On Thu, May 22, 2008 at 04:04:15PM +0100, Andrew Doran wrote: > Not a problem: in an SA scheme there is no requirement for the kernel to > know about intrathread signals. Er, signals generated by pthread_kill(). Andrew