you're right.  that worked, but it turns out the FeedbackBuffer is not the approach I'm looking for.  there's no Z-depth because its in screen coordinates, and it clips everything to the viewing frustum, as opposed to giving me polygons _before_ clipping and flattening.

will maybe pursue the re-use of Mesa vertex program processing.

thanks,
mike

On 1/17/06, Brian Paul < [EMAIL PROTECTED]> wrote:
Michael Frumin wrote:
>
>
> On 1/17/06, *Brian Paul* <[EMAIL PROTECTED]
> <mailto:[EMAIL PROTECTED] >> wrote:
>
>     Michael Frumin wrote:
>      > I am attempting to capture 3D geometries from OpenGL applications for
>      >
>
>     Have you looked at OpenGL's feedback feature?  It lets you capture
>     OpenGL geometry after it's been transformed to screen space, before
>     rasterization.
>
>
>
> Wow, I wish I had known about that earlier.  That's a great idea, in
> fact it was super easy to rig something up to capture the whole buffer.
> Any suggestions as to how to easily build a parser for the binary string
> returned in the buffer, that is structured according to the grammar at:
> http://www.opengl.org/documentation/specs/man_pages/hardcopy/GL/html/gl/feedbackbuffer.html
>  (preferable in windows, but *nix would be fine just to see how it works)?

You don't really need a parser.  I think that a switch statement with
some loops (over the polygon vertices) would suffice.

-Brian

Reply via email to