[Soc-2007-dev] Friday Status Report

Aaron Moore two.a.ron at gmail.com
Sat Jul 28 09:08:08 CEST 2007


Hi,

Sorry this is so late on friday, my sleeping schedule is really weird
right now, so I just got up...

What I did this week:
  1. Went through 2 iterations of solutions to the
callbacks-can't-access-the-API problem and settled on a solution.
  2. Wrote a preliminary aqsis plugin and discovered that I still need
to solve a coordinate system problem: blender - front y, up z; vs.
aqsis - front z, up y. I put the necessary registration features in
place but haven't gotten to implementing the projection of all
coordinates into the new aqsis space yet. This works via RIB export, I
haven't gotten the result-passing functions working yet, so one needs
to render this RIB file externally.
  3. Embedded the render callbacks in the pipeline and begun coding
the internal renderer's callbacks.
        a. filled the lampren structures with as much information as
the render api currently exports.
        b. created the function to convert polygon meshes to vertrens
and vlakrens, implemented this.
        note: since I've not gotten the result functions working yet,
this does not produce an image, however I'm currently working on this,
and when I get it done, it ought to. I know this means I'm about a
half a day late on my deadline. Also, I haven't gotten a chance to
debug these, mostly because it will be much easier to do when I can
see something.
   4. added render API data to the Render structure and created a
RenderDatabase structure to move all of the render database components
of the Render structure into. Note: since I am trying to setup the
render API's version of the convert code in parallel to the current
convert code, I won't remove the render database code from the Render
structure until the new convert code is finished.
   5. finally, since the RenderDatabase is only accessible in the new
rendercallbacks file, I have to move all of the internal renderer code
out of the pipeline and into this file. I begun to do this. When all
of the code which has result sending pieces to it converted and moved
to the new file, I will be able to setup result sending and then a
render will come out. (of course I'll have to do a bit of debugging,
but...)

What I plan to do next week:

In the next day or so:
  1. Finish converting the blender internal pipeline code.
  2. Write the result sending part of the render API, and hook up the
converted internal renderer code up to it.

After that:
  3. Solve the aqsis coordinate problem. Convert rib export code to
code that directly calls aqsis and return results via the new system.
  4. Implement the rest of the render API specification (except
functions which require more specification for implementation). Add
this new information the the WIP exports: blender internal and aqsis.
  5. Write the rest of the geometry specifications. Implement the rest
of the geometry export: subdivision surfaces, nurbs, metaballs, curves
and particle systems. If after writing the specifications, this goal
seems really unrealistic, then do all mesh related ones (subsurfaces,
particles), and nurbs.

No big problems at the moment.

Aaron


More information about the Soc-2007-dev mailing list