Clippy applies wrong fix for `useless_vec` for Edition 2021

[yongqli]

Clippy applies the wrong fix to the following code:

fn g() {
    for x in vec![0, 1] {
        f(x);
    }
}

fn f(x: usize) { }

with the error:

The following errors were reported:
error[E0308]: mismatched types
  --> crate/src/lib.rs:8:11
   |
8  |         f(x);
   |         - ^ expected `usize`, found `&{integer}`
   |         |
   |         arguments to this function are incorrect
   |
note: function defined here
  --> crate/src/lib.rs:12:4
   |
12 | fn f(_x: usize) { }
   |    ^ ---------
help: consider dereferencing the borrow
   |
8  |         f(*x);
   |           +

It seems that clippy replaced vec![0, 1] with &[0, 1], which is not correct. For Edition 2021, it should replace it with [0, 1] instead.

[kehrazy]

possibly this can be decorated further with a suggestion, seeing as the current implementation is, indeed, correct - the slice seems to be preferable here.

span_lint_and_sugg(
    cx,
    USELESS_VEC,
    span,
    "useless use of `vec!`",
    "you can use a slice directly",
    "consider making the value a reference",
    snippet,
    applicability,
);

[yongqli]

Now that array.into_iter() iterates by value, vec![]s can be replaced by arrays, although this might bloat compile times by creating many new types.

[sweihub]

I have a wrong fix too, please fix the clippy,thanks!

diff --git a/src/operator/mod.rs b/src/operator/mod.rs
index 2b6d6ec..f69838b 100644
--- a/src/operator/mod.rs
+++ b/src/operator/mod.rs
@@ -565,7 +565,7 @@ impl Operator {

 fn is_valid_start(id: &str) -> bool {
     match id.chars().next() {
-        Some(c) => ('a'..='z').contains(&c) || ('A'..='Z').contains(&c),
+        Some(c) => c.is_ascii_lowercase() || c.is_ascii_uppercase(),
         None => false,
     }
 }

[kehrazy]

I have a wrong fix too, please fix the clippy,thanks!

diff --git a/src/operator/mod.rs b/src/operator/mod.rs
index 2b6d6ec..f69838b 100644
--- a/src/operator/mod.rs
+++ b/src/operator/mod.rs
@@ -565,7 +565,7 @@ impl Operator {

 fn is_valid_start(id: &str) -> bool {
     match id.chars().next() {
-        Some(c) => ('a'..='z').contains(&c) || ('A'..='Z').contains(&c),
+        Some(c) => c.is_ascii_lowercase() || c.is_ascii_uppercase(),
         None => false,
     }
 }

1. offtopic
2. clippy is right.

[sweihub]

I want letter inside [a-zA-Z]，no numbers no other characters, however clippy fix as lower/upper case only, how do you call that is right?

[kehrazy]

I want letter inside [a-zA-Z]，no numbers no other characters, however clippy fix as lower/upper case only, how do you call that is right?

a.is_ascii_lowercase() corresponds to [a-z].
a.is_ascii_uppercase() corresponds to [A-Z].
[a-z] | [A-Z] corresponds to [a-zA-Z].

How do you call this incorrect?

[sweihub]

OK, sorry, the .is_ascii_ is misleading, it should be is_alphabet_.

[kehrazy]

@sweihub Stop bumping the thread, please. Open your own issue, if you believe that the naming is not right.