Ok I’m not dead, I’m not lost, I just haven’t found the time to write a blog entry in the past weeks.
As starting point, a little update on Gentoo/FreeBSD status: I managed to fix a couple of PAM-related problems on G/FBSD overlay, and many of the fixes for OpenPAM compatibility are now in main tree (also because most of them fixes also amd64 no-symlinks profile); at the same time, I also managed to add support to /lib/security path on OpenPAM. KDE 3.4.1 has already all the dependencies fixed and the patches applied so that it works out of the tree for G/FreeBSD, fixed ghostscript and jpeg.
I gave Stephan an experimental stage for 5.4 profile, but I’m probably going to rebuilt it completely fixing dependencies so that it works better.
GCC still deletes my crt1.o file so I need to fix that, too.. too bad that there’s no sandbox available for FreeBSD.
But now I want to talk about general Gentoo patches. One of the things most annoying when looking to external patches on other distributions is that they usually change the packages to suit their needs. Sometimes they apply patches conditionally on some archs (when they support more than one arch, sure). That’s true not just for Linux distributions, but also for FreeBSD’s ports, for Fink and for DarwinPorts. This makes those patches not good for upstream merge.
I think one of the strengths of Gentoo is that our patches are usually applied unconditionally and must work on many different architectures, and now that we are working on Gentoo/FreeBSD we have at least three operating systems to deal with (Linux, Mac OS X, FreeBSD), with many and many combinations which must work with the patches. That makes the patch quite always good for every kind of merging on upstream.
Why this is important for me? Because having patch applied upstream requires less extenral patches in distributions, problems distribution-local are less likely and the code is updated cleanly by the original devs.
That simplifies the work for us downstreams and for upstream (which has to deal with less-edited code when bugs are reported), specially when are released new versions, as we doesn’t need to see if the patches still applies.
Oh well, time to continue working and submitting patches! 🙂