famous last words lol

Possibly. I am curious as to what the challenges are

For instance, Cursive has something that lets you extract to a let block

There’s some LSP stuff around that. I am sure it isn’t trivial

But, why is it impossible? at least the dumbest case

(let [x 2 y 4] (+ x y 7)) if I select the inner form, why can’t it be bound into a defn with all unknown/unresolved args supplied?

yeah not sure what the complexities are. I just kinda laughed at your wording. It’s not the best way to request features from people

This isn’t a feature request

I am genuinely curious

If it came off as a feature request, then I am sorry

it might just be that it hasn’t been prioritized over other things. there are multitudes of low-hanging fruit in every project that gets put aside for more critical things

I see that clj-refactor.el for Emacs supports extract function

maybe it is easy and could implement it in calva without much trouble. I know the devs are really focused on getting jack-in functionality figured out, new REPL, formatting, etc.

Thanks

It's possible to implement using the middleware

I'd love me some pull requests for refactorings. ❤️

In theory it should be enough to add clj-refactor as a middleware and set up the appropriate menus/keybindings.

And then for one of my favourite jokes:

What’s the difference between theory and practice? In theory, none.

Yes, there seems to be a difference between theory and practice here.

(defun cljr-extract-function ()
  "Extract the form at (or above) point as a top-level defn.

See: "
  (interactive)
  (cljr--ensure-op-supported "find-used-locals")
  (when (cljr--asts-y-or-n-p)
    (save-buffer)
    (cljr--goto-enclosing-sexp)
    (let* ((unbound (cljr--call-middleware-to-find-used-locals
                     (buffer-file-name) (line-number-at-pos)
                     ;; +1 because the middleware expects indexing from 1
                     ;; +1 more because point has to be inside the sexp,
                     ;; not on the opening paren
                     (+ (current-column) 2)))
           (name (unless (cljr--use-multiple-cursors-p)
                   (let ((highlight (cljr--highlight-sexp)))
                     (unwind-protect
                         (cljr--prompt-user-for "Name: ")
                       (delete-overlay highlight)))))
           (body (clojure-delete-and-extract-sexp)))
      (save-excursion
        (cljr--make-room-for-toplevel-form)
        (insert (cljr--defn-str))
        (if name
            (insert name)
          (mc/create-fake-cursor-at-point))
        (newline)
        (indent-according-to-mode)
        (insert "[" unbound "]")
        (newline-and-indent)
        (insert body ")"))
      (insert "(")
      (when name (insert name))
      (save-excursion
        (unless (string-blank-p unbound)
          (insert " " unbound))
        (insert ")"))
      (unless name
        (mc/maybe-multiple-cursors-mode)))))

https://github.com/clojure-emacs/clj-refactor.el/blob/master/clj-refactor.el#L2979

So it’s a mix between middle-ware and elisp 🙂

FWIW, the functionality I use the most from clj-refactor is:

1) cljr-add-project-dependency, this is huge when greenfielding https://github.com/clojure-emacs/clj-refactor.el/blob/master/clj-refactor.el#L2261

2) cljr-add-missing-libspec https://github.com/clojure-emacs/clj-refactor.el/blob/master/clj-refactor.el#L2907 which is bound to /, so when you type foo/ clj-refactor will search for previous aliases foo and insert the appropriate :require

3) cljr-clean-ns https://github.com/clojure-emacs/clj-refactor.el/blob/master/clj-refactor.el#L2742

Thanks @U04V5VAUN! Is it possible you could add this as an issue on github? I'm worried it will suffer from the 10K slack death sooner than we can pick up things where you've left them here. I pinned it and starred it in the hope that I might be able to keep it somehow, but I think that might be naïve.

https://github.com/BetterThanTomorrow/calva/issues/137

Wonderful. Greatly appreciated!

No worries 🙂

It’s a hobby of mine to follow this project 🙂