利用者:Mx/GSoC2008 Week9

提供: wiki
移動先: 案内検索

GSoC 2008 - Week 8 report

This week

I continued working on removing SWIG. Despite trying really hard to meet my weekly target (getting the removal completed), the job is still not finished. I am, like all of you, very disappointed to note that this area of the integration is not ready yet.

I understand you might ask yourself why I am running behind schedule. This is clearly not due to a lack of work on my part, as my commit logs show. I am working full-time and full-steam on the project and I don't see how I could go faster. It is not due to perfectionism either: I have actually changed my approach to the integration. Instead of redefining the API considerably, I realized this week that I first needed to integrate the old API before starting refactoring it. The clearest example is the implementation of iterators, that I originally wanted to replace with a list-based approach. I still hope to achieve that in the future but for the time being, just getting it running independently of SWIG is a more pressing issue. I also have tried to optimize my time to meet the target by not coding test cases for my API. I hope this won't affect my first running tests in the days to come.

I feel I have underestimated the amount of work to create a "native" Python API for scratch. I have realized this past week that I should have set aside at least two and a half weeks for the task, while I candidly had planned to get it done in a week. The process is slow and massive in terms of the amount of code that needs to be generated.

After these bad news, fortunately, I am almost finished. I ported almost all base classes of the API. I have 4 base classes left (which each take me about two hours of work on average). Once this is done, I just need to copy and paste the shaders', functions' and predicates' subclasses. This should be quick.


Next week

I should be done porting the API very soon. Then, I'll have to test to make sure it works. That's another big unknown. I'll then be able to test whether my approach was valid or not. I don't see any real issue, but at the same time I don't really know before having tested it.

Once I am done with SWIG (finally...), I'll work on getting lib3ds out of the system.


Issues

None


Schedule

I am behind schedule and this is still a major worry to me. I will have to reassess with Jean-Luc, what I can and cannot deliver for the end of the summer. Whatever remains will be implemented after the GSoC session.