[Soc-2018-dev] YimingWu's week 12 report for NPR project

n w xp8110 at live.com
Sat Aug 4 03:31:58 CEST 2018


GSoC Week12 Status

------


Hello everyone, here's my week 12 report! Sorry for the delay, I've been busy with school project since August. (but I hope time difference will offset it back...)


Things have done


1. A basic user document is finished, see D3572 for the patch (patch only includes text).
2. Fixed software rendering animation sequence data cache, now sequence can be rendered without a problem.
3. Fixed software mode triangle culling errors. Triangle that cross the near clipping plane used to be a problem, but now everything works great. (yeah!)
4. Fixed render-while-drawing-viewport conflict.
5. Fixed some user interface panel missing problems, but haven't fixed them all.
6. Thanks to Joshua, GCC compiler now working.
7. Also fixed some minor memory leaks, so now LANPR renders everything without building up a large trunk.


Here's software mode animation rendering, including chained result. This is the first time  I've seen LANPR itself in action rendering animation. turns out results are good.

http://www.wellobserve.com.img.800cdn.com/NPR/DevLogs/Img/lanpranimation2.gif

http://www.wellobserve.com.img.800cdn.com/NPR/DevLogs/Img/lanpranimation3.gif


My plans on future developments


I plan to stay around **(a lot) longer** for LANPR project, because there are indeed many people looking forward to see this in master, if I continue improve the code, we don't need to halt the project.


Still helps are also needed as I'm not that familiar with modules in blender, so when anyone sees anything that he/she can improve, feel free to do so.


It's been very exciting to participate in GSoC project. Also, I'd like to thank every person in blender's community who helped me make this project possible.


If anyone could check those first...


Currently there are these problems which should be solved first:


1. In lanpr_ops.c line 2088, there's a very strange situation related to threading. See comment for info.
2. When using DPIX mode to render images, after F12 , then viewport will display wrong results. (But F12 is always correct.) I'm suspecting the state for half-pixel texture sample in OpenGL is not reset after drawing, because DPIX uses every single pixels on data texture to produce results.


YimingWu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.blender.org/pipermail/soc-2018-dev/attachments/20180804/b8f51b04/attachment.html>


More information about the Soc-2018-dev mailing list