[news.eclipse.tools.emf] Re: [ATL] OCL checker for EMF models in transfo

[Frédéric Jouault <frederic.jouault@xxxxxxxxxxxxxx>]

Hello Damien,

Firstly, I added [ATL] as a prefix to the subject of this post because I 
am answering from the ATL point of view :-).


Secondly, here is a possible approach to do what you described:

1- Write the OCL constraints as a transformation (e.g., in ATL). Source: 
Ma, target: a diagnostic (e.g., a model containing the list of failed 
constraints). Examples: 
http://www.eclipse.org/m2m/atl/atlTransformations/#ATL2Problem, 
http://www.eclipse.org/m2m/atl/atlTransformations/#KM32Problem,
http://dev.eclipse.org/viewcvs/indextech.cgi/org.eclipse.gmt/org.eclipse.gmt.tcs/dsls/TCS/WFR/TCS2Problem.atl?view=markup


2- Use a query to check if there are errors.


3- Run the actual transformation if there is no error.


2 and 3 are for instance used in the TCS compiler, at lines 122 to 142:
http://dev.eclipse.org/viewcvs/indextech.cgi/org.eclipse.gmt/org.eclipse.gmt.tcs/scripts/AMMAScripts/build.syntax.xml?view=markup

Note that because we use TCS to parse the source model, we can apply 
markers for the detected errors using the am3.applyMarkers task.


The TCS compiler that I mentioned above is implemented using the AM3 ant 
tasks:
http://wiki.eclipse.org/AM3_Ant_Tasks

Note that it is also implemented as an Eclipse plugin. This means that 
you can also implement such a scenario in Java.


Regards,

Frédéric Jouault


Ed Merks wrote:
> Damien,
>
> Comments below.
>
>
> Damien Thivolle wrote:
> > Hello,
> >
> > I'm relatively new to Eclipse and the modeling technology. I want to 
> > write a transformation from a meta-model MMa to a meta-model MMb using 
> > ATL. The problem is that the transformation will only concern a subset 
> > of MMa.
> There's a newsgroup for ATL.  I've included it on the "to" list of the 
> reply.
> > I've been hinted to write a checker in OCL that would just tell me 
> > whether a model Ma (conforming to MMa) is included in the subset I'm 
> > interested in.
> >
> > In the end, if the checker answers true when asked if the input model 
> > is correct, the transformation would take place, if not an error 
> > message would be yielded.
> >
> > I've been looking for some clues as to where to start and I was 
> > expecting to find something suitable for a newbie like me. Like, I 
> > define a file with a set of OCL constraints, right click on my model 
> > and choose "Validate constraints from file...". But I guess it's not 
> > gonna be that easy.
> >
> > More generally, I'm having a hard time finding my way around Eclipse, 
> > EMF and such... I've been reading all the material I could find about 
> > using OCL to check constraints on EMF models, and it's like Chinese to 
> > me.
> Note really.  If it were Chinese it would use an alphabet you'd not 
> recognize.  :-P
> > For example, I've seen a lot of EMF models with annotations named 
> > "OCL" and having several entries, but I can't really find anything 
> > that would explain me what to do with those.
> Did you look at the Eclipse help?
> > I'm just saying is that I find it hard to understand this technology 
> > and I'm wondering if it's normal or if this is where my comprehension 
> > reaches its limits.
> There's an awful lot of technology to learn all at once for sure.  And 
> the documentation is sometimes a little lacking or hard to find...
> > Best regards,