Consider excluding Duration's associated constants from identity_op lint

[dtolnay]

This came up in the tracking issue for std::time::Duration::SECOND and friends: rust-lang/rust#57391.

The following code using 1 * SECOND arguably reads more clearly in English than thread::sleep(SECOND).

#![feature(duration_constants)]

use std::thread;
use std::time::Duration;

fn main() {
    thread::sleep(1 * Duration::SECOND);
}

warning: the operation is ineffective. Consider reducing it to `Duration::SECOND`
 --> src/main.rs:7:19
  |
7 |     thread::sleep(1 * Duration::SECOND);
  |                   ^^^^^^^^^^^^^^^^^^^^
  |
  = note: #[warn(clippy::identity_op)] on by default
  = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#identity_op

The same pattern is common in Go which uses a similar mechanism for time constants. From a brief search, here are some articles that use 1 * time.Second or 1*time.Second in Go:

• https://go.googlesource.com/proposal/+/master/design/12914-monotonic.md
• https://blog.cloudflare.com/the-complete-guide-to-golang-net-http-timeouts/
• https://dzone.com/articles/practical-golang-getting-started-with-nats-and-rel
• https://divan.dev/posts/go_concurrency_visualize/
• https://dev.to/dizdarevic/golang-the-next-language-to-learn-for-developers-pc5
• http://devs.cloudimmunity.com/gotchas-and-common-mistakes-in-go-golang/
• https://medium.com/code-zen/concurrency-in-go-5fcba11acb0f

This may indicate that this pattern does not fall under the "obscures what’s going on" description of the identity_op lint.

clippy 0.0.212 (5d78250 2019-03-07)

[oli-obk]

I agree, we should maybe just whitelist multiplying with constants, since no matter what the constant is, the user is probably doing that multiplication out of a reason.

[Alexendoo]

No longer linted: #8204