how to make helix work with c++ and cmake

[veranovus]

I'm trying to make a c++ app using cmake as the build system but it seems like helix is failing at understanding my include paths for some reason.

Is there a way to get rid of those red dots without including headers with absolute path?

[the-mikedavis]

clangd looks for a compilation database (compile_commands.json or compile_flags.txt) that tells it how to compile the project and find includes. See https://clangd.llvm.org/design/compile-commands#where-do-compile-commands-come-from

I'm not very familiar with this myself but you should be able to setup CMake to emit this https://cmake.org/cmake/help/latest/variable/CMAKE_EXPORT_COMPILE_COMMANDS.html, or use Bear

[sinefabula]

For small programs I keep a simple compile_flags.txt in my $HOME directory which just lists common search paths (I'm on macOS so I need to specify paths to boost etc.) and my usual compile flags. It looks like this:

-std=c++17
-Wall
-I/opt/homebrew/opt/boost/include
-I/opt/homebrew/include
-L/opt/homebrew/lib

For larger projects, I usually have a simple script which just runs cmake with -DCMAKE_EXPORT_COMPILE_COMMANDS=ON to generate compile_commands.json and then deletes the CMake cache.

[veranovus]

Thanks you so much after exporting the file from cmake, moving the compile_commands.json to the root of my project directory solved the problem.

[goyalyashpal]

it's also not detecting my #include iostream and namespace std, hence, all the cout etc are also shown undeclared

[AlexisTM]

In my case, I just added the following in my helix .helix/languages.toml.
This ensures the compile_commands.json file is always up to date 👍🏼

[[language]]
name = "cpp"
language-server = { command = "clangd", args=["--compile-commands-dir=compile_commands_directory"] }

[ZennMystic]

So I come across this post when I was looking into trying out Helix for C/C++ programming. I have clangd and llbd-vscode working. But what now?

Way back when I used Vim, I have a key bind of alt-m that would run ./build.sh %... % being the C file in question. The build script would have the compiler flags I wanted and the compiler I wanted to use and were I wanted the output file to go with its name all that. But I read Helix does not have % for filename.

How do you do the same type of thing in Helix?

is this were:

[[language]]
name = "cpp"
language-server = { command = "clangd", args=["--compile-commands-dir=compile_commands_directory"] }

comes in?

Right now I am doing ctrl-z running my build script then fg to go back to /helix.

I am confused to say the least. So any help would be good. Thank for your time.

[AlexisTM]

When you use cmake, you can generate the compile commands file (a big JSON file). This file stores all compile commands for every single target of your project.

You can use: -DCMAKE_EXPORT_COMPILE_COMMANDS=ON in your cmake configure step so on the next build you get a compile commands file. You can then link the build directory containing that file with the configuration above.

https://cmake.org/cmake/help/latest/variable/CMAKE_EXPORT_COMPILE_COMMANDS.html

[ky3ow]

Don't know how to make it single keybind but you can do :sh ./build.sh <C-r>%
:sh is alias for run-shell-command
<C-r>% inserts % registry(full file path) to command prompt

[ristomatti]

In my case no additional config was needed, the actual problem was this https://stackoverflow.com/questions/74785927/clangd-doesnt-recognize-standard-headers