This is a continuation of a review from the hotspot repo: http://mail.openjdk.java.net/pipermail/hotspot-runtime-dev/2013-October/009657.html The hotspot fix has now reached JDK8-TL so I can now push the test and JDK fix.
The problem has to do with stepping into the String constructor and asking to see the value of the String. JDWP does this via GetStringUTFChars which crashed if the String didn't have a value field. That call now returns NULL and needs to be checked in JDWP. I have also added a test that breakpoints in String's constructor and verifies that it is ok to get the value of the String. webrev: http://cr.openjdk.java.net/~sla/8021897/webrev.00/ bug: https://bugs.openjdk.java.net/browse/JDK-8021897 Thanks, /Staffan
