Hi Marcus,
On Wed, 2005-05-04 at 14:03 +0200, Marcus Lindblom wrote:
> Problem solved, and of course it was the simplest thing.. again.. sigh..
>
> I was neglecting to do CPEdit(chunkmaterial, chunksfieldmask), although
> I'm currently running single-threaded and all. (It worked without that
> in many other places, so I didn't think of that until now.)
Gah! That's the first thing I though of, but it was too trivial to
mention... :(
> This write-locking via objects is a bit user-error-prone. I wonder if
> there was some better way to do it.
Yup.
> Perhaps forcing (or at least providing a method that allows) access to
> the object to happen between a begin/end-pair via a function on a FCPtr
> that returned some smart-pointer/lock object which allowed access to the
> non-const member functions of a field container.
...
>
> Perphaps you've already thought about something like this?
Indeed. That's going to be the major change for 2.0. Given that the
access to all variables is filtered via accessor methods anyway they
will keep track of changed fields. The only place where this is
performance-critical is for large multi-fields like vertices and
normals. For those the difference will be const or non-const iterators.
So if you're iterating over MultiFields it's a good idea to think about
const-ness of your iterators and start using const interators where
possible.
Yours
Dirk
--
-- Dirk Reiners OpenSG Forum [EMAIL PROTECTED]
-- The OpenSG Open Source Scenegraph: http://www.opensg.org
-- Join the list at http://lists.sf.net/lists/listinfo/opensg-users
-------------------------------------------------------
This SF.Net email is sponsored by: NEC IT Guy Games.
Get your fingers limbered up and give it your best shot. 4 great events, 4
opportunities to win big! Highest score wins.NEC IT Guy Games. Play to
win an NEC 61 plasma display. Visit http://www.necitguy.com/?r=20
_______________________________________________
Opensg-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/opensg-users
