Hi Robert,

I believe the mailing-list is the prefered way of discussing issues. You've
rejected my last PRs (on master and OSG-3.4 branches)

src/osgUtil/CullVisitor.cpp: 1004
I proposed to replace :

        if (!updateCalculatedNearFar(matrix,drawable,false)) return;
by
        if (drawable.isCullingActive() &&
!updateCalculatedNearFar(matrix,drawable,false)) return;

But I didn't give context enough... I'm still trying to get the osgDB_logo
not being culled. The way it is done is that there is a Drawable created
with the image that ha a special drawImplementation method to draw directly
on the framebuffer.

This drawable has it's culling disabled (with no geometry) still has a
bounding box (at the origin).
When passing through the CullVisitor it's BB get's to update the near far
planes before my proposed change... Which could lead to the Drawable being
culled when the scene is not centered around the origin (which is my case
most of the time being using osg and osgEarth for country size models).


The change was then : When a drawable doesn't want to be culled don't use
it to update the calculated Near/Far planes...


For context here are your two comments on these PRs :

Toggling culling off is not toggling off the compute of the near/far, it's
> two very different things so it's inappropriate to put in this extra
> isCullingActive at this point. There is already an isCullingActive() above
> this code block. If there is an issue with the bounding volume then this
> should be addressed elsewhere.
>

and

I don't understand why you put the isCullingActive into the a line about
> updating the near far values. There is a isCullingActive() earlier in the
> method already.


Maybe now it is clearer ? I agree it is not really satisfactory, do you
have maybe another idea on how to solve this issue.

BTW updateCalculatedNearFar return value is used to discard a drawable if
the whole object is behind the eye point

--
nǝıɥʇɐƜ
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to