KDE Commit-Digest for 17th December 2006

In this week's KDE Commit-Digest: A new interface element, the 'viewbar', makes its debut; continued work on Flake and Kross-based scripting within KOffice. Continued refinements in KSysGuard. Much work to improve support for VPN connections in KNetworkManager, with KNetworkManager being moved from playground/ to extragear/. Kaffeine begins porting to GStreamer 0.10. KGeography extends its global coverage with a handful of new country maps. KWin4 and Kolf begin their transition towards improved and scalable (SVG) graphics. Commits start to flow in the Student Mentoring program. Support for bullet-aliased passwords across KDE.


Looks kool.

By gerd at Sun, 2006/12/17 - 6:00am

Defintely. Hope it will make its way in kdelibs soon.

By AC at Sun, 2006/12/17 - 6:00am

Hey Casper, this is great!
Now we're this close to having true Adobe style virtual sliders for any type of parameter in KDE.

Go! Go! Go!

By Gabriel Gazzán at Tue, 2006/12/19 - 6:00am

There are 2 bug killers named Sebastian Trueg. Cloning is definitively the solution for the bug squashing team.
Light sabers are too old school..

By shamaz at Sun, 2006/12/17 - 6:00am

Heh, I investigated, and it seems Sebastian Trueg (or himself and a crazed fan ;)) is operating from two bugzilla accounts this week: trueg k3b org and sebastian trueg de



By Danny Allen at Sun, 2006/12/17 - 6:00am

Not to critize, but isn't the whole point of phonon to support apps like kaffeine? Shouldn't that effort be spent on moving to phonon?

By AC at Sun, 2006/12/17 - 6:00am

I totally agree with you, when i read this i immediatly said "WTF!? What about phonon then!? It is going to be the same sh**t as arts and that's why kaffeine is not going to be ported to it?!"

By erne.castro at Sun, 2006/12/17 - 6:00am

phonon is completely unlike arts in every way since they are completely different types of solutions.

to application developers they look similar: the provide multimedia APIs. from there things differ dramatically.

phonon provides UI elements and is much more "KDE-esque" in API design. however, it provides approximately zero actual multimedia playing/catpuring functionality internally; for that it relies on things like NMM or gstreamer or whatever Phonon backend you're running talks to.

aRts, on the other hand, was the sort of animal you'd write a Phonon back end for. it was also less KDE-like in its API and didn't provide much in the way of UI help.

By Aaron J. Seigo at Mon, 2006/12/18 - 6:00am

It think it was more about: why gstreamer backend in Kaffeine when gstreamer should be addressed from Kaffeine via the phonon layer.

that is:
Kaffeine --> phonon -> gstreamer, nmm, ...

the story is read as
Kaffeine --> (phonon -> gstreamer, nmm, ...),

By fred at Tue, 2006/12/19 - 6:00am

I don't think Kaffeine is part of kdemultimedia, so they can do whatever they want really. Kaffeine with a regular gstreamer backend might be good so that GNOME users use it without needing more than kdelibs and Qt.

By matt at Fri, 2006/12/22 - 6:00am

> Kaffeine with a regular gstreamer backend might be good so that GNOME users use it without needing more than kdelibs and Qt.

Since Phonon is part of kdelibs, there will be no difference dependency-wise

By Kevin Krammer at Fri, 2006/12/22 - 6:00am

You forget that kaffeine has a release cycle different from that of KDE.
You want to wait over a year for kaffeine to support the newer gstreamer?
Porting to phonon would also mean porting to KDE4/Qt4.
Also I don't think that phonon's API is completely finalized nor is it ready for apps to start their porting, Of course someone please correct me if I'm wrong.

By Stephen at Sun, 2006/12/17 - 6:00am

You are wrong :)

Seriously; Phonon is one of the first components in KDELibs that was deemed stable various months ago. Using the framework is very much appreciated. If you find something missing Matthias told us he is very much looking for constructive critisism.

So, in effect, I think that starting work on a KDE4 based version should happen rather sooner than later. As 'sooner' will allow you to actually give feedback if something in the libs API is not perfect.

