00:00 <headius[m]> that object would be very tightly scoped to just the method and the calls it makes, perhaps easier for escape analysis to eliminate

00:01 <headius[m]> this would also naturally fit into any upcoming value type support

00:02 <chrisseaton[m]> Thanks for the responses I need to run

00:03 <headius[m]> enebo, kares I'm going to go ahead with that fix for now since it will resolve the issue and avoid closing resources in the classloader

00:04 <headius[m]> we can't avoid issues from other classloaders being closed, however, unless we explicitly avoid caching... but that requires modifying every classloader access

00:04 <headius[m]> it's a nasty, nasty bug