@PyGamer0 me too

do we want the whole discussion moved too?

chris mccandless into the wild

yes

@lyxal go on

yes

→ 42 landlubbers moved to Off-Topic TNB

:61221749 I bought it just lazy to use it

@RadvylfPrograms gosh fricking dang it I was just about to do that :p

42 unfinal nonanswers

I've been move-ninja'd by caird and HN plenty of times, it's your turn n00b :p

still the only RO to not have moved any messages from here :thumbs-up:

@PyGamer0 why the brainfrick is this starred

it is what it is

things just get randomly starred and no one really knows why

all you can do is /shrug and accept the pentagons

can you 12 and unstar that :P

because its not even funny

thanks to whoever to did it

I didn't do it lol

Messages should require two independent hidden stars before showing up on the starboard

@RadvylfPrograms I agree

that's lowkey a good idea

→ 38 "shapes" moved to Off-Topic TNB

@RadvylfPrograms good idea

congration, youy movd msaeges!

I done did

and now because it's 12:26am and because I spent 10 minutes with all those messages selected ready to act, I'mma go to bed

o/

bye

o/

Me gonna go too

replit is cool but I wish it didn't disconnect every 10m and make me have to sit and wait doing nothing for 30 seconds

also they've managed to torture poetry so badly it takes another 10m trying to install dependencies

@Ginger I know I'm supposed to not be here right now but have you ever considered using python anywhere?

nope

You should

And why not.

It doesn't disconnect

It stays permanently on

Plus github integration is easy

because replit lets me make infinite repls in lots of languages

and because something something sunk cost fallacy

But lambda.chat is python. PA is a perfect candidate for temporary hosting

Plus you get cool terminal stuff, meaning you can use poetry from the command line

I like what I have now and don't want to bother moving it

Your loss :p

if it ain't broke don't fix it

@Ginger sounds broke to me

¯\_(ಠ_ಠ)_/¯

@lyxal i read poetry as poverty. . .

Whoa...TIL about for await

@Ginger you mean "replit has designed itself for vendor lockin"?

probably

@pxeger To be fair it's probably more like "replit isn't meant for serious projects"

I dunno about yall but something seems a bit wrong here

can't quite place what it is tho

looks fine to me

@Ginger shadows are too heavy for such tight spacing between cards

looks fine imo

@pxeger actually yeah, I agree

oh was "something seems a bit wrong here" a joke? lol

.M.

actual design feedback should go here, of course

the bane of my life is JS deciding to randomly fetch over HTTP... for some reason

why is it doing this? what did I do to incur the wrath of the JS gods?

Like...idk...security?

Why on earth are you using HTTP?

whoops typo

Oh nvm

why

Looks like a 30x might be being returned, downgrading to HTTP

lemme check

yup :|

TIL document.URL was a thing

because of course JS has two ways of doing the same thing

@pxeger only 2?

@RadvylfPrograms Then she will tell the hooman, "No no, first make it sunny and then open the door" :P

@flawr If you're using is, everything in the right-hand side has to be bound to a value first, I think. That's why it works when you give it K and ask for F, but not the other way around.

In theory, clpfd mentioned by Unrelated String should help, but I don't understand it well enough to get it to work. I'm more familiar with Brachylog, which I think does some extra magic on top of clpfd.

Hmm. Also, it looks like it's backtracking over possibilities it shouldn't.

This is pretty clunky, but it seems to work. Dunno what the error message is about. :P

Actually, since the challenge in question guarantees that you'll be given an actual Fibonacci number, the J < F+2 check isn't necessary (removing it causes an infinite loop when F is not a Fibonacci number).

Better

@RadvylfPrograms why is a redirect downgrading to HTTP not considered a security issue, but fetching an image over HTTP is?

What? No I'm saying maybe replit's redirecting to HTTP, which is being considered a security issue

@DLosc I thought something like that!

@DLosc right, I first tried to reverse it, and that works for the first two but not for the sum at the end

@DLosc what the freck?

@flawr The problem there was that you'd said fib(0,0), but you hadn't said that fib(F,0) should fail for any nonzero F. So for example if the recursion tries fib(3,0), it should fail because 3 is not the 0th Fibonacci number, but instead it matches the fib(F,K) case and keeps recursing infinitely into negative values of K. (Assuming that the is problem is resolved first.)

@flawr Do you have a more specific question? :P

@DLosc oh now I think I start to understand what the issue was!

@RadvylfPrograms I meant "considered a security issue [by the browser]"

Yeah, so did I

Am I missing something?

@RadvylfPrograms yes, the time when we were young and without burden or worry.

It's warning because it's trying to load JSON data over HTTP when it's on an HTTPS site

regardless of the fact the HTTP is because of a redirect

@flawr oof

If the original HTML document had been redirected to HTTP, it wouldn't have warned at all (except if HSTS is enabled)

@DLosc I'm wondering whether it is possible to unclunkify it a little bit

@pxeger Oh, so by "a redirect downgrading to HTTP" you're talking about the page itself?

Well, in general

Probably for backwards compat and since HSTS exists

The warning would equally be generated by <img src="http://example.com"> without any redirect

I feel like we're talking past each other

@DLosc But thanks a lot for all your help so far!

I've got a challenge idea but I barely know how to describe it LOL

wait i dont have time to try and explain, ill just sandbox later oops

@flawr Looks like a bottom-up approach to the recursion ends up slightly shorter than your original double-recursion approach (and much faster for big numbers). Here's what I came up with ungolfed.

And here's a golfed version using this golfing trick.

@DLosc what does fib(F,0,A,_) :- !, F=A. even mean? Why the ! make sense here? (As I understand it, F = A would never be unified.

