HI Roland,

On Fri, Apr 23, 2010 at 4:00 PM, Roland Smeenk <[email protected]> wrote:
> Sorry, no updates since submission.

OK.  My plan is to review the code with several passes through it,
looking to understand the approach taken and the implementation
details, and also looking to how we can incrementally work towards
getting shader composition integrated.

There are few changes such as comment/indentation that have been made
in your code, or in svn/trunk that make it look like there are more
critical changes than there really are, so what I'll do is merge these
first, if I modify your files I'll post these.  I'm tempted to create
a special branch in svn to test out these changes as well.  As I work
through I'll be able to provide more feedback.

In terms of design observations so far, the general approach you've
taken has similarities to the designs I have previously worked upon -
for instance the the shader generator attached to osg::State, the
shader cache, and the role of shader modes are similar.  One little
difference was that in my design work I was exploring whether it'd be
possible to reuse the exist OpenGL mode's used in osg::StateSet, and
extend there use in to custom modes that sit beyond the standard
OpenGL mode range.  I never got to actually implementing this though,
so I still don't know if it's practical.  For performance of the mode
to shader selection I was also considering a mapping of modes to a
bitmask, and the use of this bitmask as the key for lookup..  Again
how practical this is I don't know.... on actual implementation would
tell us.

Right back to review.
Robert.
_______________________________________________
osg-submissions mailing list
[email protected]
http://lists.openscenegraph.org/listinfo.cgi/osg-submissions-openscenegraph.org

Reply via email to