Nathan Ridge

Ah, I didn't realize it worked by putting OpenCL code inside a string; I thought it was extensions to the C/C++ syntax itself. (Just goes to show how little I...

@DaanDeMeyer @sam-mccall The patch at https://reviews.llvm.org/D114003 has been accepted -- should we merge it to mitigate this issue for the time being?

> I could have sworn I pushed that patch to the repo, maybe I messed up somewhere? Ah you're right! It has indeed been [committed](https://github.com/llvm/llvm-project/commit/5a6dac66db67225e2443f4e61dfe9d2f96780611), the commit message just didn't...

> If I have an error such as a missing reference to another function inside `do_stuff`, clangd doesn't issue any error. I guess there are two places where we could...

On a related note, I would be remiss if I didn't take this opportunity to plug **C++20 Concepts**, which allow you to describe constraints on your template parameters in code,...

It would be helpful to have a smaller code example that triggers the crash than the entire esp-idf project. > And the content of compile_command.json is wrong, but I think...

> Where can I get clangd with debug info To be honest, I'm not sure about this myself. When clangd crashes for me, I do get symbols in the backtrace....

Relevant place in the code is [here](https://searchfox.org/llvm/rev/22b5adff7189f3ca547f38d5e87a7fa9dda9c509/clang-tools-extra/clangd/FindSymbols.cpp#210). What is currently being printed is specifically the function's **type**, hence the lack of parameter names.

Since the crash is occurring inside the implementation of a clang-tidy check, you should be able to work around it by disabling this particular check, e.g. by adding the following...

Probably somehow related to clangd's preamble optimization.