(Still based on the idea that we need a predicate to generate the nth Fibonacci number and then another predicate to try all possible values of n until we find the right one, though. Therefore, still clunky.)

@flawr ! means "after this, don't backtrack on this predicate." In this case, it means "If the call to fib matches the fib(F,0,A,_) definition, don't ever try the fib(F,N,A,B) definition." That's necessary to prevent infinite recursion.

Another way of accomplishing the same thing in this specific case would be to test whether N>0 in the fib(F,N,A,B) definition: TIO

... Which, now that I think of it, saves bytes in the golfed version.

@DLosc but I thought for F = A to apply you'd have to put the exclamation mark on the right of it, no?

(apparently this is not true, but I can't figure out where I'm going wrong)

@DLosc I think you should post this:)

@flawr No, it keeps on going, it just can't backtrack past the ! if it ever tries to backtrack.

@flawr Thanks, will do :)

oooh

! is like a one-way valve, if you will

i lied im going to try and describe the challenge idea here so maybe someone will say "thats obviously stupid because <x>" or "that is known as <x> so you should google that for better terminology" before i write up a really badly worded sandbox

yeah

so i was looking at that and someone pointed out to me that theres one pair of pokemon where both of them have no effect on eachother

Any feedback? (in particular, the issues I've highlighted at the bottom)

rather than the usual, like, one has a great effect and the other has no effect

and i was wondering if you could make a meta chart, where the axes are "interaction 1 way" and "interaction the other way"

so if Rock covers Paper, and Paper is covered by Rock, you could input (covers, is covered by) and the output would be (Rock, Paper)

but maybe if you input like, (covers, covers), thered be no output

since theres no pair where X covers Y and Y covers X

so youd be given the list/table of interactions e.g [x verb y, x verb z, y verb x...] as input? and maybe be tasked with making the meta-table?

im having a rough time describing it apologies if im just rambling lol i just didnt want to have to sound this stupid in the sandbox

x A B C
A 1 1 2
B 2 1 1
C 2 1 2

=>

x | 1                 | 2
--+-------------------+------------
1 | (A,A),(B,B),(B,C) | (A,B)
--+-------------------+------------
2 | (A,B)             | (A,C),(C,C)

something like this but less ascii artish

and possibly less redundant across the diagonal in the output

CMQ: Opinions on what the 20 most useful pieces of punctuation (excluding periods and spaces) are, for my attempt at making a chat program on top of Radvylf's Chat on Chat program (currently I have ,!?'"();:/*-+=&%$@<>)

‽ … – —

¤

i love currency symbol so much

does the order matter?

order is irrelevant

ok fair

my first things that id say are definitely agree that ,?() are incredibly important as they provide very helpful semantic clarification

Interrobang is an almost good suggestion honestly, the biggest problem is its hard to type on most keyboards

Adám's the only person I've ever seen use one, and ?! is basically the same

fair enough

Currency symbol could be nice

yeah ,()?! are all definitely essential, as well as ;:'" mostly

i do have the dollar sign in there i believe

how about _

^

not sure how i feel abt it but its worth considering

^^ (^)

the ones I'm least sure about are definitely &%$@<>

+-*/_= are all pretty important

how about ^ LOL

@'s basically useless, I'd replace with |

@RadvylfPrograms definitely, i want at least basic math to be typeable

another good consideration is ability to make smiley faces

ಠ is essential then

yeah | is probably more useful than @, and i think & really isnt useful enough to justify

yea cause you just say "and" & it gets the point across

how useful is a percent sign really

not very

Modulo for programming discussions, otherwise not much

50% of statistics are made up though

what to replace it with though.... ^ maybe?

yea

^

Or _ if you don't have that

^ is good for math, chat lingo, and making the ^_^ face

the only other thoughts i have for things i could be missing are `[]{}`, but i dont have room for all of those

Yeah, and they're not super important

its ok if you dont have all the braces

one pair is fine

fair

btw what is this whole chat on chat thing

i saw messages which appeared to be encoded into japanese chars but i dont know the like.. context of why theyre encoded

,!?'"();:/*-+=_^$|<> seems good i think

chat.stackexchange.com/chatonchat

You go to chat.stackexchange.com/chatonchat with a userscript, and you can chat on chat

oh, huh

but why does it need to be like, encoded

why not

Since 1. that makes it way more fun to design and 2. I'm going to be adding additional features to it that chat doesn't already have

right ok

got it

makes sense now

@RadvylfPrograms there's always the chat improvements userscript for additional features

No like, features which require an additional channel for communicating information between the users

E.g., editing messages at any point after they were posted

(But it is better in the non-additional-information-communicated sense too, since it has CommonMark and MathJax support)

@RadvylfPrograms How does thtat work?

Protocol version cc 02 includes an ID with every message. Another message can be posted referencing that ID to edit its contents.

Hm oaky, but how do you authenticate that?

The client just ignores any edits which wouldn't be allowed

But the protocol doesn't?

No, since there's no server that could do that

And no need to

So... I could send a maliciously encrypted message in the regular chat room and it could edit anyone's?

No, because the client will see that you didn't send the original, and ignore it

Oh :P

tfw the art department sends an email in Comic Sans

Postpostmodernism :P

Todo: anarchy chat client that lets anyone edit other peoples messages and the names attached to those messages

Let’s see people sign chat messages with private/public key encryption to protect against edits

@ATaco 0_0 that would be pure chaos lol

@ATaco thats a horrible idea, i love it

my great idea for a chat client was this: no names, no editing, no separation of different users, you get a line that no one else can write on and you write a message on it, once you post it youre kicked to the next available line

This is getting close to 4chan

Although what you're suggesting is probably more anarchy-y than 4chan

i remember when tumblr let you edit other peoples posts

that was a fun time