Home » Modeling » Papyrus » Compare of Notation file
| | | |
Re: Compare of Notation file [message #1071823 is a reply to message #1071165] |
Sun, 21 July 2013 08:49 |
Marc-Florian Wendland Messages: 83 Registered: January 2013 |
Member |
|
|
Hi Camille,
maybe I do not understand, but what do the di-files (the file that contains
the specification of the editor's state) have to do with the notation file,
except that they refer to the diagrams? If one adds/moves something in a
diagram, the di-file will remain unchanged and synchronization with the
repository will not show any differences, even though the notation-files
might have some conflicts. And conflicts happen rather often in the notation
file.
Or do you mean the "virtual" Papyrus node in the Resource Explorer?
I can tell you what I did, so that you could perhaps guide me:
1. Synchronize with repository on the "virtual" Papyrus node in the Resource
Explorer/Navigator.
2. Switch to synchronization perspective
3. The synchronization view shows changes in the UML and the Notation file
4. Opening the compare editor on the UML file works as expected
5. Opening the compare editor on the Notation file does not work, i.e., the
editor remains empty.
I can switch between Papyrus and textual comparison and the textutal
comparison works, althought this is not very helpful.
I have installed EMF compare, GMF compare, UML compare and Papyrus compare.
Any ideas? Did I do something wrong?
Regards,
Marc-Florian
"Camille Letavernier" wrote in message
news:ksbf2m$i9m$1@xxxxxxxxe.org...
Hi,
I think the graphical comparison is triggered only when you compare two DI
files; not when you compare two notation files.
Camille
|
|
| |
Re: Compare of Notation file [message #1077878 is a reply to message #1072178] |
Fri, 02 August 2013 08:56 |
Marc-Florian Wendland Messages: 83 Registered: January 2013 |
Member |
|
|
Hi Camille,
I still was not able to open the graphical compare for the Papyrus notation
file. Again, I did click the di-file and compared them with each other. But
it only displayed the differences in the di-file, but no graphical
comparison.
So, I confirm that semantic comparison on UML models works for the team
provider (synchronization view) and for 'compare with'->'each other'.
Semantic comparison on the di-file works as well for both cases.
The notation file comparison does not work for either cases.
I would highly recommend/appreciate if a tighter integration with the team
provider will be tackled for SR1. I find it very important for Papyrus to
support versioning and conflict resolving using a SVN/GIT repository.
Actually, I think this is a crucial functionality for doing serious modeling
tasks with Papyrus.
Finally: I read in the Eclipsepedia that the Papyrus EMF Compare Integration
was necessary, because GMF compare was build on the assumption that each
diagram is stored in its own file/resource, right?
Regards,
Marc-Florian
"Camille Letavernier" wrote in message
news:ksir8g$tl2$1@xxxxxxxxe.org...
Hi,
The EMF Compare 2/Papyrus integration is still experimental, so I guess it
just doesn't work for your case. However, I'm not sure what exactly is
supported, or supposed to be, and what still needs to be done.
What should work is:
- Semantic comparison on UML models (Select two or three UML models and
"Compare with each other")
- Graphical comparison on DI models (Same thing)
I think there is also an integration with the Team logical model, which
means that when you commit a *.di file, the associated *.uml and *.notation
files are included in the commit.
We plan to have a stable version of the EMF Compare 2 integration in Kepler
SR1 (0.10.1).
Regards,
Camille
|
|
|
Re: Compare of Notation file [message #1108271 is a reply to message #1077878] |
Fri, 13 September 2013 15:59 |
Camille Letavernier Messages: 952 Registered: February 2011 |
Senior Member |
|
|
Hi Marc-Florian,
I've discussed with Mikaël Barbero (From the EMF Compare team). So here's the status:
- EMF Compare is properly integrated with Team
- EMF Compare is properly integrated with Papyrus (It supports the Papyrus logical model)
- EGit is not properly integrated with Team yet (But most patches are ready). Especially, it doesn't support the Team logical model yet.
However, comparing two DI or two Notation files should trigger the graphical comparison (If there are graphical changes). I've tested it again this week and it worked fine. The only issue I had is when I changed my content-type settings in Eclipse (Because Team relies on that to find the appropriate merger).
Quote:Finally: I read in the Eclipsepedia that the Papyrus EMF Compare Integration
was necessary, because GMF compare was build on the assumption that each
diagram is stored in its own file/resource, right?
Maybe in EMF Compare 1.x. In the current version (Kepler), the EMF Compare/Papyrus integration does two things:
- Implement the Papyrus logical model (If you compare two notations files, it also includes the di files in the comparison. If you try to commit the notation file, it also commits the uml file for consistency).
- Implement the Specific difference dependencies in Papyrus. Most dependencies can be computed from the Model/Metamodel, but a few ones are specific to Papyrus.
We're also planning to include Papyrus-specific labels (Diagram icons, Stereotype icons and labels, etc.) in the compare tree, as well as the Papyrus CSS support during the comparison.
Most things which were done in Papyrus in the 1.X version are now implemented directly in EMF Compare (Specific diff dependencies for UML, Undo/Redo, Usability actions such as Expand/Collapse all, ...). The integration in Papyrus is now really thin (Which is a good thing).
Quote:I would highly recommend/appreciate if a tighter integration with the team
provider will be tackled for SR1. I find it very important for Papyrus to
support versioning and conflict resolving using a SVN/GIT repository.
Actually, I think this is a crucial functionality for doing serious modeling
tasks with Papyrus.
Definitely. But this is a huge work. And Mikaël is really doing great here. Just be patient The Kepler SR2 version should be really great, and the Kepler SR1 version is already very impresive.
Camille
Camille Letavernier
|
|
|
Goto Forum:
Current Time: Mon Sep 23 08:38:03 GMT 2024
Powered by FUDForum. Page generated in 0.03322 seconds
|