Nathan Ridge

Likely related to clangd's preamble optimization (see e.g. https://github.com/clangd/clangd/issues/200#issuecomment-554973739). Assuming the real file in question has some `#include` directives at the top, is it important that the `#pragma` cover them?...

> For a more long-term solution: Would it make sense for `clangd` to end the preamble when encountering the first `#pragma clang diagnostic push`? I guess that depends on how...

It seems to be the combination of `-target aarch64-linux-gnu` and `-mabi=lp64` that clang doesn't like. The first one causes the type of the `TargetInfo` to be `AArch64TargetInfo`, which then [only...

> How to dump the error log? It depends on your editor, please see https://clangd.llvm.org/troubleshooting#gathering-logs

I'm trying to understand what is being proposed here. Let's take a concrete example of an input file `input.yacc` containing a C++ fragment, and an output file `output.cpp`. Are you...

Would the transformation be limited to cases where the source fragment in the input file appears verbatim in the output file? If so, then I think I see how this...

I think it rather depends on #412. As I understand it, #412 itself is mostly about diagnostics involving functions like `make_unique`.

> Forwarded parameters are now being resolved for inlay hints following [llvm/llvm-project@a638648](https://github.com/llvm/llvm-project/commit/a638648fef76146634c2199ce7b90c4bc6bfed01) Nice, thanks very much for your work on this! > I'll take a look at signature help next,...

clangd considers the two `vdev_init` functions to be the same symbol (its index acts like a project-wide linker in that sense). However, if you perform "find references" on one copy...

> Is there a particular reason that clangd is making these assumptions? The initial design of clangd's indexer predates my involvement in the project, so I can't give an authoritative...