利用者:Brita/GSoC BGE cleanup and support/reports/week10

GSoC14 - Weekly Report #10

July 19 - 25th - BGE Cleanup and Support

previous | next

This Week

The plan for this week was to wrap up the scene graph: 2-3 bugs left in the tracker and pending questions from the mailing list; and to start looking at physics bugs and code.

• I spent easily about 1/3 of my time just communicating: replying to emails, the tracker, the mailing lists and following up news.
• scene graph: I reviewed some parts of the code, cleaning up comments and style.

Then I picked the first bug with a problematic parent transform -> T33512. I reviewed more parts of the code looking for a wrong matrix multiplication.. something transposed (because the GE math library is misaligned with the rest of blender and OpenGL)...

I found the math library used for the ge rather ill suited for the job. I reviewed my math concepts, as it is the second day with the bug already. I asked for help, and here I have very positive feedback with sergey helping me a lot plus bastien, campbell and joshua leung all discussing the problem.

Turns out that the way the ge does the hierarchical transform math - component wise instead of with direct matrix multiplication - will always lead to this kind of problems. It needs to be all changed, and that is not a quick task, specially if merged with replacing the math library.

So I spent 3 days with this problem, plus some time to document it, because I have no time left on my soc to solve it.

The scenegraph developer documentation is basically finished. I moved there the rest of the questions I posted on the mailing list early in the week. Aligorith was very helpful with a fast complete response, but then I am lacking feedback from someone from the gameengine side.

The documentation has a 'questions' and 'problems' section including the problems above with the transform and the math library. I also made a related task on phabricator, I am still trying out the best way to coordinate/communicate.

• I had no time left to start the physics (just read some more things).
• The input bug from last week: putting it into a new design task.
• My branch is undergoing review and it should be merged soon.
• I updated the wiki with the end term goals and planning. I have basically 2 weeks left, because, for the pencils down week, I will be at Siggraph. Also, there is a trello now.

In short, goals are scene graph, physics, less bugs and more documentation.

After this week, I can no longer do the scene graph *and* the physics. I had not considered that the scenegraph problems could take so much time to solve. Most bugs in the tracker are either invalid or really difficult/time consuming to solve :( I decide for the physics, because it would yield more results as there is currently no maintainer.

General Development

• Scene Graph cleanup: 88f2b48, 86d8f10, 8468a28, 514c53e, 52d2099

Fixed (Reported)

None

Fixed (Unreported)

None

Closed / Archived / Invalid

• T30082 - Wrong blf font resolution with non-standard focal length/orthographic scale

Total: 0 fixed, 1 closed

Next Week

• tackle the physics bugs in the tracker. There are currently 46.
• start working on the physics documentation and demo files.