aalloc was a failure and prone to mis-type; I’ve started on galloc, which is supposed to be even more type-safe and flexible, yet still segfaults on me as well.

In the meanwhile, to not stay bored, I enhanced the Unicode (MirOS OPTU-8/OPTU-16) function suite in MirBSD and hacked an implementation of Plan 9’s Rune functions — dubbed p9¾_*.c due to usage of our “internationalisation like Plan 9, just on the next layer, and within the confines of UNIX®” approach to internationalisation. These (as well as some well-known ones like wcsrtombs(3)) should be OPTU safe, and we now have a macro telling us if a wide character is part of the OPTU Raw Octet codepoint range (in the CSUR PUA assignment).

I’ve contacted Bruno Haible again because his libutf8 misses the Unicode Title case (complementing Upper and Lower case)… I’ve got an idea how to implement it but would like to double-check with him to ensure nothing breaks. The Plan 9 functions need it (these two are currently implemented as stubs that just throw ENOCOFFEE).

The number of manual pages also raised…

I built XFree86® with a (slightly beefed up) Reiser CCCP in use as sole C Præprocessor now… and have yet to notice failures in operation. Now we’ll have to find out what else stuff is there depending on the existence of /usr/bin/cpp, possibly switching it, like the <sys.mk> mirmake(1) file, to ${CC} -E - (which is not 100% compatible, because cpp can also read from stdin without the dash, whereas, with the dash, neither can read from only a file). X11 etc. can be switched to /usr/libexec/cpp then.