Nathan Ridge

This observation may only be tangentially relevant, but there's precedent for treating paths as distinct types in the [Rust standard library](https://doc.rust-lang.org/std/path/struct.Path.html). EDIT: And, in fact, even in the [C++17 standard...

There is a [workaround](https://github.com/clangd/vscode-clangd/issues/687#issuecomment-2365439284) that came up in one of the linked issues, copying it here for visibility: > What I'm seeing in the log, is that vscode is sending...

> Is there any plan or timeline for addressing this issue? There is work in progress towards a fix at https://github.com/llvm/llvm-project/pull/136439. Meanwhile, the workaround I recommend is to normalize the...

This is caused by the clang parser's typo correction feature. Since `some_b_value` cannot be resolved, and `some_a_value` is close in edit distance, it gets typo-corrected `some_a_value`, and thus shows up...

I agree. As a user, I think what I'd like to see is to have clangd categorize the typo-corrected reference as a "tentative" reference (as opposed to one it's confident...

I don't think we'd want to disable typo correction altogether by default. That has impacts on other features than rename, in particular diagnostics -- in the event of an actual...

In clangd 16, the destructor is now also marked as `declaration` (thanks I believe to https://reviews.llvm.org/D136212), and in addition both the constructor and destructor now get a `constructorOrDestructor` modifier (https://reviews.llvm.org/D134728)....

> Do you have any knowledge about overloaded operators? I would like to extract that information from clangd but currently I don't see any way other than duping the AST...

I'm going to reopen this to track the proposed change to the semantic token type for constructors and destructors (from `class` to `method`).

Can you clarify how you obtained the screenshot on the top vs. the screenshot on the bottom (I assume one of them is clangd, and one of them is something...