Harm,
While I await the the results of your testing, my understanding from
talking to Ed and Nick recently and listening to their explanations of
this change to the Councils back in June and again in October was that
the runtime will continue to work on JVM 1.4. So as long as you do not
regenerate your models, the code will continue to work on JVM 1.4
without change or reduced functionality. I also assume, but have not
verified, that you could always re-generate your models using an older
EMF and thus continue to use JVM 1.4 with newer models.
The part that puzzles me most is how our process broke down - the EMF
team clearly brought this JVM 4/5 issue forward over seven months ago
so why is it suddenly a surprise and a blocking issue for Europa? I'd
like to resolve that mystery so that we, collectively, don't have
similar problems from other known-in-advance project interactions.
- Bjorn
Harm Sluiman wrote:
I believe you have it correct Bjorn.
The concern is that given that many of the higher level tools (TPTP,
WTP,
BIRT) use EMF, the "reduced functionality" may be rather significant.
It seems that with a common component like EMF requiring 5 the
majority
of configurations above base JDT will require 5. This can make the out
of the box experience for a default 1.4 user problematic.
|