[Bf-committers] RNA property updates vs anim sys

Campbell Barton ideasman42 at gmail.com
Wed Nov 24 18:10:38 CET 2010


On Wed, Nov 24, 2010 at 5:08 AM, Matt Ebb <matt at mke3.net> wrote:
> Hi,
>
> What's the deal with RNA property update functions and the animation system?
>
> I would have expected the animation system to run the update functions
> when an rna property is updated via an fcurve/driver/etc, but they're
> not. Why is this, and can it be fixed? It's pretty bizarre that you
> can drag a property button, see the value in the button change, and
> see the results update (since the rna update function is being
> triggered) but if you do the same thing via an fcurve, see the same
> value in the button change, no update functions get run.
>
> This is throwing a spanner into the works of this ocean sim work,
> since I need to be able to update the simulation data when parameters
> (currently belonging to a texture) are changed, eg. 'time'. It works
> fine when you're dragging values in the property editor, but if you
> animate it, nothing updates. It's not just updating the sim either,
> the update functions are used to send depgraph update notices which
> aren't being sent either.
>
> RNA set functions are not appropriate here - I'm not doing anything
> fancy to the values as they are set, it's just setting DNA directly
> like most normal RNA properties - the issue is that I'm telling it to
> update when it's changed, but it's not doing that.
>
> How can I make this work?
>
> Matt

>From what I can tell, the reason update functions are not called is
because it would slow blender down too much. Especially for arrays -
quat's colors, rotations. Where updating a transformation or material
3-4 times will immediately make blender noticeably slower.

One possible solution is to queue unique update functions for each
update, per ID.
A function pointer GHash & a check/w previous comparison should make
collecting unique functions quite fast. Then these can run at the end,
per ID.

-- 
- Campbell


More information about the Bf-committers mailing list