Comparison with guile-git

[nicolas-graves]

I have a question regarding the comparison with guile-git. I found that the way guile-git defines the procedures that are exported much more elegant than in guile-ssh.

Do you know if there is a "hard" reason behind that ? Could we imagine migrate from one to the other ? Maybe we could ask Ludo for advice?

[artyom-poptsov]

Hello. This is a little bit vague. Could you provide examples here what you consider more elegant? Do you mean that Guile-Git uses FFI and all that?

[nicolas-graves]

FFI and bytestructures mostly.
Most functions there look more or less like this :

(define stash-apply
  (let ((proc (libgit2->procedure* "git_stash_apply" `(* ,size_t *))))
    (lambda* (repository index #:optional (options (make-stash-apply-options)))
      "Apply a single stashed state from the stash list of REPOSITORY.
INDEX is the position of the stash in the list (0 is the most recent stash).
OPTIONS is an optional stash apply options structure."
      (proc (repository->pointer repository)
            index
            (stash-apply-options->pointer options)))))

which, as a guiler, am more comfortable with