Get in contact with KDE 4, now…

KDE 4 is bringing a lot of new or improved frameworks with it, on all levels. CMake, Qt4, Phonon, Solid, all the other one in kdelibs and kdepimlibs… Although the first developer snapshot was already released a few days ago, some new frameworks have yet to settle out of the vapour. πŸ™‚

Just make use of this fact. Everyone has a chance to adapt her code to all the new things in smaller steps instead of one giant one. I do so these days by porting the vapour framework of mine, previously called contacts framework. I took some hours this weekend to continue, spread the old code over all places that looked suiting, did some porting efforts and finally reached this far:
Former contactcards are already working a little.

Oh, what luck there is in seeing your baby live and grow. Even if it mostly cries currently. But this should improve πŸ™‚

This code is only on my harddisk for now. I hope I will soon be able to expand the framework down the stack to meet the successor of KABC and Akonadi, so it makes sense to import it into kdepimlibs. Will discuss this at least at Akademy, if not earlier.

Mobile phones getting more interesting…

For now I managed to get around my own location tracking and disturbing facility. Especially, as I do not favour to put my personal and private data like contacts and dates into control of a system I cannot control. But the number of occasions where a mobile phone would have been handy is rising. And now:

The Greenphone is expected to ship in September.

CTO Benoit Schilling added, “I’ll tell you a secret. Getting the phone into open source developers’ hands is exactly what I want to happen.”

So… Open source developer will meet in September. At Akademy, in Ireland. And Ireland is known as green island. Why do I expect now to see some Greenphones there? πŸ˜›

Met to hack… and met the heck.

Yesterday tokoe and Josef Spillner came over to my place to do some collective KDE hacking for a day. There is nothing better to have now and then some knowledgeable people right next to you to pose your questions to. πŸ™‚

I personally missed close to all my targets for that day, though. I stumbled across a strange error in the build system, which causes the omit of the creation of moc files in some modules like kdepimlibs. Not good, isn’t it? I spend half of the day trying to find the reason, but without success. The bug is now waiting on the kde-buildsystem mailing list for someone with more cmake insight to take care of it. Interested to find out why an AND conjunction of two expressions with result TRUE sometimes results in FALSE? Please go right there and test your skills!
Update: After some more research I now found the bug, the problem is solved. CMakeLists.txt contained lines like macro_bool_to_01(LDAP_FOUND HAVE_LDAP 1), but 1 makes no sense there and seems to make cmake go crazy. Changing it to macro_bool_to_01(LDAP_FOUND HAVE_LDAP) gives a working build again, finally πŸ™‚

So no work for support for the GGZ Gaming Zone into the contacts framework. No work on extending the framework for services operating on groups. And no initial/experimental work for contacts in Akonadi. Instead I started to simply port the contacts framework to KDE4, so the service framework is at least ready for integration with others.

At least tokoe and Josef had more success with their plans and did some development on Akonadi and Kung. πŸ™‚

When in the evening Josef started to investigate where the city council of Dresden is spending our money (a 3D model of Dresden) and watched the sample video it was obvious that finally our hacking power was spent for the day. πŸ™‚

As a last task Josef teached us about all the possibilities how to adjust your personal KDE commit messages filter.

Tokoe proposed to do such meetings monthly or so, at changing places, an idea which was welcomed by Josef and me. So if you live in Dresden (or around, like Jena and Berlin πŸ˜‰ ) contact us for setting up further hacking meetings! Next one might be in the middle of september, to prepare for Akademy.

Web Services…

Doing something like a service oriented architecture one cannot ignore the current phenomenon called Web Services. Old ideas refreshed with even bigger protocol overhead, some say.

Webservices are of interest for the contact framework, too. Want the next connection via public transport to a friend? Locate a place in Google Maps or get a route from yours to it? Just get the webservice description in the Web Services Description Language and do your client code.

You surely do not want to mess around with this manually. Thanks to Cornelius Schumacher, Tobias KΓΆnig, Josef Spillner and others KDE4 will hopefully come with some tools for you. πŸ™‚

Until then one could also (ab-)use parameterized urls or fill forms by script. Like for a very useful (or not? πŸ˜‰ ) service to query Wikipedia for info about a date/day, as below. Seems Matthias Ettrich’s initial post to create a Kool Desktop Environment does not make up too much for an entry in history yet.

What happened else that day?

For those websites where one cannot put the data in the url (forms with Post, not Get) I have a problem right now. Via DCOP I did not manage to have Konqueror open an url, then executing a script to fill the form and to post it (using evalJS). I simply do not know when the page is loaded, so the script can be executed, as there seems to be no proper signal emitted. What a pity, this one would have been really handy for the connection web formular of the local public transport, which I tend to use quite often if biking is no option (a car is comparatively of low use within Dresden).
And embedding in an own frame to do cross-scripting onLoad does not work, too (thank you phishers, bah). Copying the whole form into an own page feels like loosing, too much work on the client side (or for the fun of it currently πŸ˜‰ ).

Seems I have to wait for KDE 4 bringing Kung & Co. Well, I hope a prototype service connector works in time for the rainy and cold season or I have to stay at home πŸ™‚