Also; I was under the impression that Kaffeine didn't release that often. The latest (according to freshmeat) is already some 7 months ago.
In that light, I'd say they devs should finish up the KDE3 version; put it out the door as soon as possible, and focus on the KDE4 one.

But thats just my opinion :)

By Thomas Zander at Sun, 2006/12/17 - 6:00am

mine, too! get Kaffeine the perfect player it always wanted to be...

By eMPe at Sun, 2006/12/17 - 6:00am

Kaffeine is mostly a Xine frontend.

I don't see any problem in porting kaffeine to phonon after kde4 is released and proven.

By otherAC at Sun, 2006/12/17 - 6:00am

one catch: it's hard to prove a library without apps using them.

4.0 will also be less fun to use if there are fewer apps, meaning fewer people are likely to spend time with 4.0 apps and features .. meaning it's harder to prove.

i sounded the horn about a year ago that apps should probably do one or two more kde3 releases and then look to do a kde4 release at that point... so i understand the ideas behind the conservative approach; but we're pretty much at the point that working on your kde4 app makes sense. =)

By Aaron Seigo at Mon, 2006/12/18 - 6:00am

> but we're pretty much at the point that working on your kde4 app makes sense. =)

Are we? For what length of application release cycles? I can tell you that I won't start porting my apps until there is a clear message, that KDE 4 will come out in 3 months or so. For others the threshold will be different, depending on how often they usually release new versions. But I seriously doubt, too many outside-of-KDE's-release-cycle-apps will start getting ported as long as there isn't even a tentative release schedule for KDE4.

Of course we don't want to shut off KDE4 development too early, and probably it would be highly unlikely for a release schedule to actually be met on the spot. Still, if at least there were some goal posts such as "API changes (not additions) in kdelibs should be finished around MM/DD/YYYY. This deadline may be pushed ahead by two weeks at a time, as long as a majority on kde-core-devel thinks more changes are needed. First beta release should be prepared around MM/DD/YYYY"), etc. Then application developers would at least have something to base their guesses on.

By tfry at Mon, 2006/12/18 - 6:00am

3 months to port and adapt. i suppose either your apps are simple, you have a good amount of time to work on them or you're just a really fantastic developer. =)

and yes, the threshold will vary from person to person. but for any non-trivial app i think now is a good time to get going on kde4. you'll want time to adapt to qt4 and the new stuff in kde4.

it will also help us test API changes, porting tools, etc. IOW, it will help get KDE closer to the 4.0 mark.

and yes, i agree that we need a clear release schedule at this point. the time has come for that too.

By Aaron J. Seigo at Mon, 2006/12/18 - 6:00am

>>Also; I was under the impression that Kaffeine didn't release that often. The latest (according to freshmeat) is already some 7 months ago.

According to[1] last update was november 26
and according to the website of kaffeine[2], the release before november was september 08.

I think you should use more reliable sources :o)


By otherAC at Sun, 2006/12/17 - 6:00am

> You want to wait over a year for kaffeine to support the newer gstreamer?

personally, i think that one more kde3 release would make sense for an app like kaffeine and then a start on the kde4 port .. it won't be over a year before a 4.0 release is made.

By Aaron Seigo at Mon, 2006/12/18 - 6:00am

only if phonon can offer verything that kaffeine needs.
dunno if that is the case?

By otherAC at Mon, 2006/12/18 - 6:00am

given what kaffeine does and what was in the phonon demo at aKademy (e.g. a video playing with audio controlled by a seek and volume slider ;) i'd say "yep".

i understand the backends need some more work, but i'm guessing that if apps started using it more and dropped their own backend efforts to coordinate on one set of phonon backends that would help things a lot.

By Aaron Seigo at Mon, 2006/12/18 - 6:00am

> given what kaffeine does and what was in the phonon demo at aKademy (e.g. a > video playing with audio controlled by a seek and volume slider ;) i'd
> say "yep".

