Re: weak pointers, cbdata, refcounting

From: Henrik Nordstrom <hno@dont-contact.us>
Date: Fri, 26 Aug 2005 21:38:00 +0200 (CEST)

On Mon, 22 Aug 2005, Alex Rousskov wrote:

> (*) For example, multiple inheritance does not seem to be supported by
> cbdata due to (void*) casting:

Not only the cast. cbdata is really designed for simple structs and
depends on being able to find the cbdata info from the object pointer. In
multiple inheritance this isn't so easy as the object pointer in an
inherited method may be the sub-object of the inherited class..

It has however for a long time been a desireable goal to unify refcounting
and cbdata, keeping only a single concept.

So in short I am all for adding an invalidation concept to the refcount
class we have today, perparing for the migration away from the current
C-style CBDATA into a more OO-friendly design. Then perhaps move on to
have weak pointers ontop of this.

Regards
Henrik
Received on Fri Aug 26 2005 - 13:38:03 MDT

This archive was generated by hypermail pre-2.1.9 : Wed Aug 31 2005 - 12:00:06 MDT