[Soc-2009-dev] weekly report
jingyuan.huang at gmail.com
Sat Jun 13 20:37:06 CEST 2009
First of all sorry about the delay. I will send my report on time in
What I did:
Last week's progress was slower than I'd like.
* undo/redo light strokes: considered done, though it can be more efficient.
* save/load/reinit light env: could be considered done, however not
enough testing has been done. I also ran into a bug that cost me a lot
of time. I will mention it later.
* start jobs for coeffs computation: remained untouched.
** UI & rna: I'm slowly adding things to rna. I think there is a
learning curve that I need to catch up.
What I will do next week:
* continue UI & rna. I'd like two buttons to change SH properties, a
button to recompute sh coefficients, buttons to change light env type
and load a probe image, and a button to save light to hdr file.
* improve computation job (unfinished task from this week)
What I will do the week after next:
It will be a choice between integrating ray trace functions to get
shadow coefficients and intergrating Matt's patch to generate
renderings using SH. I personally prefer the latter since it will
complete a workflow even though the display in light paint mode
doesn't have shadow.
1. I think I need some time to understand rna structure better. I will
seek for help on IRC or here if I run into something that I don't
2. I ran into a bug with one file
to reproduce the bug:
a. Load bunny.blend. Switch to light paint mode.
b. (optional) pick a colour and paint the scene.
c. Save the file as blah.blend
d. Open blah.blend.
e. System crashes
backtrace tells me that it fails in render routines, which puzzled
me. Several attempts to find out the real cause have failed. It'd be
great if someone else can take a look at this and see if I didn't
something blatantly wrong.
3. Is there some testing functions available to check if values are
written to and loaded from blend files correctly?
According to my original schedule, I will have a build at around July
6th which has the basic painting features. I'd like to push that line
forward to the end of June (June 28th maybe) so that I can have people
to try out and report bugs back. I also think that I should start
light node asap after I get the build out.
Computer Graphics Lab
University of Waterloo
More information about the Soc-2009-dev