No, this is only a Kaffeine side-effect. Kaffeine is meant to be a front-end to the zillion xine options. Media playback is only one of these options.

I seriously hope that Codeine is ported to Phonon, personally I do not much care about Kaffeine.

By Davide Ferrari at Mon, 2006/12/18 - 6:00am

if phonon can't support what kaffeine does (and i'm a daily user of kaffeine; i've even contributed patches to it), then we need to know *now* not after 4.0 comes out.

kde4 media apps really shouldn't need to write to a given backend like we've done in the past. it's fragile, it's time consuming and isn't portable.

By Aaron J. Seigo at Mon, 2006/12/18 - 6:00am

I did that gst-0.10 port to see it in action.
Was not so much efforts (a few hours) to see that it's still far from xine.

Moving to Phonon is in our todo, off course. But at that moment we are concentrating on a thread-safe solution for xine's X vo plugins (xcb seems to be the way to go).

But after having given kde4 a try, i do not expect it to be released soon, so we should have some month for the kde4 port :)
And btw, atm #phonon has nearly zero traffic :)

Oh, and yes Aaron, Kaffeine does a little more than playing a video file with sound control and seek bar. For example, get a dvb-t stick (or more), plug it in and start Kaffeine ;)

By Christophe Thommeret at Mon, 2006/12/18 - 6:00am

i'm aware of the various things kaffeine does. it even plays cd's. whether that's a good thing or not is another question ;) but there it is: lots of things it can do and the developers like it that way. cool.

if there are things that phonon isn't covering for kaffeine, we need to know sooner rather than later so it can be determined if there is a good way to provide support for that (or not)

i really like how people are pretending to kick the tires, but are doing so without so much as really looking at them let alone hitting them with their shoes. ;)

By Aaron J. Seigo at Mon, 2006/12/18 - 6:00am

>we need to know sooner rather than later

I, of course, understand that, just let us some time to look at phonon (and kde/qt4). I've just compiled kde4 to give Juk/Phonon a try but unfortunately Juk failed to compiled, so rather than spending time on that i searched the web for a Phonon test app but had no luck. And zero traffic on #phonon didn't help that much. So will have to do it myself. np. Just need time.

And btw, our work on xine' X video out plugins should benefit phonon-xine as well.

Sorry but with my poor english, i don't understand that tires/shoes "métaphore". ?

By Christophe Thommeret at Mon, 2006/12/18 - 6:00am

Experienced drivers can gauge the pressure by kicking the tires and feeling how hard or soft they feel.

This metaphor has been in use since long to describe when one tries some device for the first time.

By Gr8teful at Tue, 2006/12/19 - 6:00am

First I don't think that you can compete with kaffeine in the simple-video-player market. The key features are imho morely stuff like watching dvds and dvb.
Of course we don't want to lose basic capabilities and there will be a kde4 kaffeine with support for phonon ... but this doesn't mean that we can drop our current backends.
It doesn't seem like phonon supports dvd menus or transparent osd (nearly needed for dvb) in the near future (i may be wrong) - furthermore we want to provide continuous releases to the systems our users have at the moment. I think this is also part of the reason we don't live in kde main but in extragear ...
Currently some fresh air has come and development will intensify, so expect good stuff soon ;-)
I hope this could clearify the situation a bit ...

By Christoph Pfister at Mon, 2006/12/18 - 6:00am

so one of the kaffeine backends would be a front end to backends? *heads spins* that seems a little ...... inefficient, from both a coding and a performance stand point.

what would be great is if the work everyone puts into perfecting their backends goes into perfecting the backends for phonon. now, the question is if that is realistically achievable.

you mention transparent OSDs. i don't see what that has to do with phonon at all? this sounds more like a job for a nice OSD class in kdelibs, unless i'm misunderstanding what you're driving at.

dvd menus sounds like something that the video widgets should support out of the box where available (e.g. not every phonon backend may be able to support that).

as for continuous release, yes, of course that makes sense. the question is when one can/should switch over to kde4 development so that users aren't stuck halfway between kde4 and kde3 due to your app. part of this is that the kde project needs to put out some better timelines for app devels such as yourself to sync with, but getting app developers to work on kde4 apps as we edge closer to a 4.0 will actually help us get 4.0 out sooner with greater confidence.

imho it would absolutely suck if 4.0 released with a phonon that couldn't do what kaffeine needs simply because the kaffeine developers didn't engage with phonon until it was too late. remember, kde is your platform, too, simply by nature of your app using it. just because your app is a 3rd party app doesn't mean you have no ability (or even to some degree responsibility) to improve or influence the parts of the project that you rely on.

