Investigation into Linux test failures and possible improvements

Thanks for looking into this, suggesting specific changes, etc!

Assuming that it doesn’t hamper the team’s efforts, I applaud the idea of refactoring Monty’s interfaces to centralize common functionality. Aside from reducing the incidence of cut-and-paste programming (and related problems), this can provide a good place to extend capabilities, ease cross-platform development, etc.

Speaking of which, I’ve commented before on the idea of centralizing all of the CMP I/O:

What are the prospects for doing this? Also, could you comment on the amount of OO-fu you’ve seen in the code base? Here’s a related post: