SIGTSTP propagation for 0.80 ?

[David Golden]

On 2005-09-22 10:08:22 -0400, Pete Wyckoff wrote:

> I guess I'm still tempted to do what PBS tradition has always done
> and work with you to adapt your original patch that directly
> suspends all the tasks under the control of mpiexec. 

Hi,

Haven't wandered off - just slightly stuck:

It seems to me the only* [sane] way to do [what I think is] the "right" 
thing in the concurrent interactive case is to fork off a concurrent master
server as a detached daemon, distinct from all concurrent clients, because 
for return of control to the shell, there *has* to be a  stop of the 
Ctrl-Zed i.e. TSTPed process as shells apparently conventionally
waitpid()/WIFSTOPPED():
http://www.gnu.org/software/libc/manual/html_node/Stopped-and-Terminated-Jobs.html

* Unless there's some way to avoid the need for the concurrent
master in the first place  - but I believe it's already
been said it's a TM API limitation that means you _have_ to do it
all from one process?

While you or I could presumably make mpiexec do the fork, is it
a direction you want to go?  Could/should prep a patch anyway and
see what you think, I guess...