By Aaron J. Seigo at Tue, 2006/12/19 - 6:00am

After reading your KDE4 porting arguments I realize some of these apply to my application as well. I'm working on KMess (, a MSN Messenger client for KDE.

Over a half year or so we'd like to implement multimedia stuff. That would involve Solid to access webcams, Phonon to handle audio streams in both directions, and perhaps Decibel as well. I'd love to leave KDE3 with another good release, but implementing these features without Solid/Phonon seams like a waste of time too. A release schedulle or SVN status page would help a lot.

This still leaves me with the question whether KMess should be ported to KDE4, or try to implement webcam/audio for a final KDE3 release.. :-/

By Diederik at Tue, 2006/12/19 - 6:00am

I think it's time for a clear roadmap for kDE4, so that developers can decide whether it's a good idea to release new stuff in kde 3.x series or port the application to kde4.

By otherAC at Tue, 2006/12/19 - 6:00am

John Tapsell committed a change to /trunk/KDE/kdebase/workspace/ksysguard/gui/SensorDisplayLib/

Show the CPU usage as just "32%" "5%" and "0.4%", so to just 1 or 2 s.f.
As suggested by a "ben" on

Woohoo! Thanks a lot for listening, John!

What does s.f. mean, "significant [f,d]igits"?

By ben at Sun, 2006/12/17 - 6:00am

Significant Figures (might be more of a British thing ;))


By Danny Allen at Sun, 2006/12/17 - 6:00am

"Significant figures" I presume.

By Robert Knight at Sun, 2006/12/17 - 6:00am

By John Tapsell at Mon, 2006/12/18 - 6:00am

Nice! Those little cpu/mem usage bars look really nice! Makes it much easier to see what is hogging all your cpu or memory at a glance.

By Leo S at Mon, 2006/12/18 - 6:00am

Good stuff, John. Way to go :)

By Mark Kretschmann at Mon, 2006/12/18 - 6:00am

I noticed that guidance and knetworkmanager didn't use Solid for power and network management. What's up with that?

By Hagai at Mon, 2006/12/18 - 6:00am

I quess guidance and knetworkmanager is for kde 3.5.x and you can't use Solid before kde 4.

By kanttu at Mon, 2006/12/18 - 6:00am

guidance simply hasn't been ported to KDE 4 yet.


By Simon Edwards at Mon, 2006/12/18 - 6:00am

SVG as we know is supposed to be the solution for the problem of keeping multiple icons of varying sizes for the same widget. But even SVG has issues looking nice when they are reduced to 22x22 or even 32x32.

I was thinking about CorelDraw this week-end and recalled it had the capability of mixing SVG and BMPs in the same drawing. What if SVGs degraded to BMP at small sizes?

By Peter at Mon, 2006/12/18 - 6:00am

In terms of icons, it is a given that more simple and recognisable SVGs/BMPs be made for smaller sizes. At least thats what is done with Tango.

I have not heard about storing the smallscale BMPs as part of the SVGs however. Will need to check the SVG spec about that.

By Schalken at Tue, 2006/12/19 - 6:00am

Not sure how long ago the entry was (not thaaat long ago my fuzzy ol memory recalls) but Zack Rusin blogged ( about this and proposed a pretty clever solution that mimicked some of the techniques used in keeping fonts readable at different sizes (not a GFX guy, won't try and go into detail about what he said for fear of mucking it up)

By borker at Tue, 2006/12/19 - 6:00am