r? @petrochenkov

rustbot has assigned @petrochenkov.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

rust-analyzer is developed in its own repository. If possible, consider making this change to rust-lang/rust-analyzer instead.

cc @rust-lang/rust-analyzer

The job pr-check-2 failed! Check out the build log: (web) (plain enhanced) (plain)

[RUSTC-TIMING] indexmap test:false 0.566
    Checking postcard v1.1.3
[RUSTC-TIMING] postcard test:false 0.266
[RUSTC-TIMING] syn test:false 2.942
error[E0063]: missing field `entire` in initializer of `proc_macro::bridge::DelimSpan<_>`
   --> src/tools/rust-analyzer/crates/proc-macro-srv/src/server_impl/token_stream.rs:104:31
    |
104 |                         span: bridge::DelimSpan {
    |                               ^^^^^^^^^^^^^^^^^ missing `entire`

Okay, it looks like for whatever reason pr-check-1 expects entire not to be there, but pr-check-2 expects entire to be there. They must not be using the same source??

-_-

locally I am able to execute ./x test without any issues

There is actually a script to run the CI jobs locally with Docker: https://rustc-dev-guide.rust-lang.org/tests/docker.html#testing-with-docker

From what it seems at first glance, the tests are just running ./x check, but there might be some extra config being added that makes the tests run which isn't the default.

There is actually a script to run the CI jobs locally with Docker: https://rustc-dev-guide.rust-lang.org/tests/docker.html#testing-with-docker

From what it seems at first glance, the tests are just running ./x check, but there might be some extra config being added that makes the tests run which isn't the default.

I wondering if it's because rust analyzer is a separate repo? Is pr-check-2 pulling the source from the repo?

rust-analyzer is a subtree intentionally to avoid issues like this: you can edit the code in this repo if it needs to be fixed to get both to compile, and then at the next regular merge someone will make sure the repo and the subtree match.

rust-analyzer is a subtree intentionally to avoid issues like this: you can edit the code in this repo if it needs to be fixed to get both to compile, and then at the next regular merge someone will make sure the repo and the subtree match.

Which is what I did, I am just not how or why one check is using the older version not in the PR. I tried looking at the CI directory for those briefly nothing immediately sprung out at me.

Yeah, that's mostly why I recommended trying the docker approach because it will give you a better idea of what's running and thus what's failing. Since you can investigate the containers yourself, it'll tell you what's wrong.

That is weird ... rust-analyzer should be building against the intro library sources ...