I have not tried out the scenario, but it looks similar to bug
https://bugs.eclipse.org/bugs/show_bug.cgi?id=421706 which was causing an
explosion in the felix resolver for the number of solutions (Candidates)
that are being tried to find a consistent class space.  It would be
interesting if you could try this against one of the latest I-Builds.

Tom





From:   Mickael Istria <[email protected]>
To:     [email protected],
Date:   12/04/2013 08:08 AM
Subject:        [equinox-dev] Strange Heap overflow with Luna M3
Sent by:        [email protected]



Hi all,

I think we've met an interesting issue in Equinox/Felix while preparing
migration of JBoss Tools to Luna. In some circumstances, the target
application will flood the Heap by adding and retaining many
org.apache.felix.resolver.Candidates in the main Thread.
This error doesn't seem to happen when starting Eclipse SDK, memory seems
fine. However, we reproduce it everytime we use our target-platform, which
contains Luna M3 + some additional stuff.
We're tracking the issue here: https://issues.jboss.org/browse/JBIDE-15807
, but I start to believe that we're facing a potential bug in
Equinox/Felix. So I'd be glad if you could give some hints.
Here are the steps In order to reproduce it:
* Get the following target platform in your workspace
https://github.com/jbosstools/jbosstools-target-platforms/blob/4.40.x/jbosstools/multiple/jbosstools-multiple.target

* Enable this target platform
* Try to run Eclipse workbench (using this target platform).
--
Mickael Istria
Eclipse developer at JBoss, by Red Hat
My blog - My Tweets_______________________________________________
equinox-dev mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/equinox-dev

<<inline: graycol.gif>>

_______________________________________________
equinox-dev mailing list
[email protected]
https://dev.eclipse.org/mailman/listinfo/equinox-dev

Reply via email to