Re: misc. properties

[Tim Janik <timj gtk org>]

On 2 Nov 2001, jacob berkman wrote:

> On Fri, 2001-11-02 at 15:05, Havoc Pennington wrote:
> > 
> > My main worry is that we add the extra notify signal on all buffer
> > changes, and also connect a handler from view to the buffer's
> > "changed" signal (having a handler immediately slows things down). So
> > that's a bit of overhead added to every buffer mutation.
> 
> there is only a notify when it is changed via the property, right?

no, if change notification doesn't occour everytime the contents actually
changed, your code is buggy. that is, for any situation where
  g_object_get(o,"Prop",&X,NULL);
  [...]
  g_object_get(o,"Prop",&Y,NULL);
it's true that X!=Y, you either should have gotten notification before
getting Y, or changed ontification should be pending, regardless of
how/what caused the property to change.
note also that it is _not_ an error to receive change notification, even
if X==Y.

> 
> jacob

---
ciaoTJ