|
|
|
|
|
|
|
|
|
|
|
|
|
Re: [ATL] how to extends ATL language with own library ? [message #536421 is a reply to message #536339] |
Fri, 28 May 2010 07:26 |
Sébastien Gandon Messages: 184 Registered: July 2009 |
Senior Member |
|
|
Sylvain EVEILLARD wrote on Thu, 27 May 2010 19:04 | Well you still have the possibility to post-process the transformed model to finalize it with the things not doable in ATL.
|
I don't know exactly how this would be possible because you end up with one model with as not relation with the input model.
I do not forsee anyway to match the instances of the ouput model with the ones in the input model.
Anyway merci Sylvain.
[Updated on: Fri, 28 May 2010 07:27] Report message to a moderator
|
|
|
Re: [ATL] how to extends ATL language with own library ? [message #536438 is a reply to message #536421] |
Fri, 28 May 2010 09:11 |
Sébastien Gandon Messages: 184 Registered: July 2009 |
Senior Member |
|
|
Ok,
After a good night sleep I decided to keep on testing this conversion process.
And guess what I managed to make it work.
That was due to a change that I made previously in the input metamodel to work around this problem.
But the conclusion is that, if you have a input field of type int and an output field of type long, you can create an operation to convert on type to the other. It worked.
Another conclusion is that if you hold a copy of the input metamodel you can change the input field type from int to long and the problem is resolved without any conversion
But I just realized that I tried that before and it did not work, I now know why, the ATL Regular VM fails but the EMF VM succeed in using Long-Long. Don't ask me why.
Thanks any way because i'll use the delagate technique for some other processings anyway.
[Updated on: Fri, 28 May 2010 09:26] Report message to a moderator
|
|
|
Powered by
FUDForum. Page generated in 0.05047 seconds