10% of calls to .*iter() are redundant

[Veedrac]

The source has just over 200 calls to iter, iter_mut or into_iter to things that take and T: IntoIterator. This is, by my count, about 10% of them!

Some examples are left.extend(Some(x).into_iter()) instead of left.extend(Some(x)) and for &(s,t) in v.iter() instead of for &(s,t) in v.

I thought this would make a quick, trivial change to warm myself up to the build system and such, so I've already got a branch removing it. I thought I'd ask if it's actually desirable to do so, though, since it's a lot of seemingly-meaningless code churn.

If this is wanted, the docs should probably also be updated to emphasize that many calls to .*iter() aren't needed. Much of the documentation for iter() and co. uses them in places where they can be elided.

[steveklabnik]

the docs should probably also be updated to emphasize that many calls to .*iter() aren't needed.

We do already specifically talk about this in the doc, but you have to remember that rustc's codebase is older than IntoIterator! :)

Generally speaking, we accept PRs that update the code to something more idiomatic. I don't speak for everyone, but I would be 👍 on it.

[steveklabnik]

After pinging #rust-internals, it would seem that others agree. Please send a PR! Thanks :)

[Manishearth]

Rust and Servo are bad examples for idiomatic rust code because of the hysterisis of rust upgrades.

I haven't seen new code doing this, so I don't think it's a pervasive problem -- just isolated to rust. 👍 to fixing it 😄