On 01/11/16 14:42, Jason Mehrens wrote:
The only other fix would be to collect all errors thrown (using addSuppressed) during close and throw only after the we have attempted to close all handlers. Then we would retain the behavior of always throwing errors.
Yes - that's also a possibility. Would that satisfy the issue with JavaMail as described in https://bugs.openjdk.java.net/browse/JDK-8152515 ? It is still unsure what should be done for OOME in this case though. I think I might still be attracted by the solution that brings the least code & behavioral change - so let's see if we can get away with special-casing shutdown to trap errors... best regards, -- daniel
