Assume conversions to byte slices are always successful

[koute]

AFAIK there isn't much point in returning a Result when converting from &[T] to &[u8] as those conversions should always be successful.

[sdroege]

This breaks the API, but that's fine. Thanks a lot!

Can you bump the version to 0.3.0 in Cargo.toml in a second commit? Then I'll merge it

[koute]

Done! I've also updated the changelog.

Please don't release this on crates.io yet though. I have one more potential breaking change I'd like to create a PR for; I just wanted to wait for this one to get merged. (:

[sdroege]

Done! I've also updated the changelog.

Great, thanks :)

Please don't release this on crates.io yet though. I have one more potential breaking change I'd like to create a PR for; I just wanted to wait for this one to get merged. (:

Sure, what other change is that?

[koute]

Sure, what other change is that?

Extending the Error type so that the error message is more descriptive, e.g. in case of an alignment mismatch the error message would be something like this: (I don't mind bikeshedding the exact message if you'd prefer something different)

cannot cast a &[u8] into a &[u16]: the slice's address is not divisible by the minimum alignment of u16 (= 2)

instead of the current "Wrong Alignment".

[sdroege]

Yeah that makes sense. Please put all the logic into the Display impl though and keep the Error type itself allocation-less :)

[koute]

Yep. That's the idea. (: