File: method-ambig-one-trait-unknown-int-type.rs

package info (click to toggle)
rustc 1.85.0%2Bdfsg3-1
  • links: PTS, VCS
  • area: main
  • in suites: experimental, sid, trixie
  • size: 893,396 kB
  • sloc: xml: 158,127; python: 35,830; javascript: 19,497; cpp: 19,002; sh: 17,245; ansic: 13,127; asm: 4,376; makefile: 1,051; perl: 29; lisp: 29; ruby: 19; sql: 11
file content (36 lines) | stat: -rw-r--r-- 1,017 bytes parent folder | download | duplicates (11)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
// Test that we invoking `foo()` successfully resolves to the trait `Foo`
// (prompting the mismatched types error) but does not influence the choice
// of what kind of `Vec` we have, eventually leading to a type error.

trait Foo {
    fn foo(&self) -> isize;
}

impl Foo for Vec<usize> {
    fn foo(&self) -> isize {1}
}

impl Foo for Vec<isize> {
    fn foo(&self) -> isize {2}
}

// This is very hokey: we have heuristics to suppress messages about
// type annotations needed. But placing these two bits of code into
// distinct functions, in this order, causes us to print out both
// errors I'd like to see.

fn m1() {
    // we couldn't infer the type of the vector just based on calling foo()...
    let mut x = Vec::new();
    //~^ ERROR type annotations needed
    x.foo(); //~ ERROR type annotations needed
}

fn m2() {
    let mut x = Vec::new();

    // ...but we still resolved `foo()` to the trait and hence know the return type.
    let y: usize = x.foo(); //~ ERROR mismatched types
}

fn main() { }