> On Jun 2, 2016, at 3:11 PM, Daniel Fuchs <[email protected]> wrote: > > Hi Mandy, > > Looks good to me. Though I wonder whether a better fix would be > to not tweak the bci for native methods in the native call that > sets the value to 0 (I assume this is what happens somewhere > since bci is initialized to -1 in Java). >
I think you are suggesting to do the fix in the VM not to set StackFrameInfo::bci if it’s a native method and hence it’d be left as -1. I would prefer to do as much work in the java side as possible. Mandy > best regards, > > -- daniel > > On 01/06/16 20:46, Mandy Chung wrote: >> http://cr.openjdk.java.net/~mchung/jdk9/webrevs/8157892/webrev.00/ >> >> This webrev addresses two issues [1][2]. >> >> It’s a simple fix to return proper filename that may be available for native >> method and fix getByteCodeIndex to check for a native method and returns >> proper value. >> >> Mandy >> [1] JDK-8157892: StackFrame::getFileName returns null when a source file >> exists for native methods >> [2] JDK-8157977: getByteCodeIndex method from StackFrame does not return >> negative number when StackFrame is a native frame >> >> >
