On 20 Oct 2007, at 10:23 , Stéphane Ducasse wrote:
I see. Adrian how do you detect the problems? running an OCL constraints interpreter?
this is good that you use now constraints.
Zuviel der Ehre :)
Its all hardcoded right now, just hacked in the constraints of EMOF that apply to the subset we actually use (see http:// smallwiki.unibe.ch/moose/mse/fm3meta-metamodel)
For Smalltalk, there is an unused Constraint class, which duplicates RB's Condition class, that could be used to write model constraints, written in Smalltalk and looking like OCL :) Would be nice to do.
cheers, AA