call-process-region is called with 0 as the
fifth arg ,so Xemacs should not be waiting around for the process
to terminate.
Here's another data point - during this period of time when XEmacs is
not responding, my process table shows a zombie sendmail process:
sirius:cgw$ ps ax|grep mail
29221 ? S 0:00 /usr/sbin/sendmail -q15m -bd
31845 ? Z 0:00 (sendmail <zombie>)
31846 ? S 0:00 /usr/lib/sendmail -oi -f cgw -oem -odb -t
Here's another run, in "forest" (-f) format, slightly edited
29221 ? S 0:00 /usr/sbin/sendmail -q15m -bd
31994 ? S 0:00 /usr/lib/sendmail -oi -f cgw -oem -odb -t
187 2 SW 0:00 (bash)
298 2 SW 0:00 \_ (startx)
299 2 SW 0:00 \_ (xinit)
300 ? R 7:09 \_ /usr/X11R6/bin/Xwrapper :0
303 2 S 0:42 \_ wmaker
4125 2 SW 0:00 \_ (sh)
4127 2 S 6:04 | \_ xemacs
31993 ? Z 0:00 | \_ (sendmail <zombie>)
4156 p2 SW 0:00 | \_ (gnuserv)
Process 31994 is the process actually handling the delivery of the
message. I don't yet understand what process 31993 is.