path: root/meeting-logs/kde-project-meeting-summary-20090521.txt

21.5.2009 - KDE Meeting
Roll-call: wired, alexxy, scarabeus, dagger, hwoarang, tampakrap, bonsaikitten, krytzz, yngwin, civil, papillon81, reavertm, lxnay, cryos, jmbsvicetto

Cryos legitimely excused from not attending much, new baby on the route. Grats to him! :]

Doc handling
- doc == aplication handbook or handbook, to be decided...
- enable +doc/handbook by default for kde-base
- aplication api documentation - scarabeus write mail asking for some global useflag for it on -dev
- rename doc useflag to handbook with 4.3 again :D
- make it more handleable by eclass rather than in the ebuilds
- lxnay volunteer to do the packages update in overlay
- so priority coruse is:
  - mail to dev asking how to do it
  - wait a bit and if nothing constructive comes up do the rename for 4.3

kde3
- kill arts, all misc apps needs it removed and be stabled before 3.5.10 stabling
- stable 3.5.10, all kde related misc apps needs to be revbumped/verbumped and stabled before it
- tampakrap starts handling 3.5.10 stabilisation - stable bug asap 15.6. deadline
- writing doc about kde3/4 mixing - tampakrap

kdeprefix
- long discussion about support of kde4 +kdeprefix install in kde3 and gnome, will chat with reaver about it on aproperiate bug
- add ewarn for user when installing with +kdeprefix so we assure he knows what he is messing with. (controled same like live warning)

kde 4.3
- libknotification already handled, pdepend for kdelibs.
- policykit looks ok
- so no much work for 4.3 itself

kde/kde3 useflag
we voted about it, result:
kde - latest supported kde 4, 5, whatever
kde3 - for now kde 3 series, when kde5 is out there will be kde4 useflag and so on,...
as long as new version is expected to be highly experimental we will have kdeX where X is the version number. When proper support in portage arrives it mutate into kde and older kde mutate to kdeY where Y is X-1

phonon
- ship snapshot into the tree with kde 4.3
- separating xine part to be able use qt-phonon instead of normal phonon with kde - probably reaver

CODE:
improve it, and add requirements we find out that are needed
since next meeting the code will be considered final, and not folowing it will be punished

relwithdebuginfo:
deffered after some work on it, not worth efforts

GUIDE:
tampakrap promised to write the guide for kde3/4 mix that will cover also kde4 installing

kdebindings:
scarabeus + reaver: invent some logic there; delegate work to other HTs

sabayon:
discuss the topic more with joost_op at some other more convinient time
so far done ->
bugs from ppl with @sabayonlinux.org will be handled legitimely as our bugs
and we will reflect them as HTs for kde team so we dont have recheck reported things (aka we trust sabayon devs :})

------------------------------------------------
Qt topics
------------------------------------------------

Rollcall Qt herd
----------------

- hwoarang, tampakrap, yngwin present
- carlo absent
- recruits wired, Pesa and spatz present


Meeting length
--------------

The KDE Project meeting was going on for too long, all agreed. It was suggested
that it would be good to have more frequent and shorter meetings. Also, Qt
topics could be discussed early in the meeting. Details will be discussed at
another time.


Phonon issues
-------------

KDE herd has asked us to introduce a kde useflag for ebuilds depending on::
	|| ( x11-libs/qt-phonon media-sound/phonon )
so that media-sound/phonon can be preferred for KDE users, as a workaround for
current portage shortcomings. yngwin will work out a proposal for affected Qt
ebuilds.

Also, the suggestion was offered to split the backends from phonon into
separate packages (gstreamer from qt-phonon, and xine from kde's phonon), so we
could have one phonon core package (most likely x11-libs/qt-phonon). It was
decided that this is worth looking into.


Recruits
--------

As to the status of recruits for the Qt team: 

- wired is done with the quizzes, so he only needs grilling by a recruiter and
  can be expected to become a full dev within the next few weeks
- Pesa has been very active already in contributing
- spatz just joined us as newest recruit
- sping is busy with GSoC and will continue recruitment process after that, he
  is especially interested in Qt3/KDE3 maintenance


Qt status in tree
-----------------

- bug 266201: 4.5.1 is going stable, but arches are taking their time (only
  alpha and ppc so far)
- bug 270475: bug in the eclass about the platform switch, which affects
  chroots and some arches like ppc, a solution is being worked on in the 
  overlay
- bug 235685: webkit sigbus on sparc, we will try to get upstream to fix it, in
  the meantime we can use a patch on sparc only (and maybe on other arches like
  alpha)
- bug 270769: ppc rendering fix will be fast-tracked for stabilization
- bug 209626: make qt eclasses ready for eclass-manpages, hwoarang offered to
  take care of this
- bug 224951: qt4-qtruby has been hardmasked for a while, so we should fix or
  remove the package; decided we'll work on it to see if it's fixable,
  otherwise ask ruby herd to agree with removal; yngwin will commit his
  intermediate work to overlay
- bug 236341: Pesa and hwoarang will work on removing automagic deps from PyQt4
- bug 43827: qvfb and related embedded ebuilds could be proxy-maintained, we
  will approach users that might be interested


Overlay
-------

As mentioned during the discussion of the KDE overlay's CODE document, we
should start using similar commit policies, especially starting the commit
message with $PN.

Both 4.5.9999 and 4.9999 versions of Qt ebuilds in overlay are actively
maintained and used, so status is OK.

A lot of packages are being moved from overlay to the official tree, so work
there is progressing well. Pesa will work with ali_bush to get the latest
version of qtjambi into the tree.


Eclasses
--------

There has been discussion on -dev ML about blocking mixed Qt versions. Paludis
doesn't handle the blocks and dependencies the same way portage does, but it 
was concluded that the proposal currently in overlay, which works fine with
portage, is the best solution so far, so we will go ahead and implement that in
the official tree as well.

It was also decided to remove the custom-cxxflags useflag and the default
strip-flags from the qt4-build.eclass, as testing has shown that Qt is not as
sensitive to optimized flags anymore. We will let this change coincide with
committing the 4.5.2 (next release) ebuilds into the official tree.

There has been a lot of development going on in the qt4-edge.eclass (the
overlay version of qt4.eclass). We have implemented additional default phases
for src_configure and src_install, the eqmake4 function has seen a lot of
improvements, and there is experimental functionality for handling translation
files.

We decided to add that as a new eclass to the official tree, once the ongoing
work on eqmake4 and translations crystalizes. We will then mark the old eclass
as deprecated and open a tracker bug for migration of packages to the new
eclass. There was some brainstorming about naming the new eclass, but
bikeshedding is to be continued outside of the meeting.


Elected Qt team lead
--------------------

As we are now a fast growing team, yngwin felt the need to bring up the issue
of elections for a Qt team lead, as he assumed the position when no one was
looking after qt herd. The decision was this is not needed, as the team members
are unanimously happy with the current situation of yngwin being the de facto
leader.