Re: [Gnash-dev] GC: status report & call for test

[strk]

On Fri, Jul 06, 2007 at 09:21:51PM -0600, Eric Hughes wrote:
> At 08:41 AM 7/5/2007, strk wrote:
> >Users (ActionScript coders) can still change this relationship.
> >a function's "constructor" member is a pointer to self, but
> >might be changed to point to something else:
> >
> >        A = function() {}
> >        B = function() {}
> >        A.prototype.constructor = B.prototype.constructor;
> >        delete A;
> 
> I had commented briefly about this situation a while ago, when it was 
> certainly premature then.  My proposal to this would be to initialize 
> ".constructor" members with a weak pointer.  The assignment in line 3 above 
> would overwrite that weak pointer with an ordinary GC pointer.  Thus in the 
> ordinary case, reference counting would destroy a constructor member when 
> its function goes out of scope, but in the general case the GC would catch 
> it.
> 
> The reason that this may still be premature is that it presumes a proper 
> managed_ptr class with multiple variants hiding behind it (say, gc_ptr, 
> rc_ptr, weak_ptr) so that this upgrade in capability happens silently at 
> the point of assignment.

We're still talking about non-intrusive flavors of this pointers I guess,
which would mean reviewing the whole code base to *never* pass dumb
pointers around, right ?

--strk;