You can switch it on again by sending #restoreAcceleration, but this won't
fix the underlying problem that #intoWorld: isn't called.  Try the attached
changeset.

Joshua

On Sun, Aug 10, 2003 at 01:02:17PM +0200, Martin Kuball wrote:
> OK, I was able to almost hunt down the problem. When I open a 
> B3DSceneExplorerMorph I can activate hardware acceleration. When I grap the 
> morph and move it around, acceleration is disabled instantaneously. And I 
> can't switch it on again. The reason seems to be that in B3DSceneMorph>> 
> outOfWorld: the method suspendAcceleration is called. But the corresponding 
> method intoWorld: where the setting is reversed is never called.
> 
> Martin
> 
> 
> Am Donnerstag, 7. August 2003 00:07 schrieben Sie:
> > > By the way I filed in your fix for removing the ogl window
> > > when deleting the morph. Now it seems that I'm not able to
> > > activate the hardware acceleration any more. Is this possible
> > > or am I missing something else?
> >
> > Depends on what you mean by "it seems that I'm not able to activate [it]"
> > ;-) Can you elaborate a bit on this? How do you try to activate HWA? What
> > happens when you try?
> >
> > Cheers,
> >   - Andreas
-------------- next part --------------
'From Squeak3.6beta of ''4 July 2003'' [latest update: #5371] on 10 August 2003 at 11:10:38 am'!

!Morph methodsFor: 'dropping/grabbing' stamp: 'jcg 8/8/2003 01:34'!
justDroppedInto: aMorph event: anEvent
	"This message is sent to a dropped morph after it has been dropped on -- and been accepted by -- a drop-sensitive morph"

	| aWindow partsBinCase cmd aStack |
	self formerOwner: nil.
	self formerPosition: nil.
	cmd _ self valueOfProperty: #undoGrabCommand.
	cmd ifNotNil:[aMorph rememberCommand: cmd.
				self removeProperty: #undoGrabCommand].
	(partsBinCase _ aMorph isPartsBin) ifFalse:
		[self isPartsDonor: false].
	(aWindow _ aMorph ownerThatIsA: SystemWindow) ifNotNil:
		[aWindow isActive ifFalse:
			[aWindow activate]].
	partsBinCase ifFalse: [self intoWorld: self world].
	"Note an unhappy inefficiency here:  the startStepping... call (inside #intoWorld:) will often have already been called in the sequence leading up to entry to this method, but unfortunately the isPartsDonor: call often will not have already happened, with the result that the startStepping... call will not have resulted in the startage of the steppage."

	"An object launched by certain parts-launcher mechanisms should end up fully visible..."
	(self hasProperty: #beFullyVisibleAfterDrop) ifTrue:
		[aMorph == ActiveWorld ifTrue:
			[self goHome].
		self removeProperty: #beFullyVisibleAfterDrop].

	(self holdsSeparateDataForEachInstance and: [(aStack _ self stack) notNil])
		ifTrue:
			[aStack reassessBackgroundShape]
! !