Thanks Don. It was really confusing to me for awhile trying to figure
out what was going on.
In a related visibility issue, the default MapMouseInterpreter doesn't
take into account whether a graphic in a list is visible or not inside
getGeometryUnder(). I had to override that method for other reasons, so
I went ahead and added that check since that's the behavior I want (in
my case, a user shouldn't be able to 'click' on a graphic that's
invisible).
It's arguable whether this modification would be better made in
OMGraphicList's "findClosest()" method - and the more I think about it,
the more I feel it should be: Suppose there were 2 overlapping graphics
positioned in the OMGraphicList, with one of them invisible. (It's quite
possible in the application I'm working on for this to happen
frequently). If findClosest() happens to return the invisible graphic,
there would NEVER be a way to click on the (correct) visible one,
because it would never be found as the 'closest'.
-Peter
-----Original Message-----
From: Don Dietrick [mailto:dietrick_at_bbn.com]
Sent: Friday, September 24, 2004 4:03 PM
To: Peter Solderitsch
Cc: openmap-users_at_bbn.com
Subject: Re: [OpenMap Users] Visibility and OMScalingRaster
Hi Peter,
Thanks for pointing this out. I actually think it's wrong that the
visibility is changed at all. Instead, the internal bitmap image
should be set to null if the scaled image doesn't overlap with the
projection window, which would signal the superclass render method to
not bother trying to render it.
The problem with using the visibility variable is that if the layer
author makes an OMScalingRaster invisible for some reason, and the
OMScalingRaster is offscreen, when it comes back on screen it will be
made visible again automatically, even though the author didn't change
the visibility setting.
I've made the change and tested it, I'm going to be updating the
repository shortly.
- Don
On Sep 24, 2004, at 2:05 PM, Peter Solderitsch wrote:
>
>
> All,
>
> Apologies if this has been brought up before, but I didn't see
> anything about it in the mailing list archives.
>
> I believe that OMScalingRaster (and its subclasses, like
> OMScalingIcon) have a bug that prevents objects of that type to ever
> be successfully set as invisible.
>
> I've been trying to track down behavior where I would set an
> OMGraphicList as invisible, and all the graphics in the list except
> for the scaling rasters would be correctly invisible, but the rasters
> remained on screen. Doing some digging, it looks to me like the
> scaleTo(Projection) method in OMScalingRaster.java does not make an
> attempt to preserve the visibility set on the graphic prior to the
> scaleTo() method being called; rather, it always sets visibility to
> 'true'.
>
> I fixed my immediate need by subclassing OMScalingIcon (which was the
> class I was using), and overriding scaleTo() to first remember the
> prior visibility state, and then re-set the visibility after the
> super.scaleTo() has been called. However, I believe this is probably
> a
> good candidate to get fixed in the main source branch.
>
> Thanks,
> -Peter
> --
> [To unsubscribe to this list send an email to "majdart_at_bbn.com" with
> the following text in the BODY of the message "unsubscribe
> openmap-users"]
>
>
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Don Dietrick, dietrick_at_bbn.com
BBN Technologies, Cambridge, MA =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
--
[To unsubscribe to this list send an email to "majdart_at_bbn.com"
with the following text in the BODY of the message "unsubscribe openmap-users"]
Received on Fri Sep 24 2004 - 16:11:51 EDT