i've submitted patches that fix the race condition i was seeing (update tasks that would finish so quickly that the parent would get confused and think an update process was still running), which was signal related, but not how i thought. i also updated the signal semantics to what i think are completely protected behaviours. i did use sigset instead of the equivalent sigaction, but that shouldn't be a big deal. i sent it directly to proff, so i don't know when they'll show up (or if they will). i can s send them to the list if enough people are interested.
