[read]: for tables with a known `format` field, parsing should fail on mismatch

[cmyr]

Thinking about this more, though, I think we should be doing more: I think that for any table with a defined format we should actually be checking the format at parse time and returning an error if it isn't the expected value; this would mostly be future-proofing.

Originally posted by @cmyr in #949 (comment)