[STUMP] Hangs with run-shell-command (was: Lisp crashes with stumpwm)

[Milan Zamazal]

>>>>> "JS" == Julian Stecklina <address@hidden> writes:

    JS> Milan Zamazal <address@hidden> writes:
    >> The only remaining problem is that if using run-shell-command in
    >> the timers then StumpWM hangs about once or twice per day, with
    >> both the threaded and non-threaded SBCL versions.  No error
    >> message, no backtrace, no way to handle windows and SBCL has to
    >> be killed from another terminal.  Well, the obvious work around
    >> is not to use run-shell-command which is possible in most cases.

    JS> Could you post this to the SBCL people? Perhaps you can build a
    JS> minimal example outside of stumpwm. I guess, it would be quite
    JS> helpful in solving this problem.

I'll try when I have some time...

>>>>> "VM" == Vitaly Mayatskikh <address@hidden> writes:

    VM> I thought it was fixed in 1.0.16:

    VM>   * bug fix: copying output from RUN-PROGRAM to a stream
    VM> signalled bogus errors if select() was interrupted.

The problem I describe above happens with SBCL 1.0.16 (from Debian).

>>>>> "SB" == Shawn Betts <address@hidden> writes:

    SB> It's moot now but I want to point out that multiple timers isn't
    SB> exactly the source of the issue. stumpwm figures out when the
    SB> next timeout is due and uses clx's event timeout to schedule a
    SB> wake up at that time. It then does what needs to be done (clear
    SB> the message bar, update modeline, etc) and then figures out when
    SB> the next timeout is due. and so on. I felt labeling this as
    SB> "multiple timers" was misleading.

I see.  Good to know.

Regards,

Milan Zamazal