I typically don't use the exn functions, but is it problematic that we're changing what exn a bunch of functions is raising? I'm thinking about the case when people are actually relying on catching them.

Edit: oops, accidentally fat finger closed the PR...

It should probably be noted as a breaking change, but these are all functions that have non-exception-raising alternatives. It doesn't really make sense to catch these exceptions, since you should then use the alternative non-exception-raising function instead (that definitely doesn't mean that people aren't still using it though).

So, what about splitting this up, first adding the panic function, and then doing a separate discussion/PR on whether the current exn methods should use panic under the hood?

So, what about splitting this up, first adding the panic function, and then doing a separate discussion/PR on whether the current exn methods should use panic under the hood?

Done. Internal use split out to #79

Thanks @glennsl !