretched KDE transition to ~2 months. Not to mention 
complexity in packaging multiple programming languages with many different 
packaging practises. It's pretty clear that the kdebindings source package (in 
its current form) has to go away for the benefit and sanity of everybody 

Fortunately, there has been some positive movement towards split source 
packages upstream. While split kdebindings won't officially arrive until 4.7, 
I think it should be quite feasable to build 4.7 Beta tarballs (or previous 
git snapshots) against our 4.6.x. I had quite much success with PyKDE4 [1]. 
Packaging was much simplified in the process. It should no longer kill 

That being said, I personally have no (even remote) intention to maintain 
other bindings except Python (that is python-kde4 + krosspython). To be 
honest, the only reason I care is a huge number of rdepends inside and outside 
KDE SC and popularity (importance) of Python bindings in general.

So that leaves Mono, Smoke and Ruby bindings up in the air. They need 
_dedicated_ maintainers with upload power (DD or DM) who would package 
standalone version, _fix bugs_ even when they're nasty (be prepared) and keep 
bindings updated in case new KDE SC release / language related transition 
broke them. Packages can be maintained within or outside Qt/KDE team as 
maintainer(s) prefer.

Otherwise, I will seek those bindings to be removed from the archive once I 
upload standalone PyKDE4 bindings:

* Mono bindings have no reverse dependencies. No problem here.

* Ruby bindings have a single reverse dependency - omaque [2] (maintainer 
CC'ed). In my opinion, this leaf package could be removed from testing/archive 
until somebody steps up to maintain Ruby bindings. omaque popcon is very low.

* Smoke bindings have a single reverse dependency - qt4-perl [3]. This one is 
a bit more complicated as debconf build-depends on the latter. Fortunately, 
that build dependency seems to be redundant as debconf builds fine without it 
[4]. Btw, I'm fully aware that removal of qt4-perl would kill Qt 4 Debconf 
frontend. People, who don't want that, just have to repackage and commit to 
maintaining Smoke bindings.

I'm not sure about the timeline yet (but definitely after KDE SC 4.6.3 
migrates to testing). I think Python stuff could really be uploaded soon as 
it's mostly ready. However, uploads could be delayed if Mono, Ruby or Smoke 
needed more time to be prepared.

No new unstable uploads of monolithic kdebindings is planned.

[1] http://anonscm.debian.org/gitweb/?p=pkg-kde/kde-sc/pykde4.git;a=summary
[2] http://packages.qa.debian.org/o/omaque.html
[3] http://packages.qa.debian.org/q/qt4-perl.html
[4] http://bugs.debian.org/629405

Modestas Vainius <mo...@debian.org>

Attachment: signature.asc
Description: This is a digitally signed message part.


Reply via email to