Incremental regeneration

[ckald]

Jekyll 3b has an amazing feature: incremental regeneration. The website I have that was generating for a couple of minutes now updates in a second.

We could do something similar: build a dependency graph of the project (including folder structure and cross references) and regenerate only those files that really need to be updated.

This is an interesting engineering problem, but I am not sure if this is really required considering the orientation of Pyccoon on small projects

[cryptonomicon314]

The project on which I use Pyccoon is small, but has lots of static data files which must be copied (mainly to generate the documentation and for tests), and it takes a non-negligible amount of time even if only a single file has been changed.

I think it's an interesting engineering problem too, and I would be interested in working on it.

Link to my project: http://cryptonomicon314.bitbucket.org/msdiff/documentation.md.html

[ckald]

Is it still slow with the latest master? I think I fixed the problem when Pyccoon tried to parse static and binary files — performance went through the roof even without incremental regeneration.

[cryptonomicon314]

I don't know. I'll try it after dealing with the markdown extension.