[ https://issues.jboss.org/browse/JBIDE-21857?page=com.atlassian.jira.plugi... ] Max Rydahl Andersen commented on JBIDE-21857: --------------------------------------------- Okey, I now grok this issue better and I think it is either expected behavior or due to us handling the restart our self now. The issue boils down to (variation of [~fbricon] comment): 1) jvm started and debugger connected (all is good) 2) user make a method body change - debugger tells jvm - all is good. no restarts needed. 3) user make a structural change - debugger tells jvm - jvm does not support it. user is warned. (debugger and jvm is out of sync) 4) restart just the deployment - change is now visible (debugger and jvm is still out of sync) 5) do a method body chance - debugger tells jvm - debugger still out of sync, user asked to restart. #5 *could* just be our custom code for handling the restart on failed change that is not handling this case properly. In the past (1+ year ago) we didn't do that. Before we raise this to be a JDK limitation is there a way we can disable that custom logic and see if eclipse plain hotswap failure logic works ? [~rob.stryker] ? -- This message was sent by Atlassian JIRA (v6.4.11#64026)