[GRASS5] Re: [GRASSLIST:4009] d.zoom compared to d.rast.zoom

GRASS GRASS developer
1

On Wednesday 03 July 2002 03:21 pm, Roy Sanderson wrote:

Dear Grass-users

There seems to be a difference in the behaviour of d.zoom (GRASS 5)
compared with d.rast.zoom (GRASS 4x) in its treatment of regions in zooming
out. d.rast.zoom gives the option of rejecting a window, without affecting
the region, which is very useful to view a small area of a map in detail
before continuing at the original region. GRASS 5x d.zoom seems to modify
the region, even if the 'unzoom' option is selected having originally
zoomed in. There does not appear to be a facility to quit without
modifying the region.

Or am I just asking another silly question?

No, you are right. I have modified d.zoom, because I hated to accept/reject
each selected region.
We could add "previous region", but we don't have enough buttons on mouse.
This problem appears on many places.

Would it be possible to catch keystrokes by
R_get_location_with_pointer()? It would be also useful to have some
way, how to break running R_get_location_with_pointer() from GUI.
(Glynn?)

Radim

2

Radim Blazek wrote:

Would it be possible to catch keystrokes by
R_get_location_with_pointer()? It would be also useful to have some
way, how to break running R_get_location_with_pointer() from GUI.
(Glynn?)

Actually handling KeyPress events wouldn't be that hard, but:

1. Arranging for KeyPress events to be reported is non-trivial. Button
events are always reported to the window under the pointer, but key
events are normally dispatched according to the keyboard focus, and
focus management is non-trivial.

2. How would they be reported?

3. The modal input model implied by R_get_location_* is a hangover
from the days of the Tek-4015 etc. I'm reluctant to put much effort
into extending it, as it's likely to be the first thing to go in the
event of any significant UI update.

--
Glynn Clements <glynn.clements@virgin.net>