Community
Participate
Working Groups
Created attachment 223805 [details] Reproducable test cases Not mark a model as "modified" just because you switch to another diagram. You know that you haven't modified the model, but a user can not be sure about it. So... You are asked to save this unmodified model, and thereby creating a new version of it. (new timestamp) The problem is the same in Indigo and Juno versions of Papyrus. Attached are workspaces with backup projects in them, plus a ppt slide.
In the sprint planning presentation 11-Feb you refer to Bug 398712. This one is also related to the newly submitted Bug 400809. The "modified model" indication also happens when you open a new diagram or close a diagram, when you stack diagram windows etc. According to Bug 400809: Such private user information should not even be stored in the model.
Current status: - Opening/Closing a page (Table, Diagram, ...): Marks the model as dirty. Related to Bug 400809. - Switching tab: Does not mark the model as dirty anymore. - Manipulating empty tables: Model is dirty. It can be saved, but will remain dirty. This is a bug from the Tables, where a new command is triggered after each save action. I think this is an EMF Facet bug. However, this is a deprecated component and it probably won't be fixed. Closing the empty table (Or adding elements in the table) will fix the issue. - Opening SysML Diagrams: there seems to be a "version migration" when SysML IBD diagrams are opened, which marks the model as dirty. The migration action is always executed, while it should probably be executed only once. I will open a specific bug for this one. There may still be a few actions which mark the editor as dirty whereas they should not. They need to be identified.
I have tested this in Kepler RC4. Switching to another diagram does not "touch" the model. (= need to save it) ...but closing a diagram or opening one do! This problem is described in Bug 400809. ...and this is related to the 3-file Papyrus model discussed recently.
> I have tested this in Kepler RC4. > Switching to another diagram does not "touch" the model. (= need to save it) > ...but closing a diagram or opening one do! > > This problem is described in Bug 400809. > ...and this is related to the 3-file Papyrus model discussed recently. This is now fixed (Details in Bug 429239). I leave this bug opened because of subtask Bug 403432
..and still open because of Bug 421411.
Remaining subtasks have been fixed