Tom Murphy <[email protected]> writes:
> Tom Murphy wrote:
>> Hi,
>>
>> Here's an updated diff from Omar Polo's addition of group-last
>> command to cwm. I've been using it without issues and it's
>> really handy to be able to switch back to the previous
>> workspace you were on with it.
>>
>> Many thanks to Omar Polo for doing all the original work. I've
>> just updated the diff so it applies cleanly.
Thanks for reviving the patch.
I've kept using it this whole time and still found it super useful!
>> OK?
>>
>> Tom
>>
>> [...]
>
> I discovered something strange while testing this.
> X Windows appears to bug out and ends up showing one
> of the windows across all groups until cwm is reloaded.
>
> To reproduce:
>
> 1. Bind group-last to any key you like (I used 4-Tab)
> 2. Open up a window in group 1
> 3. Switch to group 2 and open up a different window there
> 4. Switch back to group 1
> 5. Jam on the 4-Tab key (or whatever you bound in step 1)
> 6. After a while the windows flicker and then you get one
> of the windows stuck as sticky.
> 7. $HOME/.xsession-errors contains error messages like these:
>
> [ 10/10/21 22:29:12.946 x_create_picture_with_pictfmt_and_pixmap ERROR ]
> failed to create picture (X error 9 DRAWABLE request 139 minor 4 serial 4284)
> [ 10/10/21 22:29:12.962 paint_one ERROR ] Window 0x0060000c is missing
> painting data.
> [ 10/10/21 22:29:13.079 x_create_picture_with_pictfmt_and_pixmap ERROR ]
> failed to create picture (X error 9 DRAWABLE request 139 minor 4 serial 4501)
> [ 10/10/21 22:29:13.094 paint_one ERROR ] Window 0x0060000c is missing
> painting data.
> (etc... with different numbers after 'serial')
>
> I wonder if I triggered some X Windows bug by not using this function
> as intended? (toggling it over and over would not be normal use)
>
> It does work on normal desktop usage, as most people won't be trying
> to break it like this.
>
> Any ideas? Seems strange behavior to me.
I think it has happened in the past to me too, but I can't always
reproduce it (even if I hold 4-Tab for half a minute, sometime it
doesn't get stuck.) I've been using this patch for a long time, so I
don't know if it's something that happened also with an unpatched cwm.
Just one small clarification to your description: (point 6) a window can
get "sticky", but not in the usual cwm sense. A stuck window is not
able to get focus, if you click on it you get the usual cwm menu as if
the window wasn't there, and you can't focus or resize it or anything
else. But after a restart, it's "live" again.
I also don't think the error messages you posted are generated by cwm.
I've managed to reproduce it once in the current session and I don't
have anything like that in .xsession-errors. (I don't have anything
that resembles a cwm error either in .xsession-errors, and I'm running
cwm -vvv)
Excluding the common debug3: xev_handle_{enternotify,propertynotify,...}
all that remains are a couple of
error 8: BadMatch (invalid parameter attributes) request 142 minor 6 serial
163124
error 9: BadDrawable (invalid Pixmap or Window parameter) request 139 minor 4
serial 163125
error 4: BadPixmap (invalid Pixmap parameter) request 54 minor 0 serial 163325
that don't seem to be logged by cwm and neither to be related to the
error (they are logged both long before and after the accident.)
> Thanks,
> Tom
