A corpus of examples to test parser implementations against

This piqued my interest, and it would help a lot when implementing a parser to have some more examples to test against, things like:
* some invalid files, with different errors
* some files with different mixing and matching of features used
* more?

I like the overall design and features.

Regards