On Wed, 10 Jan 2007 22:32:11 +0100, Steffen Mueller <[EMAIL PROTECTED]> wrote:
>Jan Dubois schrieb: >> All the "other" signals should be delivered through the "safe" mechanism >> though. > >I have to rely on my memory now, but aren't kill(0, $pid) and kill(9, >$pid) the only signals win32's kill supports? That would leave no "soft" >kill. I may be totally wrong, though... It also supports SIGTERM (Ctrl-Break) and SIGINT (Ctrl-C) for subprocesses. SIGBREAK is mapped to SIGTERM, and all other signals are mapped to SIGKILL. But for pseudo-fork processes it looks like all signals are "supported" now, at least with the 5.8.8. Supported in the sense that they will invoke the corresponding signal handler in the child interpreter. I didn't test this right now, so this is just educated guessing from looking at the sources. Cheers, -Jan
