Should 'ExactlyOneError' implement std-'Error'?

[njam]

The exactly_one() function was added in #310:

exactly_one(self) -> Result<Self::Item, ExactlyOneError<Self>> 

For using the early return operator ? on its result, it would be nice if ExactlyOneError would implement Error. Or is there any reason not to?

cc @Xaeroxe @Centril @bluss

[Xaeroxe]

Fine by me

[jswrenn]

@njam Would you like to file a PR? This would be sufficient:

impl<I: Iterator> Error for ExactlyOneError<I> {}

[Xaeroxe]

I want to PR for a small change to the innards of the thing, so I'll get it.

[Xaeroxe]

See #484

[tech6hutch]

I was just wanting this today, since the lack of Error impl makes ExactlyOneError not work well with anyhow. Is there any idea of when this enhancement will make it into the cargo version?

[jswrenn]

Thanks for poking me on this, @tech6hutch. I've been very busy. :( I'm going to try to chew threw the remaining items in our release milestone this week.