Other |
---|
tbdResolved |
Duplicate :
|
"When AttachCurrentThread() fails you should not trust the value in the penv argument, even though I could not actually find a relevant passage in the JNI "spec" book which states that to be the case (it does include code snippets that promote that use, but confusingly also snippets that do not check the return value before dereferencing the penv argument)." This should be cleaned up.