Adam “replaced” Hoka, who has advertised and bug-spotted (and fixed) in the past, begun a Haiku port of mksh. He has been given CVS commit access for that purpose (possibly extended later), and as such, is now the 8th developer (counting both active and passive ones) to The MirOS Project.

Apparently, the Haiku port now compiles but does not run. Possibly another buggy operating system to add to the list.

I somehow end up fixing pcc each time I try to build it. But code size on i386 has much improved recently, although still nowhere near gcc ‘-Os’ already much better than its ‘-O2’ counterpart, both gcc3 and gcc4. I’m still waiting for more bug fixes, and more of my fixes integrated into upstream, before updating the pcc mirport though.

While I removed the use of __typeof__ from the mksh(1) source, SUNWcc still explodes on it unless we use HAVE_EXPSTMT=0 *sigh*

There are new “hosted” subprojects in our CVS repository. These are for each committer to place his/her stuff into, which, in contrast to other stuff in the contrib module, does not necessarily need to have something to do with BSD, Unix, shell, etc. However, existing mismatched code (the .exe files of my keyboard layout and Benny’s polymers code) will not be moved, because that doesn’t make sense with cvs(1).

Today, my health has been somewhat better, but I still feel as if I had a cold but hidden, so it does not impact me that much but I cannot lose it either. Headaches started at about 20³⁰, which I consider good, even though I now cannot comply with Jonathan’s wish of having a look at llvm-gcc for spotting a bug in libobjfw.