Issue 698 in moose-technology: Mondrian does not properly raise MOElementSelection

Status: New
Owner: ----
CC: [hidden email]
Labels: Type-Defect Priority-High Component-Mondrian Milestone-4.6

New issue 698 by [hidden email]: Mondrian does not properly raise  
MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

When in Glamour, often when selecting a Mondrian element does not populate  
the selection because the MOCanvas does not raise MOElementSelection.

The reason for this seems to be a race condition due to changes in the  
announcement framework. In the new announcement framework from Pharo 1.3,  
the order of announcements is not guaranteed.

The issue seems to be due to the logic of the announcement:

MOCanvas>>mouseUp:
..
(lastEnteredGraphElement isNil not and: [
                (lastEnteredGraphElement class ~~ MORoot and: [
                        lastEnteredGraphElement isNotSelected ])])
                                ifTrue: [
                                        selectionEvent := MOElementSelection event: anEvent on:  
lastEnteredGraphElement.
                                        anEvent commandKeyPressed ifTrue: [ selectionEvent setAsMultiple ].
                                        self announce: selectionEvent.
                                        anEvent wasHandled: true.
                                        ^ self  ].

Sometimes, "lastEnteredGraphElement isNotSelected" returns false because  
somehow the element gets selected beforehand. This is most likely related  
to the fact that the selection of a node actually happens by handling the  
same MOElementSelection announcement:

MOCanvas>>announce: anAnnouncement
        ...
        anAnnouncement element isNil ifFalse: [
                anAnnouncement element announcer announce: anAnnouncement ].
        self announcer announce: anAnnouncement.

MOViewRenderer>>setUsefulHandlersForNodes
        ...
        self interaction on: MOElementSelection
                do: [:ann |
                                ann isUnique
                                        ifTrue: [ self root allSelectedNodes
                                                                do: [ :node |
                                                                                node
                                                                                        deselect;
                                                                                        announce: (MOElementDeselection new) ] ].
                                ann element select.
                                self updateWindow ].


_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #1 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

Any idea how I can reproduce the problem? Best would be to have a yellow  
test

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #2 on issue 698 by [hidden email]: Mondrian does not properly  
raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

I do not know how to write the test case.

To reproduce it:
- open a MoosePanel
- select a group of classes
- select the system complexity tab
- select a couple of classes. You will see that in many cases, the figure  
gets selected and it usually moves, but the next pane is not updated.  
Sometimes, it works though, that is why you have to click on multiple  
classes to see the problem.

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #3 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

This is an important issue for 4.6. Alex, do you have time to look into it?

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #4 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

I will have a look at it. But it may not be easy to fix this. I will work  
on it asap even if several deadlines are approaching. I will let the  
mailing list know about my progresses

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #5 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

This is not reproducible in Pharo 1.4. Apparently, the problem comes from a  
shift by -1@0 when I click (it goes through MOCanvas>>mouseOver:)
Really weird behavior.

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #6 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

Is this problem still around?

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #7 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

The problem is still around with the Jenkins VM, but works fine with the  
Elliot's Cog VM.

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #8 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

did you check the event that is generated?
If you look in eventModel on Squeak source, you will see that there is  
listener that print event on the transcript now since this is VM code and  
you will not see anything in the image.
So bad idea. Did you contact esteban?

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Hi Esteban,

We have a serious problem with Mondrian, and we believe this comes from the way events are handled by the virtual machine. When drag and dropping a morph, it happens that the morph is dragged by 1 pixel to the left. As a short test, do a "Morph new openInWorld", grab the morph and when you release it by pressing the left mouse button, it moves by 1 pixel.

The problem is still around with the Jenkins VM (CogVM), but works fine with the Elliot's Cog VM.

Any idea where this comes from?

Cheers,
Alexandre

On 4 Oct 2011, at 02:44, [hidden email] wrote:

--
_,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:
Alexandre Bergel  http://www.bergel.eu
^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;._,.;:~^~:;.





_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Hi Alex,
I don't understand well the problem, or I can not reproduce it... I'm doing what you say and I don't see a problem there.
Said so... carbon and cocoa event handling procedures are complete different, so... yes, it is possible that there are some glitch somewhere... if you can provide me some more info, I can continue digging into it.

cheers,
Esteban

El 04/10/2011, a las 8:37a.m., Alexandre Bergel escribió:


_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Hi,

It is indeed a bit erratic. At the latest, I can show it to you first hand at Smalltalks.

Cheers,
Doru

On 4 Oct 2011, at 21:08, Esteban Lorenzano wrote:

--
www.tudorgirba.com

"Every thing should have the right to be different."




_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Updates:
        Labels: -Milestone-4.6 Milestone-4.7

Comment #9 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

(No comment was entered for this change.)

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Updates:
        Labels: -Milestone-4.7

Comment #10 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

I have the feeling we will not fix this issue, so I am removing the 4.7 tag.

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Comment #11 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

We should try to move to Roassal.

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev

Updates:
        Status: WontFix

Comment #12 on issue 698 by [hidden email]: Mondrian does not  
properly raise MOElementSelection
http://code.google.com/p/moose-technology/issues/detail?id=698

(No comment was entered for this change.)

_______________________________________________
Moose-dev mailing list
[hidden email]
https://www.iam.unibe.ch/mailman/listinfo/moose-dev