> I really hope that the memory issues reported by Jurgen after a valgrind run are the > reason of all the current problems (including chdman getting miscompiled by Apple GCC > with default OPTIMIZE value ), so that we can fix all of them at once
Nothing Juergen reported is anything like fatal. He and everyone seems to assume that the Linux maintainer wouldn't already have tried Valgrind and submitted a patch called, say, "fix trivial uninitialized variable in chd.c" and that sort of thing.