Inidentally, ADC News #292 deals with this:

   http://developer.apple.com/qa/qa2001/qa1123.html

This and the included link to the unix faq seem to imply that it is hard to impossible to do this portably...

Joern

Joern Eyrich wrote:
> 
> Hi,
> 
> I'm neither an export on unix intricacies nor on osprocess, but I won't spare you my comments anyway... ;-)
> 
> John M McIntosh wrote:
> > Actually it's there because the Smalltalk code is checking to see if
> > the pid is valid. It also does this because if you save an image and
> > restart then we need to confirm if the pids we think we have in the
> > squeak code actually exist or not.
> 
> Hm, I would prefer to invalidate the list of remembered pids on image startup. even if there is a process with this id, you don't really know if it's the same process. and even if it is,
> 
> a) if the image-saving-squeak has been terminated in the meantime, the child process is bound to have been reparented to init and squeak wouldn't get a SIGCHLD anyway, would it?
> b) if the image-saving-squeak is still running, now two squeaks will think it's their child process, which seems of dubious usefulness to me
> 
> I'm not sure I understand why you are looking for the death of a specific child, though; I can't look at the code, I've only found UnixOSProcessPlugin.c on minnow, and if I download it, it's corrupt...
> 
> To me, Richard's suggestion looks ok.
> 
> If you really have to get the pid info, however, maybe the sources to Darwin's 'ps' command will give a clue:
> http://www.opensource.apple.com/cgi-bin/registered/cvs/Commands/NeXT/adv_cmds/ps.tproj/
> 
> Joern