by lonniev@gmail.com » Mon Feb 18, 2019 10:54 am
Saulius,
I like that 2009 claim of yours that one can run the QVTO transforms within the current MD editing session. I'm curious, though, how that is actually done. The provided sample (the simple UML to RDB one) does not perform its transformation on the current model but, rather, loads an external model file, transforms that, and writes the in-memory generated output model to an external model file.
I understand that the special string "md:model" tells the QVTO script handler to pass the current MD model to the transformation. This works and I can observe (via logging) that my transformation runs. However, if the transformation's out model argument is also set to "md:model", I do not see the current model update with the artifacts that the transformation appears to be creating.
Is there another sample somewhere (back in the LT labs, maybe) where a QVTO script, run directly from within the Macro Engine, without a wrapping JS (or Groovy or Ruby or Python) script, generates new model content for the current model?