| [news.eclipse.tools.emf] Re: EMF performance |
Romain,
Comments below.
I don't think I understand the question. What two things are you comparing? Reflection delegates through the generated accessors so it will be slower than direct access by definition.
Hi
I would like to know, between the EObject reflective API and the Java implementation of a Ecore metamodel, which one use the less memory and/or is the fastest.
Yes, they're not very useful anymore...
I looked at the org.eclipse.emf.test.performance project on the CVS but it seems deprecated/obsolete.
I'd suggest you measure. A generated implementation will be faster than a dynamic one and reflection is an additional layer that adds cost...
I searched on the internet and can't find really useful informations.
If somebody can help :)