Add manual_bits lint

[paolobarbolini]

Closes #6670

---

changelog: new lint: [manual_bits]

[rust-highfive]

Thanks for the pull request, and welcome! The Rust team is excited to review your changes, and you should hear from @flip1995 (or someone else) soon.

Please see the contribution instructions for more information.

[m-ou-se]

Oh nice! Thanks for implementing this! (I'm not a clippy reviewer. Just happy to see this getting added. ^^)

[ghost]

It looks like it triggers for size_of_val as well. I don't think it should do this.

#![warn(clippy::size_of_as_bits)]

use std::mem;

fn main() {
    let x = 0u32;
    println!("{}", 8 * mem::size_of_val(&x));
}

warning: you seem to be using `mem::size_of::<T>()` to obtain the size of `T` in bits
 --> src/main.rs:7:20
  |
7 |     println!("{}", 8 * mem::size_of_val(&x));
  |                    ^^^^^^^^^^^^^^^^^^^^^^^^ help: consider using: `u32::BITS`
  |
note: the lint level is defined here
 --> src/main.rs:1:9
  |
1 | #![warn(clippy::size_of_as_bits)]
  |         ^^^^^^^^^^^^^^^^^^^^^^^
  = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#size_of_as_bits

[ghost]

There is also an issue with type aliases.

#![warn(clippy::size_of_as_bits)]

use std::mem;

type Word = u32;

fn main() {
    println!("{}", 8 * mem::size_of::<Word>());
}

warning: you seem to be using `mem::size_of::<T>()` to obtain the size of `T` in bits
 --> src/main.rs:8:20
  |
8 |     println!("{}", 8 * mem::size_of::<Word>());
  |                    ^^^^^^^^^^^^^^^^^^^^^^^^^^ help: consider using: `u32::BITS`
  |
note: the lint level is defined here
 --> src/main.rs:1:9
  |
1 | #![warn(clippy::size_of_as_bits)]
  |         ^^^^^^^^^^^^^^^^^^^^^^^
  = help: for further information visit https://rust-lang.github.io/rust-clippy/master/index.html#size_of_as_bits

I think the suggestion should be Word::BITS instead.

[camsteffen]

The name is a bit funny to me. Perhaps size_of_to_bits or bits_from_size_of.

[ghost]

I like manual_bits for the name.

[paolobarbolini]

I applied all of the above suggestions

[bors]

☔ The latest upstream changes (presumably #8236) made this pull request unmergeable. Please resolve the merge conflicts.

[bors]

☔ The latest upstream changes (presumably #8210) made this pull request unmergeable. Please resolve the merge conflicts.

[flip1995]

I also like the manual_bits name. This is similar to the manual_memcpy name (and lint).

Small comment for the error message. This will require a cargo dev bless. Otherwise LGTM.

[paolobarbolini]

👍. I also rebased on top of master since #8190 would have broken my code and bumped the clippy::version to 1.60.

[flip1995]

@bors r+

Thanks!

[bors]

📌 Commit 166737f has been approved by flip1995

[bors]

⌛ Testing commit 166737f with merge 7c82ae1...

[bors]

☀️ Test successful - checks-action_dev_test, checks-action_remark_test, checks-action_test
Approved by: flip1995
Pushing 7c82ae1 to master...

[HaoYang670]

I am afraid that in most contexts (especially in match arms), sizeof::<u8>() * 8 cannot be simply replaced by u8::BITS because the return type of sizeof is usize but u8::BITS is in u32 type.

What about using "help: consider using: u32::BITS as usize "?

[ghost]

@HaoYang670, what is the point of tacking a comment onto a PR that was merged three months ago? The correct way to get your problem addressed is to create an issue.

[xFrednet]

I'll fix the suggestion, that'll probably be faster than creating an issue for it ^^