Community
Participate
Working Groups
Hi all, While I was writing the Spanish book chapter on EuGENia, I noticed a few minor inconsistencies in the EuGENia workflow: 1. We have a menu entry called "Synchronize EMF gen model", and another called "Synchronize GMF GenModel". How about using "GenModel model" and "GMFGen model", instead? 2. The only way to invoke the "Synchronize EMF gen model" action is from the drop-down menu in Eclipse. The all-in-one "Generate GMF editor" doesn't call it, and for that reason, the Ant task doesn't either. Should we call it as well? 3. The ECore metamodel is exposed as "ECore" in some polishing transformations, and as "Ecore" in others. Should we use "ECore" everywhere? 4. The ECore metamodel is stored again in GenerateToolGraphMapDelegate. Shouldn't we set store=false in that case? 5. I already exposed with read=true and store=false the GmfGraph model in the FixGMFGen step. While we're at it, perhaps we should expose the other GMF models in the same way? Except for step 3, I think most of these changes are safe. What are your thoughts on this?
#1 sounds OK #2: yes - that's an omission in the current version #3 could break existing code. We should probably use both (ECore could be an alias so that we preserve backward compatibility) #4 rings a bell as something that was deliberate at the time - although I must admit I don't quite remember what the issue was when store was false #5 sounds good to me
I'll work on #1-#3 and #5 then, and leave #4 as is.
I have implemented #1-#3 and #5. I'll wait until we get this into an interim release to update the Eugenia article in the website.
This has been fixed in the latest interim version. Antonio: it should now be OK to update the Eugenia article accordingly.
Fixed in 1.0
