Cas,
   Something you mentioned rang a bell:

@ anything from STDIN ? Why does "uname -a" not have this problem ?

  It worries me as well. It's even more worrisome that if you do do
  remsh;info="foo -l user ls"
  and hit CR once or twice, you get the output.....
  So someone is waiting for something.

There's an issue when a son process is calling XCONTRAP while sharing the
same $stdin as the parent.  It treats it as an I/O to $stdin so it blocks
until the pending I/O on the parent completes.  Could this be what's up
with remsh?  Just a guess.

Rat
 ------------------------------
Rich Trapp
OMNIDEX Product Manager - IMAGE/SQL
Dynamic Information Systems Corporation  http://www.disc.com
Work: [log in to unmask]  Play: [log in to unmask]
Voice: 303-444-4000  Fax: 303-444-0208