i wonder if i can come up with a good tinylisp solution to this challenge

dealing with ragged lists does seem to be something tinylisp is great at

@des54321 Dew it ^_^

CMQ: Best color of ethernet patch cable?

working on it now

main problem is theres no great way to compare lists with ints

however (l list int) does evaluate to (), instead of 1/0, so there might be something i can do with that

@Neil nil is already defined in the standard library as ()--did you mean treat () as the empty lazy list, or did you mean create a new value that can be used as the empty lazy list?

@DLosc im pretty sure its working, now im going through the pain of typing out a gazillion test cases into tinylisp format

posted

@DLosc yes I mean treat it as an empty lazy list

@RadvylfPrograms Okay so apparently using the 192.168 IP instead of my server IP gets me 40x faster speeds

@allxy hack font? lol

@RadvylfPrograms how is my codepage ^?

@Neil Oh, okay--so basically what you're saying is lazy-nil doesn't have to be a function, it can be a sentinel value that's explicitly checked for. Yes, that seems to work, and it's a bit simpler.

(def lazy-nil (q lazy-nil))

(def lazy-empty?
 (lambda (lseq)
  (equal? lseq lazy-nil)))

(def lazy-head
 (lambda (lseq)
  (if (lazy-empty? lseq)
   lazy-nil
   (lseq 0))))

(def lazy-tail
 (lambda (lseq)
  (if (lazy-empty? lseq)
   lazy-nil
   (lseq 1))))

(def lazy
 (lambda (ls)
  (if ls
   (substitute (q &ls)
    (quote ls)
    (lambda (switch)
     (if switch
      (lazy (tail &ls))
      (head &ls))))
   lazy-nil)))

(def lazy-cons
 (lambda (val lseq)
  (substitute (q &lseq)
   (quote lseq)

@PyGamer0 Some of those characters aren't going to render the correct width in most fonts, but looks good

Is this going to be a 7-bit language?

@PyGamer0 I sense APL and BQN influence

@RadvylfPrograms Yeah, BQN looks great in the font on its website, but then you copy-paste it over here and you have to look twice to verify it's even the same characters.

@PyGamer0 checkmarks? diode electrical symbols? sure why not, I suppose "golfing lang" and "typeable on a normal keyboard" are already contradictory goals so you might as well go all the way

I really like what Jelly does, with all of its codepage being typeable on a single keyboard layout (US-INTL)

Copy-pasting is still common, but there's another option for people who use it a lot

Or do what (Charcoal, or maybe Canvas?) does and have a readable way to type it, that gets compiled into the actual code by your online interpreter

it's Charcoal that does that

Yeah I get those mixed up lol

That's...actually pretty true lol

Although Unicode does restrict you a bit I think

Even when typeable-on-a-keyboard and renders-the-right-width don't hold you back

@des54321 i don't see a diode? lol

@des54321 uh i have mostly meaningful names for builtins

diodes

those aren't diodes

⍋⍒

It's APL's grade-up and grade-down

@PyGamer0 by names i mean symbols, your symbols mean something in relation to what they do, but at the cost of not being typable on a normal keyboard

@lyxal theyre basically diode symbols, i guess minus the line at the tip

@des54321 lol

@RadvylfPrograms ^

@des54321 'CHRISTMAS TREE WITH POLE THROUGH IT' (U+234B) and 'UPSIDE DOWN CHRISTMAS TREE WITH POLE THROUGH IT' (U+2352)

CHRISTMAS TREE‽‽‽‽‽‽

@DLosc yes you do

github.com/PyGamer0/flax/blob/main/docs/elements.txt#L173

´ is foldl1, ` is scanl1, ⋈ is pair

there is some k influence too :p

W is where

‘each right, ’ each left

@des54321 Well, those would have been the names if they'd asked me ;)

@des54321 hey i can pick all 3, i will make a keyboard layout for flax :P

@PyGamer0 okay now thats called cheating :P

Reason #23 I love LiveOverflow

I really need to stop being so hard on myself about the time I spend being unproductive

codepage feedback?

