Multiple source files, global variables

[costeastefan]

What is the recommended way of using llama across multiple source files? For example having functions that use llama variables/objects (e.g. View) across multiple files and be accessed in other files?

[bernhardmgruber]

Regarding includes, you can either use the single header

#include <https://raw.githubusercontent.com/alpaka-group/llama/single-header/llama.hpp>

in every file you use LLAMA, or just install LLAMA/use it via a package manager/CMake sub_directory and #include <llama/llama.h> wherever you need it.

If you have shared definitions, like a common record dimension, or tags, or views, or accessors, etc., you could put those in a shared header file. You should be able to use LLAMA features like you would use ordinary C++ structs etc.