23. August 2006
Until today, I had heared about m4 only that it is one of the so many special Unix tools of which I will never now exactly how they work or at least not completly. Unfortunately gtkmm uses a lot of m4 to generate the C++ code from it’s .hg and .ccg input files and I had to change some things to allow building without Atk.
So I found out that m4 is not that damn complicated and could after some manual reading easily archive the things that I wanted. Anyway, I also changed some code in *.pm files without knowing what language they use in detail. Seems like a quite easy scripting language but would be cool if someone could tell me.
23. August 2006
The –disable-api-atk in gtkmm is much more difficult than I thought. I have to hack a lot on the code generation which means I first have to learn how all this works. I uses m4 and scripting and I have to find out in detail how to add #ifdefs to the signals that use ATK.
16. August 2006
Renamed the hidonmm packages to the correct names as they were filed in the wrong folder of the repositry. I wonder if there is a way to delete the old ones.
Played around a bit with a “–disable-api-atk” option for gtkmm but did not succeed because some m4 macros obviously create code in a different scope meaning it is out of the #ifdefs.
Pesc|O made a nice patch to integrate ctags 5.6 in anjuta. It’s not yet finished but it looks promising. The guys from ctags event want to open up their development so I hope we can have a better integration and maybe a libctags soon.
15. August 2006
The C++ support for Maemo is moving forward. Today, I have uploaded the packages for libhildon-fmmm and libhildon-widgetsmm which should allow to write complete maemo applications in C++ in theory.
14. August 2006
Libanjutamm is slowly on the way. I have wrapped most important things excluding the interfaces and written a sample plugin which does not yet load because I have to fight more with the GType system. Anyway, I hope to finish this sample soon and write a script to automaticly wrap the interfaces. Happy C++ coding
gtkmm on maemo
The gtkmm package for 386 should be up now and I am working on the armel package which should be finished soon. Next is the packaging of hildonmm which should hopefully not be too difficult.
Update: gtkmm packages for armel have been uploaded!
10. August 2006
Today, I built my first debian package. I still don’t fully understand how the whole mechanism works but at least I got good results by reusing the debian/ directory of an existing package as a starting point.
Naba made a great job and implemented a good auto-indent feature for anjuta, though the supported languages are only C, C++ and Java in the moment. Libanjuta is now a bit more wrapping friendly because I added “G_BEGIN_DECLS” and “G_END_DELCS” to all headers.
I am working on a libanjutamm. I follow the excellent instructions in the Gtkmm Documentation but I still have some problems. Anyway, I hope the nice guys at #c++ will be able to help me.
9. August 2006
I set up a maemo environment on my ubuntu box today. It was quite easy as they provide a good documentation and an easy installation script. There are some things that do not work perfectly yet but I think they can be fixed.
Currently trying to set up glom put I have problems to connect to the database. Seems like libgda got installed without support for postgresql but I am not sure yet. Retrying with jhbuild now…
No volunteers for gnome-build, yet! Applied patch from Günther Brammer (thanks!) to fix a parsing bug with “+=”.
7. August 2006
For my new employer Openismus, which is the first time that I get paid for anything Gnome or open-source related, I tried to fix Bug 343012 today. The patch itself was quite simple but it was a bit difficult to find out. I had no other testcase as the provided because I am not really familiar with the gtkrc system. Please try this patch out and tell me if there are any side-effects.
Going to have my last exam for this month at 3 pm…