Subj : Re: (Spidermonkey) Converting a property name to a tinyId
To   : netscape.public.mozilla.jseng
From : zakalawe@mac.com (James Turner)
Date : Wed Mar 19 2003 02:41 pm

Brendan Eich <brendan@meer.net> wrote in message news:<3E77A8A9.8020505@meer.net>...
> > So
> >you can't write 'filtering' property ops, at least as I currently
> >understand it.
> >
> 
> Watchpoints are implemented as setter filters.  I'm not sure exactly 
> what you want to do, though.  Perhaps a concrete example would make 
> things clear.

At this risk of being laughed at a lot, I shall explain what I'm up
to. Keep in mind that, following your (Brendan's) clarification of
what's happening with propertyOps, I do now know a way to achieve what
I want, it's just a bit 'interesting'. Anyway, here is what I am up
to.

My objects have 3 kinds of properties - native/builtin ones, ones
inherited from their prototype, and ad-hoc ones. I need to be aware of
when any property has changed, and depending on the name, take action.
Notably, I need to mark the object as 'dirty', so that it's persisted
when the last reference is dropped. When I move to a distributed
object system, it's possible my jsobject will actually wrap a native
proxy which needs to marshal the property values over the wire, or I
may need to obtain an exclusive-write lock on the entity before
continuing.

So, as I said, I need a way to control *all* property accesses, to
achieve these kinds of behaviours. As i said, I can certainly achieve
this with some fiddling of how I define properties, but it would be
less than clean, so if there's a better way I'm interested.

Sorry this is dragging out rather seriously...
James

.