Yeah, the lack of proper monospacing is what really jumps out at me. :(

@PyGamer0 I don't like that

Those look really similar, it's fine for tihngs like string literals but with different operations it'll get confusing

Also ≢ renders really badly on GitHub (at least in Chrome)--looks like there's a slash through the following two characters rather than the ≡.

@PyGamer0 Are there mnemonics behind the choices of accented letters?

yeah

flax looks like a cursed mix of apl and bqn with sprinkles of jelly: +´⍳10

mmm

mmm

yeah the sbcs WORKS

time to create a new release

ooh we surpassed 2⁹ commits

just need docs search like MATL...

@PyGamer0 Whoa. Pip only just passed 2⁷.

tfw afking on an ipad

@PyGamer0 only 2^9?

Cringe.

Talk to me when you reach 2¹¹

We can't all be Vyxal

@DLosc I don't see why you can't use nil as the sentinel

@RadvylfPrograms Otherwise you'll be going out to the internet and back in to your home network

I think a good enough router should be able to prevent that from needing to happen

I'm still wondering what he means by his server ip

I mean, it should be using 192.168.x.x already

I was assuming his server also has a public IP

I can't imagine it has its own public IP

@Neil ...... Oh. I, uh, wow. I guess you can. That wouldn't have worked in my original implementation because it expected lazy-nil to be a function like all other lazy sequences, but your version doesn't need that to be the case, so... yeah. Wow.

you can then write something like (def lazy-head ((lazy-list) (q(and lazy-list (lazy-list 1)))))

bah, I still put the q in the wrong place

tinylisp doesn't have a short-circuiting and (due to limitations in the macro system), but that's pretty close to what I wrote:

(def lazy-head
 (lambda (lseq)
  (if lseq
   (lseq 0)
   nil)))

@DLosc yeah, I've just discovered that limitation

hmm, I tried (d defun (q (() args (v (c (q d) (c (h args) (c (c (q q) (c (t args) ())) ()))))))) but that also echoes the name :-(

Yeah. Again, macro limitations: since you're eval'ing the cobbled-together call to d inside the macro, it's not a "top-level call," and therefore the interpreter doesn't know to quiesce it.

The macro system I came up with for Appleseed evaluates the result after it returns from the macro, which I think would solve that problem. That's one of the things I want to change if/when I make a tinylisp 2.

@DLosc Can't it be something like (i x y x)?

@DLosc the other thing that would be nice is a way to choose which arguments are evaluated

Hm yes, for web

Short-circuiting isn't really necessary in tinylisp due to no side effects

oh, there are side effects

@allxy Well yes... but either you make it a lambda and you can't do short-circuiting, or you try to make it a macro and you can't evaluate x and y in their original scopes, which causes problems.

@allxy Short-circuiting is crucial for recursion, tho

Also, yes, disp is a side effect, and you can do some wacky stuff with d if you try hard enough (though you can't redefine something once it's already been defined).

(q (q args (body))) would be a variadic macro, (q ((arg1 (q arg2)) (body))) would be a macro whose first arg was evaluated but whose second wasn't

if the body was (i arg1 (c (q q) (c arg1 ())) arg2) that would then be a short-circuiting or (and is easier because all falsy values evaluate to themselves)

@DLosc I guess, yeah

@Neil Hm, interesting! I'll file that idea away. I also like the idea of differentiating macros from functions by nesting the list a level deeper rather than adding an extra nil at the beginning.

@lyxal :(

time to make dummy commits

or once you've written the b macro, (i arg1 (b (q (u arg1))) arg2)

@Neil hold on what am I missing that (i arg1 arg1 arg2) wouldnt simply work, why does it have to evaluate to (q arg1) to return arg1?

@Neil So I'm looking at how I wrote these in Appleseed, and I realized it's actually a lot simpler than that because the macro system is basically just a rewriting pass before the expression gets evaluated. For example, either (2-argument or) is (macro (&expr1 &expr2) (if &expr1 &expr1 &expr2))), which means (either x y) gets turned into (if x x y) and then evaluated. The only drawback is that x has to be evaluated twice.

