P4
- JPLISAgent.c L783 - need to handle JVMTI_ERROR_NOT_AVAILABLE so that we don't abort if the capability isn't available (not an issue for our implementation of course).
- JPLISAgent.c L1319 - do you need to check if originForRelease is NULL?
- does setNativeMethodPrefix implementation need any internal synchronization? What if two agents and they both attempt to set a prefix around the same time?
-This is too late for the CCC but I wonder if we should have introduced a new attribute in the JAR manifest to indicate that the agent needs to ability to set the native method prefix. Something like "Can-Set-Native-Method-Prefix" is ugly but it would be consistent with Can-Redefine-Classes. I assume in some VM implementations the capability might need to be added in the onload phase so this would be the indication that the capability is needed.
###@###.### 2005-05-24 05:43:13 GMT
- JPLISAgent.c L1319 - do you need to check if originForRelease is NULL?
- does setNativeMethodPrefix implementation need any internal synchronization? What if two agents and they both attempt to set a prefix around the same time?
-This is too late for the CCC but I wonder if we should have introduced a new attribute in the JAR manifest to indicate that the agent needs to ability to set the native method prefix. Something like "Can-Set-Native-Method-Prefix" is ugly but it would be consistent with Can-Redefine-Classes. I assume in some VM implementations the capability might need to be added in the onload phase so this would be the indication that the capability is needed.
###@###.### 2005-05-24 05:43:13 GMT