[openbox] openbox/rox scrollwheel desktop skipping

Kacper Wysocki kacperw at online.no
Mon Aug 6 19:15:57 EDT 2007


On 8/6/07, Dana Jansens <danakj at orodu.net> wrote:
> On 8/6/07, Kacper Wysocki <kacperw at online.no> wrote:
> > On 8/6/07, Dana Jansens <danakj at orodu.net> wrote:
> > > On 8/6/07, Kacper Wysocki <kacperw at online.no> wrote:
> > > > On 8/6/07, Keith W <n4phth4 at gmail.com> wrote:
> > > > > FWIW, I don't see this behavior w/ Openbox 3.4.2 and ROX-Filer 2.6.1,
> > > > > however I do see it w/ the (old) pre-release  version 3.3.994 and
> > > > > ROX-Filer 2.6. (These are on different machines w/ possibly different
> > > > > configs; I can't look into it much further atm.)
> > > > >
> > > > > My OB config on the working box is essentially stock with minor focus
> > > > > tweaks for clients (i.e., don't raise when focused). ROX-Filer's
> > > > > config in the "compatibility" section has only "Panel is dock", "Don't
> > > > > use hostnames", and "Don't use extended attributes" checked.
> > > >
> > > > My options are the same, and I also didn't have this problem with
> > > > earlier OB3.4's. Could you try openbox 3.4.4?
> > >
> > > Just to add a little information to the clutter..
> > >
> > > The reason this is happening, I guess, is because ROX pinboard is
> > > "replaying" the mouse button event, causing it to get sent to Openbox
> > > twice somehow.  You don't have these bindings in both your "desktop"
> > > and your "root" contexts, do you?
> >
> > Nope, or rather - I didn't until Michael mentioned the root context.
> > However, I just tried ob3.4.4 on my work computer and the problem
> > doesn't occur here, I'll check what's different.
>
> Ah.. when he suggesting putting it in the "root" context, he also
> meant removing it from the "desktop" context.  if they replay the
> event to the root window, then Openbox will get it twice in the
> "desktop" context but only once in the "root" context.

The root context bindings don't do anything (since the pinboard
covering the screen is not the root desktop. similar to nautilus, but
not really. If I disable the desktop context bindings I get no effect
whatsoever. Thought I made that clear :-) sorry.

> That said, it'd be good if you could figure out why your work one
> isn't getting the event thrown around by ROX in the same way.

Mystified, I looked at xev again:

ButtonPress event, serial 32, synthetic NO, window 0x3000001,
    root 0x1a5, subw 0x0, time 1030461064, (5,176), root:(846,867),
    state 0x10, button 4, same_screen YES

ButtonRelease event, serial 32, synthetic NO, window 0x3000001,
    root 0x1a5, subw 0x0, time 1030461064, (5,176), root:(846,867),
    state 0x810, button 4, same_screen YES

ButtonPress event, serial 32, synthetic NO, window 0x3000001,
    root 0x1a5, subw 0x0, time 1030461064, (5,176), root:(846,867),
    state 0x10, button 4, same_screen YES

ButtonRelease event, serial 32, synthetic NO, window 0x3000001,
    root 0x1a5, subw 0x0, time 1030461064, (5,176), root:(846,867),
    state 0x810, button 4, same_screen YES

Yep, that's two events for one scroll notch. I can't think of any
reason why I saw only one initially, except maybe evil monkeys with
too much booze and spare time on their hands. If anyone has a hint as
to why this is happening I'd be happy to receive it (the hint),
otherwise I'll slink off somewhere more appropriate.

-K
-- 
http://kacper.doesntexist.org
http://windows.dontexist.net
There is symmetry in chaos, and chaos in overmuch order.



More information about the openbox mailing list