Re: [osg-users] cull setting inheritance

2008-03-05 Thread Mathias Fröhlich
Good morning, On Tuesday 04 March 2008 16:13, Robert Osfield wrote: It a matter of expediency and history, reusing the CullSettings class in various roles rather than creating different classes that store the same parameters. The history comes down to how the OSG has evolved with new

Re: [osg-users] cull setting inheritance

2008-03-05 Thread Robert Osfield
Hi Mathias, The inheritance of state, transforms and cull settings should all be down top down, so the highest objects in the OSG world being the view(er) then the view's Camera's, then the scene graph below them - this is public face of the scene graph. The rendering traversal is just a back

Re: [osg-users] cull setting inheritance

2008-03-05 Thread Mathias Fröhlich
Hi Robert, On Wednesday 05 March 2008 10:35, Robert Osfield wrote: The inheritance of state, transforms and cull settings should all be down top down, so the highest objects in the OSG world being the view(er) then the view's Camera's, then the scene graph below them - this is public face of

Re: [osg-users] cull setting inheritance

2008-03-05 Thread Robert Osfield
On Wed, Mar 5, 2008 at 12:34 PM, Mathias Fröhlich [EMAIL PROTECTED] wrote: But the View does not have CullSettings itself. Sorry, trying to do a quick reply without doing background research. So, which one is currently the View's 'master setting' before the cull visitor will hit the

[osg-users] cull setting inheritance

2008-03-04 Thread Mathias Froehlich
Hi Robert, I am not sure about that, but to make cull setting inheritance for cameras work, I would think that we should use the cameras inheritance mask to decide which settings are innherited rather than the cull visitors inheritance mask? That is: Shouldn't the following change be applied

Re: [osg-users] cull setting inheritance

2008-03-04 Thread Robert Osfield
Hi Mathias, Could you back out a little and explain what you think is going on right now, and what you think should be happening and why, I'm afraid my brain doesn't work well with just a diff output. Robert. On Tue, Mar 4, 2008 at 12:49 PM, Mathias Froehlich [EMAIL PROTECTED] wrote: Hi

Re: [osg-users] cull setting inheritance

2008-03-04 Thread Mathias Fröhlich
Robert, On Tuesday 04 March 2008 14:07, Robert Osfield wrote: Could you back out a little and explain what you think is going on right now, and what you think should be happening and why, I'm afraid my brain doesn't work well with just a diff output. Well sorry, I thought you know your own

Re: [osg-users] cull setting inheritance

2008-03-04 Thread Robert Osfield
Hi Mathias, Thanks for the explanation. Reviewing the code now I do wonder if the inheritance isn't a bit the wrong way round. The code has the CullVisitor inheriting from the local Camera that is being traversed. I am now thinking that the camera's settings should be inheriting from

Re: [osg-users] cull setting inheritance

2008-03-04 Thread Mathias Fröhlich
Robert, On Tuesday 04 March 2008 15:16, Robert Osfield wrote: I have gone ahead an added an #if 1 #else #endif block using the new suggested code below to CullVisitor.cpp. Things appear to hang together on the osgprerender and osghud examples but.. I haven't done any exhaustive testing.

Re: [osg-users] cull setting inheritance

2008-03-04 Thread Robert Osfield
On Tue, Mar 4, 2008 at 2:54 PM, Mathias Fröhlich [EMAIL PROTECTED] wrote: Yes, the semantics of the current CullSettings api suggest your change. But no, that needlessly copies all values over and copies back what is in the mask. Not nice IMO :) The copies are done to replicate the effect