Remy Maucherat wrote: > > I did examine the Eclipse code, and it is invalid (you must > not attempt > to access other CL controlled resources based on another URL > which was > returned by the CL). > Hi Remy,
What you say sounds reasonable to me as a user of the eclipse code, looks like the developers of eclipse emf did not think of the problems such an access method can cause. Unfortunately EMF is a base for other APIs like XSD ... > However, I did add a workaround in the CL which makes your test case > work, but which has a performance penalty on first access to a JAR. I > think avoiding resource leak is more important. > Thanks a lot !! I agree. Best regards, Jan