[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Lispm FORTRAN goes to Hyperspace

Anybody seen anything like this?

I'm build a system that is --to first order-- a wrapper around a
(pre-existing) FORTRAN program.

The FORTRAN program is launched in Lisp Listener #1.  In short, it sleeps
waiting for input (variables get set using simple semaphore-like flags).

The Lisp program, running in Lisp Listener #2, is basically a command-loop
that 1] accepts (character) commands from the user and 2] sets the
variables used by the FORTRAN.

Everything runs along fine except sometimes Lisp Listener #2 simply hangs
while waiting for input from the user.

There are really no other symptoms.  It just accepts *no* input (not even,
say, M-C-Suspend).  It just sits there, often for several minutes, and
after a while seems to wake up and start processing the last command.
After which it is fine (until the next time it happens).

Typing M-C-Abort _several_ times will usually cause the "You can't
interrupt this process..." notification to come up.  However typing any
character simply causes the notification to go away without any effect.

Trying to break in by some other route, say by going into peak and trying
to KILL or RESET the process will cause _that_ window/process to hang in
the same way.

Occasionally, if my Lisp-program launches a menu, the menu will hang, too.

It seems to happen randomly, although, it happens most often if I re-enter
the system from some other system (e.g. Zmacs) (although it will hang
_rarely_ even if I stay in the program).

It happens on more than one system (i.e. a XL400 and 3670, both running
8.0).  Also, it _seems_ to happen more and more often the longer the system
has been up since a boot.

This _only_ happens when the FORTRAN is running.

So, in short, it simply hangs and there seems to be no way that I can break
in to see what it's waiting for.

Any ideas?  At this point I'll try anything.