My team is currently researching ways to relieve Ruby of some heavy lifting. The majority of the team is familiar with Ruby, but not necessarily Go. I wanted to experiment with using the C interface to keep the Ruby and Go code together, versus breaking into separate services and passing JSON (or other) back and forth. My intention was to make Go as approachable as possible.
This direction keeps the majority of the domain logic in Ruby, but allows the passing of most data types to Go. Using Ruby FFI, the code reads as if you are calling Ruby methods, but in fact, Ruby is embedding the Go and calling exposed Go methods via C.
It boils down to writing Go methods that are faster than the Ruby equivalent due to memory management, etc. Then compiling the Go to expose those methods. Then on the Ruby side, you create a module that embeds the Go library and explicitly exposes the C style methods. Ruby can then call those methods, passing things like Arrays, Strings, Integers, etc. Some data types require some conversion when going from Ruby to C to Go.
My intention is to go through the data types creating simple examples others can reference when passing data from Ruby to Go and vice versa.
I'm creating directories for each data type. Included in each directory is an example.rb file for the Ruby logic. And a Go file containing the Go logic. Also included are the two files created during the Go compilation. At the top of each Go code file I am including the command to compile the Go code.
The Ruby FFI project has a wiki that discusses some of the oddities for things like Arrays and Structs (aka Hashes). Below are references to the data types I will be working through.
https://github.com/ffi/ffi/blob/master/lib/ffi/types.rb#L74-L150
https://github.com/ffi/ffi/wiki/Types
Feel free to implement a data type or make a current implementation better.
Just fork and create a branch named after the data type and issue a PR.