rustc_on_unimplemented supports referring to trait

[estebank]

Add support to rustc_on_unimplemented to reference the full path of
the annotated trait. For the following code:

pubmodBar{#[rustc_on_unimplemented = "test error `{Self}` with `{Bar}` `{Baz}` `{Quux}` in `{Foo}`"]pubtraitFoo<Bar,Baz,Quux>{}}

the error message will be:

test error `std::string::String` with `u8` `_` `u32` in `Bar::Foo`

[rust-highfive]

r? @pnkfelix

(rust_highfive has picked a reviewer for you, use r? to override)

[estebank]

Originally found that this would be a nice addition when working on #43001, to make sure that the path of the trait is always correct on the error message.

[Mark-Simulacrum]

[00:03:27] Expected a gate test for the feature 'on_unimplemented'.
[00:03:27] Hint: create a file named 'feature-gate-on_unimplemented.rs' in the compile-fail
[00:03:27] test suite, with its failures due to missing usage of
[00:03:27] #![feature(on_unimplemented)].
[00:03:27] Hint: If you already have such a test and don't want to rename it,
[00:03:27] you can also add a // gate-test-on_unimplemented line to the test file.
[00:03:27] tidy error: Found 1 features without a gate test.
[00:03:27] some tidy checks failed
[00:03:27] [00:03:27] [00:03:27] command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0-tools/x86_64-unknown-linux-gnu/release/tidy" "/checkout/src" "--no-vendor" "--quiet"
[00:03:27] expected success, got: exit code: 1
[00:03:27] [00:03:27] [00:03:27] failed to run: /checkout/obj/build/bootstrap/debug/bootstrap test src/tools/tidy
[00:03:27] Build completed unsuccessfully in 0:01:00
[00:03:27] make: *** [tidy] Error 1
[00:03:27] Makefile:74: recipe for target 'tidy' failed

[aidanhs]

@estebank looks like the tidy failure is still waiting on you? Might be worth getting it fixed before review in case it reveals any other test failures.

[aidanhs]

Oh wait, you just pushed a change :)

[aidanhs]

Some other test failures have been revealed

[01:09:26] error[E0232]: this attribute must have a value
[01:09:26] --> $DIR/bad-annotation.rs:26:1
[01:09:26] |
[01:09:26] 26 | #[rustc_on_unimplemented] //~ ERROR this attribute must have a value
[01:09:26] | ^^^^^^^^^^^^^^^^^^^^^^^^^ attribute requires a value
[01:09:26] |
[01:09:26] = note: eg `#[rustc_on_unimplemented = "foo"]`
[01:09:26] [01:09:26] error[E0230]: there is no type parameter C on trait BadAnnotation2
[01:09:26] --> $DIR/bad-annotation.rs:30:1
[01:09:26] |
[01:09:26] 30 | #[rustc_on_unimplemented = "Unimplemented trait error on `{Self}` with params `<{A},{B},{C}>`"]
[01:09:26] | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[01:09:26] [01:09:26] error[E0231]: only named substitution parameters are allowed
[01:09:26] --> $DIR/bad-annotation.rs:35:1
[01:09:26] |
[01:09:26] 35 | #[rustc_on_unimplemented = "Unimplemented trait error on `{Self}` with params `<{A},{B},{}>`"]
[01:09:26] | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[01:09:26] [01:09:26] error: aborting due to 3 previous errors

[estebank]

r? @aturon

[arielb1]

r? @arielb1

I'll try to review this later today.

[arielb1]

I don't think rustc::traits should be having an error reporting logic distinct from typeck, but let's leave that refactoring to a separate PR.

[arielb1]

LGTM

@bors r+

[bors]

📌 Commit 05d3526 has been approved by arielb1

[arielb1]

@bors rollup