fsset@bach.lerc.nasa.gov (Scott E. Townsend) (10/18/90)
I am writing a front-end to a parallel processor attached to a Personal Iris and would like the capability to provide a wsh(1) window for each node in the system for debug output, etc. I'd really rather not reinvent the capabilities of wsh(1) like the scroll bar, logging, etc. My first guess is to have the front-end create two named pipes for I/O and then fork/system "wsh -c mycu" where 'mycu' is a cu-like program that copies stdin and stdout to the appropriate named pipes. The front-end then connects these named pipes to the appropriate node's program. I realize that I will hit an open file limit rather quickly, but I'm hoping the user will be too confused by all those windows before then ;-) Is this a reasonable approach? Do you know of any problems/secrets I should look out for? Any comments would be appreciated. -- ------------------------------------------------------------------------ Scott Townsend | Phone: 216-433-8101 NASA Lewis Research Center | Mail Stop: 5-11 Cleveland, Ohio 44135 | Email: fsset@bach.lerc.nasa.gov ------------------------------------------------------------------------