new lint: bufreader_stdin

[yawara]

changelog: [bufreader_stdin]: new lint

fixes #6755

[rustbot]

r? @Manishearth

rustbot has assigned @Manishearth.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[J-ZhengLi]

~

[J-ZhengLi]

nit: I think Buf and Reader should be separated words? no?

[Manishearth]

nah, this is BufReader + Stdin, so I'd keep BufReader as a single atomic unit in the name

[J-ZhengLi]

we should be very careful dealing with suggestions where type changes occur, this suggestion won't work if the resulting variable is used somewhere else, for example:

use std::io::{self, BufReader};
use std::io::StdinLock;

fn main() {
    let b = io::stdin().lock();
    let reader = BufReader::new(b);
    // let reader = b; // does not work
    
    i_want_buf_reader(reader);
}

fn i_want_buf_reader(reader: BufReader<StdinLock>) {
    // do something with the reader
    drop(reader);
}

this will not compile, and another example would be

static READER_ONCE: OnceLock<BufReader<Stdin>> = OnceLock::new();
let a = READER_ONCE.get_or_init(|| BufReader::new(io::stdin()));

although these are very unlikely to happen, but you can never be too careful~ So I'd say at least changing the applicability to MaybeIncorrect?

Also, it would be nice to run lintcheck to see if there are any false positive cases.

[Manishearth]

Yes, these suggestions aren't MachineApplicable. MaybeIncorrect, unfortunately.

[J-ZhengLi]

looks like an accident, make sure to change it to lowercase name