I first assumed you must be doing something
wrong, since I've installed it from that site many times before.
But instead of saying that, thought
maybe I should try it first ... and sure enough, something seems to have
I opened bug 417208 for that issue.
So if no quick response on bug 417208,
you might try being an early adopter of the kepler version, using a "near
SR1 version" of Eclipse SDK (in fact, RC4 version will be available/announced
in a few hours ... which is expected to be our final SR1 version if no
"blockers" found), and use the "kepler version" of
Idrissa DIENG <idydieng@xxxxxxxxx>
Cross project issues
09/13/2013 11:54 AM
Be sure to update b3 aggregator editor, or may see error about "Feature
'allowLegacySites' not found"
what is the recommended b3 version to use ? because it seem that the b3
Juno's version cannot be installed in Eclipse Juno SR2 & Eclipse Kepler
Le 18/03/2013 03:05, David M Williams a écrit :
The wiki  was recently updated to
direct Simrel participants to use Juno SR2 and add the aggregator using
the the "4.2" stream of the b3 aggregator. (I just noticed a
few days ago, myself, that it was out of date). http://download.eclipse.org/modeling/emft/b3/updates-4.2/
The reason this is important is that a new feature was recently added to
the aggregator's model, that we will take advantage of, where you can set
"allowLegacySites" (the default is true, to match existing default
behavior, but we set it to false for our SimRel aggregation). "Legacy"
in this case, refers to the old update manager site format. As I'm sure
you all know, it has long been a written rule that participating projects
provide "optimized p2 repositories", but now the tool will do
a little more checking for us. There has been a few times in the past,
when a project mistakenly published a site such that only the "site.xml"
file was provided, not a "content.jar/xml" file so this will
simply help catch those few mistakes as soon as they occur. (p2, and the
aggregator, handle both types of sites, but the site.xml update manager
style is less efficient and less rich).
But, it does require everyone have a current version of the tool, or else
your own "verifications" in your own b3 aggregator editor, will
fail on our Kepler model since we now specify that attribute. (That is,
it will fail right away, just because it doesn't recognize that attribute
of the model, not that there is anything wrong with your contribution).
Thanks, (and, sorry for the short notice).
And, thanks to Thomas Hallgren for adding the feature. (bug 389296)