Rust-GCC
Dataless enum variants can be matched without full path
I tried this code: https://godbolt.org/z/7eqP9vrTn
extern "C" {
fn printf(s: *const i8, ...);
}
enum Foo {
A,
B(i32),
}
fn main() {
let result = Foo::B(123);
match result {
A => unsafe {
let a = "A\n\0";
let b = a as *const str;
let c = b as *const i8;
printf(c);
},
Foo::B(x) => unsafe {
let a = "Result: %i\n\0";
let b = a as *const str;
let c = b as *const i8;
printf(c, x);
},
}
}
I expected to see this happen: compile without error Instead, this happened:
rust1: error: expected [Foo{A = 1, B (0:Int:i32:(Ref: 28 TyRef: 8[8,]):bounds:[])}] got [<tyty::error>]
<source>:13:5: error: failed to type resolve expression
13 | match result {
| ^
<source>:10:1: error: expected [()] got [<tyty::error>]
10 | fn main() {
| ^
Meta
- What version of Rust GCC were you using, git sha if possible.
No, this isn't matching the Foo::A variant. In this case A is a pattern binding. You could have written foo and it would have worked the same. Rustc gave you three warnings to notify you of this.
warning[E0170]: pattern binding `A` is named the same as one of the variants of the type `Foo`
--> <source>:14:9
|
14 | A => unsafe {
| ^ help: to match on the variant, qualify the path: `Foo::A`
|
= note: `#[warn(bindings_with_variant_name)]` on by default
warning: unreachable pattern
--> <source>:21:9
|
14 | A => unsafe {
| - matches any value
...
21 | Foo::B(x) => unsafe {
| ^^^^^^^^^ unreachable pattern
|
= note: `#[warn(unreachable_patterns)]` on by default
warning: unused variable: `A`
--> <source>:14:9
|
14 | A => unsafe {
| ^ help: if this is intentional, prefix it with an underscore: `_A`
|
= note: `#[warn(unused_variables)]` on by default
@bjorn3 thanks so much i missed this! This is a huge relief that this is the case as I was trying to figure out how that would work lol
This isn’t actually fixed; the test added in #2072 behaves differently on gccrs than on rustc: Godbolt link (Expected output is A because A is a pattern that matches anything.)
See also https://github.com/Rust-GCC/gccrs/issues/2082#issuecomment-1681246225
@narpfel yes, I had reopened https://github.com/Rust-GCC/gccrs/issues/2144 as a follow up. We closed this issue because the original code would not compile, even though it should (with warnings)
