fix nightly build issues with as_ref()

[nevi-me]

fixes #217

I spent part of the day trying to see what change to Rust introduced this issue, but came blank. I've replaced string.as_ref() with a clone, which fixes the issue. If clone is a problem, we could resolve this later (though I was able to fix the error in a different way which was a bit longer). Found a better fix of passing a &str, and annotating it as such where necessary.

cc @gwenn

[gwenn]

We don't want to clone/allocate until we are sure that:

        if line.as_ref().is_empty()
            || (self.ignore_space
                && line
                    .as_ref()
                    .chars()
                    .next()
                    .map_or(true, char::is_whitespace))
        {
            return false;
        }
        if self.ignore_dups {
            if let Some(s) = self.entries.back() {
                if s == line.as_ref() {
                    return false;
                }
            }
        }

[nevi-me]

That's fine,, please check my latest commit

[nevi-me]

Darn, it "worked on my machine", I'll investigate the failures

[nevi-me]

Nearly there, it's currently the example that's failing. Struggling to repro as I don't have the history.txt file

[nevi-me]

@gwenn is the current solution acceptable?

[gwenn]

I hope/guess you will consider #218 to be better.
Thanks for your help/time.

[nevi-me]

Yes, I just saw it, and I'll close this.

Will you be able to release a 4.0.1 today?

[gwenn]

Do you mind if a 4.1.0 is released instead ?
I am not sure that we don't have introduce any regression with old stables

[nevi-me]

Do you mind if a 4.1.0 is released instead ?
I am not sure that we don't have introduce any regression with old stables

Either one, I haven't used rustyline before, so I wouldn't know what would be appropriate. I only looked at this because it was blocking a growing number of PRs.