Is it acceptable to add new formatter for emacs-lisp, which is emacs package itself? (elisp-autofmt)

[ed9w2in6]

I have added this formatter in my own config:

(define-format-all-formatter
 elisp-autofmt
 (:executable)
 (:install "M-x package-install elisp-autofmt")
 (:languages "Emacs Lisp")
 (:features region)
 (:format
  (format-all--buffer-native
   'elisp-autofmt-mode
   (if region
       (lambda () (elisp-autofmt-region (car region) (cdr region)))
     (lambda () (elisp-autofmt-region (point-min) (point-max)))))))

I have been using this for a while, which works pretty well, and IMHO much better than the default one (Emacs).

May I ask if this is acceptable for this package to add a formatter that depends on another emacs package? And what do you expect in a pull request?

[lassik]

May I ask if this is acceptable for this package to add a formatter that depends on another emacs package?

Yes, absolutely. The following already supported formatters are Emacs-native:

• auctex
• emacs-bibtex
• emacs-bibtex-sort
• ledger-mode

Of these, at least ledger-mode is a third-party package that does not ship with Emacs.

For native formatters, the main thing is to use (funcall 'elisp-autofmt-region ...) instead of (elisp-autofmt-region ...). The later form assumes that the function is defined, which it won't be if the package is not installed.

We don't currently have a standard way to check whether or not a native package has been installed. I can add that at some point. For now, your code above is fine. Just do the funcall thing.