Re: [atl_discussion] What are the requirements for EMF?

Hi,

> As per my last email I'm trying to get ATL working with EMF
> and have fixed the ATL attached so that it now runs after a fashion.
>
> I can set a breakpoint and observe a match on the top Molecule,
> but I get no match on the inner Atoms. Is this because the
> EMF generantd XMI has omitted the schema-deducible types in XMI 2?
> Is it because the model should be below rather than at the root XMI?

When I execute your transformation, the M2M rule is executed once and
the A2A rule three times. It seems correct to me.
  
> Anyway, my result is just a typeless XMI directive without anything
> from the transformation.
>

I have the same problem here. I think this is a bug in the model
loading process under Eclipse. It appears when the source and target
metamodel is the same.

You can get your transformation to work by doing the following:
* rename the target metamodel to something like Chemistry2 in your atl file:
    - in the header: create OUT : Chemistry2 from IN : Chemistry
    - in the 'to' part of the rules: to out : Chemistry2!Molecule
* change the launch configuration to match the atl header and make
Chemistry2 point to chemistry.xmi.

This will have the transformation engine see 2 different metamodels
and bypass the bug.

We will try to correct it as soon as possible.

Thanks again for your help.

Regards,

Frédéric Jouault

           
Received on Sun Jan 09 01:03:22 2005