Hi Serge, I encountered this behavior recently as well. It seems to be a newly introduced bug. Maybe because of a change in Roassal - but I think it's not necessarily a bug there. Maybe I didn't adapt to a change correctly, since it does not seem to happen in other Roassal views. So far I haven't been able to track it down and fix it but I will work on it and update you when I found something. Any input on this is welcome of course. Cheers, Dennis On Tue, May 7, 2013 at 7:35 AM, Serge Stinckwich <serge.stinckwich(a)gmail.com> wrote: Dear all, I'm doing some experiments with QuickSilver, but I have always some Morph issue when running an expression like this one: QsSystemViewer for: 'Zinc' I expand the node and try to zoom out. After a while, I have a broken morph with red square and yellow cross. The problem apparently is always in ROView>>elementsToRenderDo: aBlock Maybe a Roassal bug. Regards, -- Serge Stinckwich UCBN & UMI UMMISCO 209 (IRD/UPMC) Every DSL ends up being Smalltalk http://www.doesnotunderstand.org/ _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

How to reproduce the error? Alexandre On May 7, 2013, at 1:35 AM, Serge Stinckwich <serge.stinckwich(a)gmail.com> wrote: -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

I think I have solved the problem. Would you mind to check please? This was tricky. The problem was due to the recent optimization I did: the rendering queue was modified while being iterated. Fortunately, I still had the implementation fresh in mind to solve the problem. Update to Roassal 1.351 Cheers, Alexandre On May 7, 2013, at 3:23 PM, Dennis Schenk <d.schenk(a)students.unibe.ch> wrote: alexandre.bergel(a)me.com> wrote: wrote: -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Btw: the same thing with padding lead to the method: removeMarginFor: elements "Assume that elements is not nil, and that all the elements have the same parent" elements do: [ :e | e allElementsDo: [ :ee | ee resizeStrategy: (e resizeStrategy paddingGap: 0) ] ] in ROTreeMapLayout. I think from a performance perspective this is not very nice as well, since we have to go over all the elements again and fix something, which could be prevented with default values or a simple handling of nil values, to ignore the padding. On Wed, May 8, 2013 at 8:58 AM, Dennis Schenk <d.schenk(a)students.unibe.ch> wrote: Hi Alexandre, Thanks for looking into this. When I do QsSystemViewer for: 'Zinc' now, I get an error that padding is nil of a ROFixedSizedParent instance I'm using when drawing a layout. I fixed that by changing: visualRepresentationOfChildNode resizeStrategy: ROFixedSizedParent instance. to: visualRepresentationOfChildNode resizeStrategy: (ROFixedSizedParent instance paddingGap: 0). But that is not very nice, I think. Wouldn't it make sense to have some default values in place? Or to allow a nil value and interpret it as 0@0? After I fixed that, the next bug appeared, again the same thing with padding being nil, but this time when creating a ROPopup. ROPopup:createElement is called, but then in ROElement:addedInAnElement: el self translateBy: el topLeft negated. self translateBy: el resizeStrategy padding. self parent: el; view: el view. el addElement: self. el adjustSizeIfNecessary The padding is again nil for resizeStrategy. For ROPopup ROExtensibleParent is used. I think there should be some check: if padding is nil or 0@0, no translation should be attempted. What do you think? Also a related question: I saw that ROFixedSizeParent was changed in the ROTreemapLayout to ROPermissiveParent: What is the difference between the two? When should one be used and when the other? Cheers, Dennis On Wed, May 8, 2013 at 1:52 AM, Alexandre Bergel <alexandre.bergel(a)me.com> wrote: I think I have solved the problem. Would you mind to check please? This was tricky. The problem was due to the recent optimization I did: the rendering queue was modified while being iterated. Fortunately, I still had the implementation fresh in mind to solve the problem. Update to Roassal 1.351 Cheers, Alexandre On May 7, 2013, at 3:23 PM, Dennis Schenk <d.schenk(a)students.unibe.ch> wrote: -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

the same parent" resizeStrategy paddingGap: 0) ] ]

Well... I think the whole story comes from the fact that the tree map is considered as a layout. And this layout is very different from other layouts (e.g., tree map layout), since edges have to be removed, zOrder updated, size of the nodes changed, interaction of the element modified (they need to forward drag and drop right?). This makes me think that maybe the tree map should maybe be a new builder (as the MondrianViewBuilder). At the moment to define your nodes, you can indeed use a padding = 0. Regarding the performance, your method #removeMarginFor: is called in the initialize of the layout. Which means that it is executed once. Even iterating over 100 000 elements is very fast. Let me know if there is something unclear. Cheers, Alexandre -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Hi Alexandre, Thanks for looking into this. When I do QsSystemViewer for: 'Zinc' now, I get an error that padding is nil of a ROFixedSizedParent instance I'm using when drawing a layout. I fixed that by changing: visualRepresentationOfChildNode resizeStrategy: ROFixedSizedParent instance. to: visualRepresentationOfChildNode resizeStrategy: (ROFixedSizedParent instance paddingGap: 0). But that is not very nice, I think. Wouldn't it make sense to have some default values in place? Or to allow a nil value and interpret it as 0@0? After I fixed that, the next bug appeared, again the same thing with padding being nil, but this time when creating a ROPopup. ROPopup:createElement is called, but then in ROElement:addedInAnElement: el self translateBy: el topLeft negated. self translateBy: el resizeStrategy padding. self parent: el; view: el view. el addElement: self. el adjustSizeIfNecessary The padding is again nil for resizeStrategy. For ROPopup ROExtensibleParent is used. I think there should be some check: if padding is nil or 0@0, no translation should be attempted. What do you think? Also a related question: I saw that ROFixedSizeParent was changed in the ROTreemapLayout to ROPermissiveParent: What is the difference between the two? When should one be used and when the other? Cheers, Dennis On Wed, May 8, 2013 at 1:52 AM, Alexandre Bergel <alexandre.bergel(a)me.com> wrote: I think I have solved the problem. Would you mind to check please? This was tricky. The problem was due to the recent optimization I did: the rendering queue was modified while being iterated. Fortunately, I still had the implementation fresh in mind to solve the problem. Update to Roassal 1.351 Cheers, Alexandre On May 7, 2013, at 3:23 PM, Dennis Schenk <d.schenk(a)students.unibe.ch> wrote: -- _,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;: Alexandre Bergel http://www.bergel.eu ^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;. _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev _______________________________________________ Moose-dev mailing list Moose-dev(a)iam.unibe.ch https://www.iam.unibe.ch/mailman/listinfo/moose-dev