Increasing Decimal Logical Types Support

[AlexRiedler]

I work a fair bit with logical types, and noticed we could probably increase support a fair bit by adding precision and scale to our decoded logical types (or even better potentially add support for bigdecimal crate).

I tried to implement this, was going to give it a PR but ran into an issue that would require significant refactoring of the code, and the public interface 😢 . In particular, the BigInt provided in Decimal type currently could be more than the number of bytes supported by the Fixed field and its unclear how to return an error in this case when encoding.

Suggestions on how to move forward on this? or should we just panic like we currently do?

AlexRiedler@0ac5e1f is my current change-set if you are interested