Clojurians Log v2

Do you ever run out of ideas?

I've been wanting a book or guide along the lines of "writing elisp for the clojure-minded". This is not that, but it might be an even better idea 🙂

(That said, for writing/maintaining elisp intended for an audience wider than the already clojure-infected, adding this as a dependency might not be preferred. This is not a criticism of clj!, just me thinking out loud about the use-cases I had in mind.)

Yeah I know. It's probably for end users only. There are Clojure to Elisp transpilers out there but I don't know how well their code is to just check that into source control

Very nifty, though, for the end user! 🙂

That's very cool!

Would love to see/know about people that migrated they whole config to it and how it plays in performance those cases

Just ask your robot to do a POC?

maybe! but that will mean lose my emacs for some time during the tests hehe

you can totally write another config and launch a headless emacs on the side?

hum I think so

you should measure the second invocation of a .clj file though, the first time it's processed, but then byte-compiled to .elc

I just measured here. the cljbang "runtime" takes 2ms to load. a compiled ns with 1000 functions 1ms

but of course it's be better to measure a real project

sounds awesome!

Allow me some level-headed reasoning. Please don't get me wrong Michiel, this absolutely looks cool and I hope it grows into something far more practical than "just cool". From pure syntactic/cosmetic point, perhaps it's not hugely defensible. Idiomatic, modern Elisp can look nearly identical. Here's elisp and cljbang:

(defun visible-major-modes ()
  (thread-last (window-list)
               (seq-map (lambda (w) (buffer-local-value 'major-mode (window-buffer w))))
               (seq-uniq)))

(defn visible-major-modes []
  (->> (el/window-list)
       (map (fn [w] (el/buffer-local-value 'major-mode (el/window-buffer w))))
       distinct))

(defun stale-buffers ()
  (thread-last (buffer-list)
               (seq-filter (lambda (b)
                             (when-let* ((f (buffer-file-name b)))
                               (not (file-exists-p f)))))
               (mapcar #'buffer-name)))

(defun stale-buffers ()
  (cl-loop for b in (buffer-list)
           for f = (buffer-file-name b)
           when (and f (not (file-exists-p f)))
           collect (buffer-name b)))

You are right @ag but I'm not good at remembering the elisp API and language. Just like bash. You know what I did for bash.

Like I write in the README: > ⚠️ I'm not sure if any of this is a good idea, but it kinda works for me.

Well, bash had no chance over Clojure. Problem with Elisp is that it is "worse is better" case. It's bad but not as bad. Like Python. I wish Python was bad enough to drive people away from it, but for most people it's just not that obviously bad.

I agree but still :)

Not trying to sell it to people who love elisp

Oh, but I don't love Elisp. I love Clojure. I merely tolerate Elisp. 😄

I wonder how it compares to https://github.com/BuddhiLW/clojure-elisp , this project is also still active it seems

@a2svior I'm sure it's an awesome project but it needs an external transpiler process which imo isn't really ergonomic while you're writing elisp functions, unless if you have way to invoke the transpiler, but this comes with a lot of tooling while elisp already has a good REPL-modeL

with cljbang everything just runs in emacs and you don't have to bring anything else

Makes sense, didn't think about that. Curious to try it out. Btw the author of the other project is also here I think @pedrogbranquinho

Cool! Is cljbang like squint and https://github.com/borkdude/clj2el is more honest to clojure? I have my own clojure -> elisp compiler. I'm still finding some parity bugs and I'm still adding benchmarks and I'm not dogfooding so interop hasn't had polish. I do hope you'll check it out when I'm done.

no clj2el is not more honest to clojure, it was also squint like but didn't even have a stdlib/runtime

more honest to clojure would be adding HAMT etc which imo is defeating the point of a thing that runs in emacs to configure emacs

@ag syntax-quote now works

Whoa, I feel like I'd have to send my dog to you, lol. I love my dog, slow down a bit with this awesomeness.

Why does hamt defeat the point? Erm, the point you're getting at is probably just different from why I wanted clojure on emacs lisp. I did it because I went bankrupt with emacs lisp again but I've never gone bankrupt with clojure so...

Now, since you brought this up, @borkdude I wonder how difficult would be to implement HAMTs in pure elisp...

Oh I'm sure it's possible, elisp is flexible enough

> Why does hamt defeat the point? Erm, the point you're getting at is probably just different from why I wanted clojure on emacs lisp. My bar isn't the high currently. A full faithful clojure runtime would isn't what I aim for here, I let elisp do the heavy lifting whenever I can.

I also want interoperability with other parts of emacs to be high, without having to go back and forth.

Yea. That makes sense. Sorry, I first read that as that hamt wouldn't make sense in absolute terms. Mb. I went as far as implementing java classes so that I could run real clojure libraries like clojure.tools.deps.

lol

Yea, idk how interop will shake out. I have singletons for true/false and nil/empty list.

that sounds way more ambitious

It has consumed most of my ai subs for over 10 weeks now.

this is so exciting!!! congrats!

Congrats!

Been using datalevin the last few months in a project at $JOB, extremely cool, and very happy to see it hit the big 1.0(.0)! Can't emphasize enough how patient @huahaiy and the rest of the #C01RD3AF336 people have been with all my embarrassingly-newbie questions gratitude

Congratulations!

Yay! Congratulations on the big release!

I haven't had the chance to use Datalevin in production yet, but I've been following this project with interest.

If you have issues or questions, please raise them as separate threads in #C03S1KBA2 or #C06E3HYPR so we can respond individually, thanks!

Thank you for fixing CLJ-2897 although, at this point, I no longer need it 😐 but I may do again in the future perhaps.

it was a good call regardless