Unfortunately, it's far past my bedtime, so I've got to go now.

o/

o/

o/

o/

@DLosc yeah I wanted to avoid that

@des54321 this is under @DLosc's proposal that the return from the macro gets evaluated, which is what I want for arg2 (because it hasn't been evaluated yet) but not for arg1 (because it has)

i managed to speedup nvim

its takes 100ms to startup now!

(i use like 40 plugins)

lazy loading is good

@RadvylfPrograms modding but with clickbait title

youtube.com/clip/UgkxYuvFlTPIIB3rcfcy5BuUWPcGbpNhm7Up

lol

@Ginger Not really

That's not what he's doing

He's reverse engineering the packets the game's sending to the server, and injecting new ones through a MitM

oh

that's cool

he's not really doing the reverse engineering, people on wiki.vg have already done that

Well, he's doing part of it

Not the hard part, but still

I had a horrible idea for beating the tinylisp solution to Shortest code to produce infinite output, but unfortunately it seems that (load ) suppresses all output from loaded files

okay errors still produce output, but now im just abusing this language

(X
(load A

when saved in A.tl, this produces several lines of errors, but not infinite for some reason

Error: referencing undefined name X
Error: referencing undefined name X
Error: () is not a function or macro
()
Error: referencing undefined name X
Error: referencing undefined name X
Error: () is not a function or macro

I guess you cant really recurse with (load) statements

thats probably a good thing

Unlike many esolangs, tinylisp is actually kinda practical :P

practical, except for the fact that you need to import the library to use mathematical operations as simple as multiplication

or, phrased another way, tinylisp+library is practical, pure tinylisp isnt

DLosc seems to be a fan of minimal languages, which I can get behind

Just tinylisp without the library can still do everything, right?

it can do everything, but you basically have to re-define most of the library

because the library is just a bunch of built-in definitions from the base operations

also of course the fact that tinylisp has absolutely no input reduces the practicality of the lang

Oh lol

although,,,, I suppose you could theoretically juryrig some form of input using (load)

CMQ: Falsey or Falsy?

hmmm thats a good one

I usually write falsy but falsey does seem better honestly

falsy

Is it me or is there never anything in the review queues? Lol.

I've certainly never seen anything in them, are they broken somehow?

@pajonk "False eye" of course, smh

I would not think so. Maybe the site has too many active users, if such a thing is possible.

It has too few active users, I think, but there are a few users who are very active in the review queues

CMQ: how do you pronounce "quine": kine or kwine?

kwine

I always thought it was kwine as well.

@Pyautogui It's been a few months since I last did a review on the meta site, and I still show up as the 5th most recent reviewer in the Reopen votes queue

Interesting. So a small minority has a monopoly on all the reviews, effectively?

Guess so

There's also just not much activity :(

Really? The site seems very active, to me at least.

@des54321 kwine

@Pyautogui It's mostly site regulars, though. Not a lot of new users coming in

Ah, understood. Thanks for the explanation.

I mean, there are new users arriving all the time, but it's not so frequent that we get First Post reviews or question closing reviews every day

Makes sense.

I'd like to see what percent of our reviews come from each category

Possibly suggested edits

That would be interesting!

@pajonk Falsey

@pajonk Falsey

@des54321 you can still input via function parameter though can't you?

@Neil yes, but theres no actual program input, you can really only send input in from another tinylisp program

yeah but I mean function parameter is an acceptable input format for code golf

true

but im saying that, its still not a practical language if it cant take anything from stdin

if you wanted a practical language you'd just use lisp anyway

I've been attempting to make a tinylisp solution to Fault-Tolerant Hello World, and I have come up with something imperfect but sort of workable, I think it fails for less than a dozen of the possible removals

okay it actually fails for 18 removals, but out of a 200 byte file thats not that bad

Impressive

I made this diagram with google drawings lol

For yet another addition to my famous™ weighted random array item answer

tfw you google "weeb worker"