Subj : Re: Function name...?
To   : netscape.public.mozilla.jseng
From : "Robert Bates" <rbates_removethis@austin.rr.com>
Date : Tue Jul 08 2003 11:36 pm

Great!  Exactly what I was looking for... I knew there had to be some way to
get back to the function object from within a native C call.  I had poked
around in the SpiderMonkey code and seen references to argv[-2] but didn't
know if that was after an internal argument array shift or not.

Thanks!

"John Bandhauer" <jband@netscape.com> wrote in message
news:beevc4$nmq2@ripley.netscape.com...
> Quick and dirty...
>
> The function object (as a jsval) is in argv[-2]. Use
> JS_ValueToFunction on it to get the JSFunction*. Then
> JS_GetFunctionName to get its name.
>
> If you don't want to work in terms of strings then you would tag
> the function object when you create it by setting a private
> property on the object using JS_SetReservedSlot (you can't mess
> with the 'main' private slot on a function object, but there are
> 2 available 'reserved' slots). See an example at:
>
http://lxr.mozilla.org/seamonkey/source/js/src/xpconnect/src/xpcwrappednativeinfo.cpp#144
>
>
> And then use JS_GetReservedSlot on the function object at call time:
>
http://lxr.mozilla.org/seamonkey/source/js/src/xpconnect/src/xpcwrappednativeinfo.cpp#44
>
> You can get the function object directly from argv[-2]...
>
>   JSObject* funobj = JSVAL_TO_OBJECT(argv[-2]);
>
>
http://lxr.mozilla.org/seamonkey/source/js/src/xpconnect/src/xpcwrappednativejsops.cpp#1257
>
> John.
>
> Robert Bates wrote:
>
> > OK, in pursuit of my Quick-n-Dirty C++ wrapper for JSObject, I am trying
to
> > figure out if there's a way (short of using the debug API) to get the
actual
> > method name invoked from the JSNative function call.  For example:
> >
> > JSBool c_method(JSContext *cx, JSObject *obj, uintN argc, jsval *argv,
jsval
> > *rval)
> > {
> >     JSString *jsMethod = ...
> >     ...
> > }
> >
> > Is there a way, or am I trying to do something not intended...?
> >
> > Brendan's made me aware that there are lost cycles if I don't do the
> > straight-away mapping, but the objects I'm exposing will only have these
> > methods called a couple times at most, and not iteratively, so
Quick-n-Dirty
> > is the name of the game right now...
> >
> > Thanks!
> >
>

.