..feeding the attention whore within

I’ve had it lying around without no changes taking place for almost a month now, so I figured that I’d just go on and release pyliblzma 0.5.2 as I never updated the cooker package to 0.5.1 (which fixes some nasty bugs) because I was planning on releasing 0.5.2 right away anyways..

I never got that far due to various reasons which made me not getting very far with anything in general, so now I’ll just release it out there! :)

This version is mainly synchronizing code with bz2module.c from python upstream adding various bug fixes, new python 2.6/3.0 specific features etc. and I also rewrote much of the code to be more in sync for easier maintenance for myself and any others.

Related to this release I added a nice patch to python in cooker to automatically disable python’s memory allocator (pyalloc) when running under valgrind to make it easier to debug, this made my work a lot easier when debugging with valgrind and should do the same for anyone else on mandriva debugging python stuff with valgrind as well. :)

Since I returned from FOSDEM I’ve also done quite some work on getting to know the the Gdium & the mips platform, merging stuff from the Mandriva Glinux mips tree to cooker and on the toolchain in cooker. Cross crompilers should now build properly on most platforms so that one can build a mips compiler on ie. x86_64, so this should be in place for the system gcc 4.3 compiler, I’ve also since imported a gcc4.4 package to start work on gcc 4.4 and to clean it up a bit and eventually get a complete cross compiler toolchain for mips with all the nice mips related improvements introduced in the upcoming gcc 4.4 release. :)

I’m not really sure on what the plans are for the mips cooker tree yet, but I’m planning on getting the complete toolchain built to build everything on my quad core x86_64 rather than on the 900 mhz loongson Gdium netbook of mine, then I should be able to easily build the full distribution in relative short time and also make it easier for others as well on non-mipsel architectures.

Working on this now I get to reuse a lot of the work and experience from the past with my (hobby;) efforts on the sparc port, so it’s a nice thing to see it not going to waste even if not being more than a sporadically maintained port on a hobby basis. :D