| Re: [papyrus-rt-dev] Textual syntax for UML-RT merged into master |
Since Xtext generates the appropriate tooling (outline, formatting, code-assist), then that may work for combined languages. Of course, it is unlikely to be as mature as the JDT or the CDT, but I don't know if it is at all possible to somehow "invoke" the JDT or CDT formatting, code-assist and other capabilities. for specific sections of the same source file inside the same editor. I'll look into it. My guess is that such "mixed" editor is not possible, so we could only have limited support for other languages within the same editor, and if the user really wants the full support of JDT or CDT, then he could write such action code in separate file(s) and the .umlrt file will have references to those files.The idea of combining the capabilities and support for the core language and an action language within the same editor is very ambitious and it would be incredible, but I think that should be a bit of a longer term goal. Even Papyrus doesn't provide any language aware support in the OpaqueBehaviour editor.
I agree that formatting and completion support for action languages are highly desirable. However, I would argue that if an action's code is sufficiently long and complex, it would be better to put it into a separate file and the action in the model would refer to that external file. I have actually done this, with PapyrusRT and with another tool I used a while ago. Putting complex action code in separate files improves readability and navigability. If we promote that approach, then the need for complex formatting and code-completion within the same editor is greatly reduced, as we delegate that task to the appropriate editor in CDT or JDT.