Today I finished-up my work writing automated tests for the Processing.js parser and a full unit test infrastructure. It has been an interesting bit of work, and resulted in a lot of good info. Today I imported Ben and Casey's sample code from their Processing books, and it was interesting to watch things pass and fail.
Of the 855 Processing files I tested today, 623 were parsed correctly (72%). This leaves 232 that fail for one or more reasons. It's actually not as bad as it sounds, since there are six main errors I see over and over, for which I've filed bugs, and a bunch of other failures that are likely related to my python/js scripts.
I'm pumped to have some new students about to start writing tests and doing more investigations into these bugs. At the same time, we've got a new project coming online to do an antlr-based rewrite of the parser.
If you've got Processing code that you want to make sure works on the web, let us know so we can add it to our automated test suite.