Subj : Re: Rhino line counting
To   : rickr_mc@lycos.com
From : Igor Bukanov <igor@fastmail.fm>
Date : Wed Sep 17 2003 06:07 pm

It was fixed in Rhino CVS tip, see
http://bugzilla.mozilla.org/show_bug.cgi?id=217951

Rick Romero wrote:
> I'm trying to use the line counting functionality in Rhino. I know
> about the setInstructionObserverThreshold, and how to extend the
> Context in order to trap these events. I believe, though, that there
> may be a bug in the current source. Check out Interpreter.java:2132-38
> 
>     case TokenStream.CALL : {
>         if (instructionThreshold != 0) {
>             instructionCount += INVOCATION_COST;
>             cx.instructionCount = instructionCount;
>             instructionCount = -1;
>         }
>         cx.instructionCount = instructionCount;
> 
> Note that after this code is executed, (cx.instructionCount == -1). It
> appears that this last line is incorrect. It essentially resets the
> instructionCount.
> 
> On a related note, does anyone have any idea what the intention here
> was? This switch case, TokenStream.NEW and TokenStream.CALLSPECIAL all
> have the same sort of thing happening with the INVOCATION_COST. Anyone
> out there able to explain the rationale here? I haven't spent too much
> time trying to figure out the programmer's intent, and would
> appreciate it if someone in the know could provide some insight into
> this section of code.
> 
> Thanks,
> 
> Rick Romero

.