Nomi

Yeah, that's fair. In rules_pico that define is associated with pico/config.h and gets propagated transitively to everything that depends on it, rather than being associated with the FreeRTOS library. I...

Aha, I think I see. Thanks for those pointers to the clangd issue; I wasn't aware of the transferCompileCommand behavior added in https://reviews.llvm.org/D116167. It looks like the root issue might...

Sorry, I should have been more clear! No, the compile_commands.json looks like this: ([full gist](https://gist.github.com/NomiChirps/834c67898314a4d4a53ddebbb481fce8)) ``` [ { "file": "test.cc", "arguments": [ "G:/code/tools/gcc-arm-11.2-2022.02-mingw-w64-i686-arm-none-eabi/bin/arm-none-eabi-gcc", "-U_FORTIFY_SOURCE", "-D_FORTIFY_SOURCE=1", "-DFREESTANDING", "-fstack-protector", "-Wall", "-fno-omit-frame-pointer", "-mcpu=cortex-m0plus",...

Clangd verbose log for rules_pico: https://gist.github.com/NomiChirps/85afddb252bb9699b2d0d59a2a3f43f3 For mingw-gcc; however, I can no longer reproduce the compile_commands.json in the table above. This log is with [this compile_commands.json](https://gist.github.com/NomiChirps/744f72a68eead84499eba3a16b035b53), which excludes all header...

Okey dokey, here's as clean and complete of a reproduction as I think I can manage: https://github.com/NomiChirps/compile-commands-testing There's no rules_pico or alternate toolchain involved, and I'm now on Linux instead...

That fixes it well enough for me! Thank you for spending so much of your time chasing this down :)

Thank you for looking at this! I won't be able to check until Monday but I'll definitely come back and take another look then. Meanwhile -- I don't think it...