Re: [RP] tmpwm fix?

[Bernhard R. Link]

* Shawn <address@hidden> [081101 18:36]:
> Can anyone who's had a problem with tmpwm pull the latest git and give
> it a go? I think it may be fixed or at least more stable than ever before.

There seem to be multiple fail modes. Sadly I currently (with the old
version) can only reproduce the one about ratpoison exiting when the
tmpwm returned to fast. The one where the prefix key fails I cannot, so
I cannot say whether it is fixed.

But the problem with ratpoison sometimes exiting because the X server
thinking there is still another window manager seems to be fixed. I
sometimes get quite a long time where it waits in that loop. I think
this is because the X server is so occupied telling ratpoison there is
still another window manager, that it does not find the time to look
for lost clients, especially for a window manager that was killed which
actually caused ratpoison to ask this question.

The following patch fixes this for me (at least as sure I can say from
not seing it again in a few dozens trys while it usually happens for
about 3 out of 4 tries without[1]):

diff --git a/src/actions.c b/src/actions.c
index 136d114..f23bd33 100644
--- a/src/actions.c
+++ b/src/actions.c
@@ -4677,6 +4677,8 @@ cmd_tmpwm (int interactive, struct cmdarg **args)
                      | SubstructureRedirectMask | SubstructureNotifyMask
                      | StructureNotifyMask);
         XSync (dpy, False);
+       if (tmpwm_error_raised)
+          sleep(1);
       }
   } while (tmpwm_error_raised);
   XSetErrorHandler (old_handler);


Hochachtungsvoll,
        Bernhard R. Link

[1] Did I already mention that I hate race conditions?