Hi Simon, Raphael
Thank you two for your quick response, and also investigating beyond
the parameters I suggested. Raphael, I just saw your response just now
and will have to digest it a bit.
> Notebook, Celeron 333, XFree86 3.3.6, Intuos A5:
> I get a context menu. [as do I] For the misbehaviour I do not need to be inside
> the selection. If I make a selection, select e.g. a paint-tool, paint
> someting (in- or outside the selection) and while pressing the pen down
> press the "right mousebutton" a menu pops up ...
I too get the menu. In contrast, if I were to perform similar actions with the
real mouse/right button, with the "wacom" device disabled, no context
> When selecting something (e.g. tearing the menu off) the marching ants die.
> When opening a new image they are alive for the new image.
I Confirm this as well. Excellent, Simon, for you have brought forth the work-around
to the problem for the poor Gimp user until this is resolved: When Marching Ants Die,
Cntl-D for a New Image (I suppose saving the image first won't be such a bad idea).
This creates a new Selection object for that image without an unbalanced pause count.
<remaining very useful observations snipped for brevity... Behaviour mirrors that of
> It is possible that the graphire Marc had is more similiar to the
> Artpad than to the Intuos. I will dig a little bit in the XFree86-Driver.
> Hopefully we can fix this in a sensible manner.
Well, Xsgi != Xfree86 AND the Leipeid GPL wacom driver != commercial Wacom
driver for SGI, and yet our observations are essentially the same, so my attention
is drawn toward how GTK/GDK abstracts these distinct devices into a normalized
display/pointing device. In other words, the problem is not specific to
X-server/tablet combinations, but how GTK (in uneasy co-existence with the
not very well understood X Input Extension) homogenizes these devices so that
what one expects of something like gdk_pointer_grab() is the same no matter what
piece of hardware is doing the pointing.
At present, we know that is not entirely true in two mixes of hardware.
However, the manner in which they are untrue appears essentially the same.
I have, however, just read Raphael's well-documented response, and he is seeing
quite different things. I need more time to reflect on his mail, but his results
that 10498 does have hardware dependencies. I think a sensible fix is not immediately
Thank you all for your observations!
Be good, be well