[GRASS5] closing monitors by click
Markus Neteler
neteler at geog.uni-hannover.de
Wed Nov 29 13:08:44 EST 2000
Hi Andreas,
On Wed, Nov 29, 2000 at 02:06:41PM +0100, Andreas Lange wrote:
> Hi Markus, Hi Justin,
>
> you can close the grass monitor on Linux with the "Kill" option too (i
> use a lesstif/motif window manager on linux that has the
> IRIX-look-and-feel). But you can not restart the same (x0..) driver
> again, as the lockfile is not removed.
Yes, that's right. The left-over lock file is causing troubles.
Suggestion: Maybe a check could be added like this:
- lock file found
- check if monitor process is (still) running
- yes: lock
- no: remove the lock file, start monitor
- lock file not found
- start monitor
I did not look into the code, but maybe this is a way to get rid
of the problem.
> On the SGI i can close the monitor with close or exit (mouse) and
> re-open it again.
Helena reported as well that SGI users don't face this problem (lucky
people!)
> This all seems to be for the same reason that d.mon -L does not work on
> linux. I see now that i was looking in the wrong place (SWITCHER.c), the
> problem must be somewhere in the code that checks for the monitors
> running. There must be a timeout/deadlock, as d.mon -L hangs forever if
> it encounters a running monitor.
> This may be because some system call have other semantics on linux than
> on commercial unix.
Sidenote: If you use the IPC based driver, the d.mon -L is working (on
Linux). With fifos it doesn't. Maybe helpful? So only the IPC-modified
files might be related to this problem (probably reduces the number of files
to be checked).
> I'll look into this, until now i can only close the monitor on linux and
> re-open it after i call d.mon -L once. This produces an error message
> and the lock seems to be gone.
>
> cu,
>
> Andreas
Yours
Markus
----------------------------------------
If you want to unsubscribe from GRASS Development Team mailing list write to:
minordomo at geog.uni-hannover.de with
subject 'unsubscribe grass5'
More information about the grass-dev
mailing list