Nicer ignore of parse errors in sBIT

[kornelski]

No description provided.

[fintelia]

It would be nice to align the style with parse_cicp, parse_mdcv, and parse_clli. And ideally the rest of the optional metadata chunks as well.

[kornelski]

Oh, that would be even bigger churn. I'd rather not.

If I were going to touch them, I'd change them all to the style of sBIT, because having parsing code that can't use Result is too cumbersome, and having extra inner helper function is unnecessary if the failure can be ignored in parse_chunk.

[kornelski]

I've considered changing error handling for all ancillary chunks, but if we haven't had reports of unnecessary failures caused by them, then maybe they're not too broken in the wild, and it's okay to keep stricter checks.

[fintelia]

Switching the other ancillary chunks is tracked in #525.

We don't have to switch everything in one PR. But I would like to decide on one approach to ignore errors from chunks and then make sure all changes move us towards that