add callback argument to lumo get-completions

[hlolli]

lumo.repl/get-completions takes two arguments.
https://github.com/anmonteiro/lumo/blob/master/src/cljs/snapshot/lumo/repl.cljs#L1252
So I added map str function into the callback, now I'm getting completions suggestions from company-mode.

---

Before submitting the PR make sure the following things have been done (and denote this
by checking the relevant checkboxes):

• [x ] The commits are consistent with our contribution guidelines
• [x ] The new code is not generating bytecode or M-x checkdoc warnings
• You've updated the changelog (if adding/changing user-visible functionality)
• You've updated the readme (if adding/changing user-visible functionality)

Thanks!

[hlolli]

Yes, the completions work, but arglists aren't appearing while scrolling the completion list via company. Could you test this and the other pull request I made, or look over it. I see that on lumo side, sometimes lumo gives wrong completions, there is to say, all possible completions, in my case after ) character.

[arichiardi]

I think it would be more clear and maybe correct to use a prn or (print (pr-str ...)) here just to make clear that this is a callback that writes out. I was even wondering why it is working now..We pass an anonymous callback that returns a list...however the return of get-completions is nil. Is it? I am confused 😀

[arichiardi]

So the callback thing is a pain (notice the nil at the end because of the print) :

cljs.user=> (lumo.repl/get-completions "redu" #(print (map str %)))
(reduce reduce-kv reduceable? reduced reduced? reductions)nil

[hlolli]

That nil should go trough the elisp read function and become ignored when it hit the case statement (because it's not in quotes and therefore treated as elisp nil). I regret making two pull requests as I see this are two different problems interleaved with one another.

[arichiardi]

I'd solve it this way:

(let [hack (atom)]
  (lumo.repl/get-completions "redu" #(reset! hack (map str %)))
  @hack)

cljs.user=> (let [hack (atom)] (lumo.repl/get-completions "redu" #(reset! hack (map str %))) @hack)
("reduce" "reduce-kv" "reduceable?" "reduced" "reduced?" "reductions")

[hlolli]

Very good! Agree, I update the pull request and use your snippet 😄 Besides the nested % sign, doesn't want to quote it against emacs format function. Use fn [] instead.

[arichiardi]

Fixed.

[hlolli]

Are you commenting on the correct pull requests? This fixes it or you've fixed it?

[arichiardi]

No sorry I have fixed the comment with the snippet, I cannot fix the PR 😄

[arichiardi]

This has another problem due to pretty printing, I will raise it in lumo.

[hlolli]

PR updated.

[arichiardi]

As per the linked lumo issue, this avoid pretty pretting:

cljs.user=> (do (set! lumo.repl/*pprint-results* false) (let [hack (atom)] (lumo.repl/get-completions "redu" #(reset! hack (map str %))) @hack))
("Reduced" "reduce" "reduce-kv" "reduceable?" "reduced" "reduced?" "reductions")

Don't know if it matters actually, I haven't tried in inf-clojure, yet.

[arichiardi]

swap! accepts a function, reset! is better.

If you set it too:

(do (set! lumo.repl/*pprint-results* false) (let [hack (atom)] (lumo.repl/get-completions "redu" #(reset! hack (map str %))) @hack))

and squash all commits into one, we are good 😄

[hlolli]

I'll do that and squash. But isn't this a bit mean to someone who wants to prettyprint the results when emacs is always setting it to false?

[arichiardi]

Uhm, Antonio suggested it but you are right, it is global...Maybe this is the moment when we introduce a post processing step to completion. First I would try if it works with multiline lists, if not we have a fn that makes everything single line but we need to introduce some glue for post processing..

[hlolli]

There are functions that can do that, but as far as I see, it returns a list of strings which is read as emacs data, so newline should be irrelevant, in my company-mode completions, makes no difference.

[hlolli]

Ok I changed swap! to reset! it works and gives completions, and also squashed the commits.

Little related @arichiardi , there's a bug in lumo

This can be reproduced in lumo like this

(let [ret (atom)] (lumo.repl/get-completions "red)" (fn [res] (reset! ret (map str res)))) @ret)

Otherwise I think @bbatsov can merge this now.

[arichiardi]

LGTM