I’m having trouble telling to clangd where to search for my includes, so I want to use the inc folder inside the root folder of my project, but I can’t manage clangd to resolve my include paths, so I tried:

• adding a compile_flags.txt containing, amongst other flags, -Iinc
• generating a compile_commands.json file from my makefile, I just started this project so I can show the full file:

[
  {
    "directory": ".",
    "command": "clang++ -Wall -Werror -Wextra -pedantic  -Iinc -MMD -std=c++20 -c src/utils/invoke-daemon.cxx -o build/utils/invoke-daemon.o",
    "file": "build/utils/invoke-daemon.o"
  },
  {
    "directory": ".",
    "command": "clang++ -Wall -Werror -Wextra -pedantic  -Iinc -MMD -std=c++20 -c src/main.cxx -o build/main.o",
    "file": "build/main.o"
  }
]

• using .clangd, but I saw that it is not a valid option since .clangd doesn’t support relative paths in flags.

Even with this compile_commands.json at the root of my project, when I try to include a file contained in my inc folder.

I after saw that the root_directory was set to /home/$USER, so I changed it to be the current working directory, but I still get the same problem, clangd seems to just ignore compile_flags.txt and compile_commands.json

So the only way I found that works to get the right include paths is to use .clangd and put absolute path in -I flag, because everything else I tried doesn’t work. And I want to include my config in git, and not having to run a script each time to generate the .clangd with the right absolute path.