Nathan Ridge

Results 1248 comments of Nathan Ridge

Could you provide a code example please, and indicate where in the code you are performing the code action?

Thanks, I see now: when invoking the quick fix for clang-tidy's "invalid case style", which is to rename the variable, only occurrences in the current file are renamed, not occurrences...

I wonder if perhaps a solution could build upon the proposal in [this RFC](https://discourse.llvm.org/t/rfc-enhancing-clang-tidy-with-project-level-knowledge/63960) to "enhance clang-tidy with project-level knowledge". Then perhaps clang-tidy could produce the multi-TU quick fix itself.

@tom-anders very interesting :) > In theory, clangd could "upgrade" the clang-tidy quick fix from a "local rename" to a "global rename". But to do this, clangd would need to...

> I can try implementing this in clangd and see what the feedback is - if it's not accpected, at least then we'll have an editor-agnostic patch that people can...

This already works in clangd 18 (added in https://reviews.llvm.org/D150124).

> While the compiler is gcc, clang-specific arguments like `-resource-dir` and `--` separator are generated, causing the compilation to fail. The log output is a bit misleading in this respect:...

> ``` > Unknown argument: '-rdc=true'clang(drv_unknown_argument) > Unknown argument: '-gencode'clang(drv_unknown_argument) > ``` In some other issues filed about CUDA (https://github.com/clangd/clangd/issues/1610, https://github.com/clangd/clangd/discussions/1863), I have seen people remove those flags using clangd's...

> The real C++ directory seems to be under `/usr/lib/gcc/x86_64-linux-gnu/12/../../../../include/c++/11`. or more shortly `/usr/include/c++/11` This may be a case of https://github.com/clangd/clangd/issues/1394 then, see workarounds in [this comment](https://github.com/clangd/clangd/issues/1394#issuecomment-1328676884).

I don't have a strong opinion on this, I'll just make two minor points: 1. Technically, `--query-driver` is not _just_ a security mechanism for the user to confirm that drivers...