Improve unresolved use error message

[kornelski]

"use of undeclared type or module foo" doesn't mention that it could be a crate.

This error can happen when users forget to add a dependency to Cargo.toml, so I think it's important to mention that it could be a missing crate.

I've used a heuristic based on Rust's naming conventions. It complains about an unknown type if the ident starts with an upper-case letter, and crate or module otherwise. It seems to work very well. The expanded error help covers both an unknown type and a missing crate case.

[rust-highfive]

r? @matthewjasper

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

[kornelski]

WDYT?

if ident.name.with(|n| n.chars().next().map_or(false, |c| c.is_ascii_uppercase())) {
    (format!("use of undeclared type `{}`", ident), None)
} else {
    (format!("use of undeclared crate or module `{}`", ident), None)
}

[estebank]

WDYT?

That might not work well with unicode_identifiers when it is stabilized, but that can be a problem for "future us". This logic would be in the "happy path", right? Looking at the code it seems we would only hit this in the case of Err(Determined) which I would expect (but don't know for sure) would an "always emit an error path".

It seems like the new sample for E0433 is giving an E0432:

failures:

---- /checkout/obj/build/x86_64-unknown-linux-gnu/test/error-index.md - Rust_Compiler_Error_Index::E0433 (line 7746) stdout ----
error[E0432]: unresolved import `rand`
 --> /checkout/obj/build/x86_64-unknown-linux-gnu/test/error-index.md:7747:5
  |
3 | use rand::Rng;
  |     ^^^^ maybe a missing crate `rand`?

error: aborting due to previous error

For more information about this error, try `rustc --explain E0432`.
Some expected error codes were not found: ["E0433"]

failures:
    /checkout/obj/build/x86_64-unknown-linux-gnu/test/error-index.md - Rust_Compiler_Error_Index::E0433 (line 7746)

You probably want to use some long-ish unlikely to ever exist crate name.

[kornelski]

@estebank since that didn't sound like an objection to the principle, I've included the heuristic.

[bors]

☔ The latest upstream changes (presumably #74862) made this pull request unmergeable. Please resolve the merge conflicts.

[kornelski]

There's something odd with the build checking whether expected error happens in error documentation:

CI check sees this:

use ferris_wheel::BigO;
  |     ^^^^^^^^^^^^ maybe a missing crate `ferris_wheel`?

Rust playground sees this:

use ferris_wheel::BigO;
  |     ^^^^^^^^^^^^ use of undeclared type or module `ferris_wheel`

The difference is due to the check:

rust/compiler/rustc_resolve/src/lib.rs

Line 2383 in 7ec1de0

let (label, suggestion) = if module_res == self.graph_root.res() {

where the graph_root is documented as "// For rustdoc.". I don't know what that's supposed to be.

The playground consistently shows a different error than Rust's CI, in all of: bin crate, lib crate, regular rustdoc comment, and compile_fail test.

For now I've relaxed the documentation compile_fail check. The docs reflect what Rust does outside the CI environment, because that's the version that is relevant to users.

[matthewjasper]

@bors r+

[bors]

📌 Commit 7ec1de0 has been approved by matthewjasper