We’re slowly getting closer to spring! It was about 12°C even in the late afternoon of the weekend, and I took out my bike on Sunday (helped a friend with cleaning up after moving on Saturday). Sadly, three DNFs (one search aborted due to the law enforcement approaching, one not even attempted due to too many muggles – although I already had logged another cache with almost the same name in exactly the same location –, and the third one not done because a certain institution’s garden has different opening times during winter period and I didn’t quite want to risk using the fire fighters’ entrance, like some other logger. Not even for a coin.
Anyway, the tpm driver bugs me a little (it’s possible to hang it from user space), and I can’t make -fwrapv default because -ftrapv wouldn’t disable it then. Unless I look more into gcc’s source again. But I hereby officially announce that code on MirBSD™ can assume wrapping semantics and 8/16/32/64 bit wide integral types, big or little endian. No 9 bit bytes, 36 bit PDP endian integers, saturation arithmetics. Ever. It’s a promise.
The IcedTea cross-compile patch and the OpenJDK BSD patchkit are not build system compatible. I probably need to go the route of using compat_openbsd(8) for it. Kurt Miller said if I sigh Sun’s agreement chances are good MirBSD support making it upstream (into the patchkit, for now, but maybe eventually into OpenJDK 7). MirUsers will just have to use a binary package I provide for bootstrap.
More… some other time. Still wish humans would hibernate too…