[Soc-2008-dev] Week 10 report
nicholasbishop at gmail.com
Sat Aug 2 03:48:05 CEST 2008
a) Previous week:
Compared to my flurry of commits last week, this week was pretty slow.
* Fixed undo for sculptmode + multires
* Memory reductions + bugfix for putting certain modifiers after multires
* Fixed multires shortcut keys
Early this week my motherboard fried, so I've been on a backup
computer until a few hours ago. Slow as hell to compile Blender (it's
actually more or less the same computer I used when I first worked on
Blender in 2006; I've gotten spoiled since!) I have my new motherboard
in now though, so everything's back to normal.
Anyway, I've been working to do something about the very slow
subdivision/level switching of the new multires (as compared to the
multires in trunk.) Basically, I'm removing some uses of CCGSubsurf in
favor of directly subdividing displacements. CCGSubsurf is still used
for outputting the final mesh, so there won't be any user-visible
differences other than speed. Directly subdividing displacements is an
approach I had abandoned earlier as too complex, but I think I can
make it work at least as cleanly as the current code.
I also had a look at file loading; I've written (but not comitted) the
basic framework needed to convert old multires files to the new
modifier format, but it still needs a somewhat complex conversion from
the old vertex ordering to the new displacement grids, which will take
b) Next week:
More work to do on the displacement subdivision code. Not sure how
much more work this'll be; handling the edges of faces is difficult.
Also, I've asked for some user testing here:
There're a number of bug reports I will look at.
If memory serves, deadline for SoC is August 18th. Looking good.
On Fri, Aug 1, 2008 at 7:31 PM, Davide Vercelli
<davide.vercelli at gmail.com> wrote:
> Week 10
>>> Note: on Sunday I'll be flying back to Italy, so I can't attend the meeting. Sorry about that! <<
> This week
> Added support for "Sphere" option and phong correction/terminator thingy/bias.
> Started working on "faster tree creation", but no code committed yet.
> It turned out to require some modifications to the kd-tree code, and I
> want to make sure that they don't break other uses of it, etc. I am
> debugging right now, I hope to commit as soon as possible.
> Next week
> Targets: finishing "faster tree creation", and starting basic,
> proof-of-concept implementation of instant radiosity.
> Btw support for Oren-Nayar diffuse is probably buggy, have to look into it.
> Pretty good, but still a lot of details to fix! In the next couple of
> days I'm moving back to Italy so I'll resume working on gsoc on
> Soc-2008-dev mailing list
> Soc-2008-dev at blender.org
More information about the Soc-2008-dev