Hi, Could somebody please review this fix for bug 4858370?
Bug: https://bugs.openjdk.java.net/browse/JDK-4858370 webrev: http://cr.openjdk.java.net/~sgehwolf/webrevs/JDK-4858370/webrev.01/ Testing done: jdk_jdi test group passes. Added regression test. There is a memory leak in the JDWP implementation where method parameters, method return values and the "this" object reference for constructor invocations are kept in memory via a global reference and get, thus, never GC'ed. The proposed fix deletes global references again in invoke_completeInvokeRequest(). The references got previously created in fillInvokeRequest() and invoker_invoke*() implementations. Thoughts? Note that I'd need somebody to sponsor the push to JDK 9 tree for me (once approved). Thanks. Cheers, Severin
