« first day (4548 days earlier)      last day (592 days later) » 

00:21
i ended up golfing it down to s.match(d+"\\b")?0:"."
c--
c--
thoughts on this?
00:37
@c-- maybe you could add a visualization of sample I/O for easier understanding (ascii-art is ok)
c--
c--
something like a screenshot of LifeViewer?
that would be great
c--
c--
ty
00:52
also an interesting test case: solutions to "unique father problem"
01:45
> When you pick a random value, make sure it's odd by setting the lower bit; even values are hardly ever prime
Swift’s Bool.random is, iirc, Int64.random() >> 17 & 1. No idea why 17 in particular
I know some PRNGs have issues with adequate randomness in their lower bits
Maybe they want to protect against that?
02:03
interesting
02:14
it is said that middle bits in the internal state are "more random" than other parts
@RydwolfPrograms Sure, but why 17 specifically? There's no comment or anything
02:33
insert xkcd random number joke here
0
A: Sandbox for Proposed Challenges

l4m2Inspired, maybe copy background from there Given a map, output a shortest possible input. Test cases: Start here _ / => / \ \ / / Start here _ /\/ => /// /\/ \/\ \/\ / / /\ Start here _ /\/\ => ...

@SandboxPosts Is which of { shortest, O(1)*shortest, O(n)*shortest, unlimited } better?
@Bubbler i need a browser extension for those
c--
c--
@l4m2 I don't get it
03:19
@Bubbler How is that relevant here?
I feel like that xkcd gets wayyy overused
 
2 hours later…
05:29
TIL at 1:30 in the morning that “we do a little trolling” is a Trump quote
It's called we do a little trolling
i read that message like five times and have no idea what it means
@noodleman it's the following line
He says "we do a little trolling. It's called we do a little trolling"
i agree
05:51
mmm gotta love -6 rep from a drive by downvote spree
Someone doesn't like either Vyncode or Luminespire
the flags actually
Well whoever it was missed a recent post :p
Maybe the lumenspire promotion could be placed outside the code block? That way you can format the link properly and it wouldn't take as much space with the blank line
It's there in case I ever decide to put control characters at the bottom of the explanation
For metadata purposes
It's very big currently, especially for short explanations
05:54
True, but it provides a guaranteed selection point
Also that's right I didn't actually put the link there only the emoji
@mousetail also also I just remembered that the answerer decides what to put in the code block
Luminespire doesn't generate the triple backtick
The link doesn't have link formatting
I'll go make a change to the generator
i finally took a look at Vyxal's flag list, and it turns out that not all of them are actually that abusive
It's hard to create that many abusive flags
though the way they're used still don't feel in the spirit of the flag decision
05:59
@mousetail
Should it be edited into existing posts or only for new ones?
like, a flag that pushes 100 at the start of the program can only be abused. There's almost no reason that you can't just do that in within the actual program
hence the downvotes (with some upvotes sprinkled in)
@JoKing fwiw the H flag has only been used about 3 times for challenges where the answer is already going to be trivial (with one exception where it's used for novelty). Version 3 won't have it because it's just not even useful
my totally objective list ends up at 16 almost guaranteed to be abused, 14 abusable, and 18 whatevers
06:09
We have 48 flag types?
Wow
oh, and i guess i downvoted a couple of thunno posts too, sorry @TheThonnu
like, surely your language has a flatten builtin, why do you need a flag for it
6
@JoKing it's a sum flag, but if you use sum on a nested list in Python it flattens it by one level
So I've used the S flag for that purpose a few times
i see, your language must lack a sum builtin then
I don't think flags for what you don't have a built-in for are good ideas
but flags for stuff you already have builtins are totally okay??
like, the sum builtin is S as well, you're literally moving code into the command line
06:16
But even if you use the old system where you have to add the length of the flags to the bytecount, flags in Thunno 2/Vyxal will sometimes save you bytes
It's not just for this system where flags are free
@JoKing what would this be for Thunno 2?
@TheThonnu I call it "syntax skill issue" :P
3
the old system still gave you a "discount" on moving code into the flags you know
you can have an entirely different set of instructions in your flags that don't have to follow the same rules as your main code
@TheThonnu 23/5/8 and one i don't know what this means
it just seems like most of these are set up as against the spirit of "these are all different languages", when they're not at all treated like that.
instead, they're "i saved a byte or two by moving them into the flags, haha"
06:57
huh, I was reading luminespire as luminesphere
07:45
@Bubbler I used to read it as luminspire
@JoKing well, in vyxal, the programme □∑ returns the correct result for a sum of numbers. But then, adding the HṪ flags make it return an incorrect result for a sum of numbers.
08:09
0
Q: Landmine Number IV

Jeremy CollpravBackground You have again x3 been given the task of calculating the number of landmines in a field. But this time, we have no landmine number, and some landmines have exploded. You must calculate the landmine score given a string of numbers. This is the landscape. You have no landmine number to s...

@PlaceReporter99 i didn't downvote that one though
i mean, i didn't upvote it either, since the flags don't seem necessary and it feels like you're trying to make it rather than actually golfing it
e.g. the code could just be empty lmao
 
1 hour later…
10:14
I don't really know what the link should point to though, I don't have seperate pages for each builtin and building all of them would be significant effort for single sentence explanations
I'd like to link to the original wiki but it doesn't have ids set in a way that allows me to easily link to one specific builtin
@mousetail you don't need to have a separate page for each. Just have like a yaml/json/whatever file mapping built in to single sentence and dynamically render the page
10:29
Main issue is the way I've set up my deployments making even one new page will require excavating the foundation of how things are set up
Actually no main issue is I'm lazy
Couldn't you just make a second html page?
And point to that url instead of adding it to index.html?
Or even a separate repo
Doesn't have to permalink to the interpreter directly
11:07
@lyxal well done on becoming a moderator on PLDI!
:)
Thanks :p
11:34
0
A: Sandbox for Proposed Challenges

I am kind of a language dev\$n\$ bottles of beer! You want to recreate the song "99 Bottles of Beer", but you don't have 99 bottles! Given an input non-negative integer \$n\$, output the lyrics to the song "99 Bottles of Beer" but starting from \$n\$ instead. For example: Input: 10 Output: 10 bottles of beer on the wal...

 
2 hours later…
13:25
0
A: Sandbox for Proposed Challenges

Jeremy CollpravChamber of Reflection Background A ray of light is fired from the top left vertex of an MxN Chamber, where M a denotes the width and N denotes the height of the chamber. The ray of light advances one grid space per second. Given that T is the time frame, calculate the number of reflections in thi...

14:24
GIVE ME BUTTER FOR FREE ROBUX
Might be answerable as actually, I think I see some patterns
4
@NewPosts that will need to be moved to trash at some point.
oh wait at the end of their second answer to the question:
>
> kickoff idk off off
>
@NewPosts CALLING @everyone FLAG THIS POST AS SPAM
No need to use all caps
sorry
yes it's been deleted!
can someone move a few of these messages to trash now?
@mousetail although that one is interesting... maybe star it in the nineteenth bakery?
Jul 7 at 14:58, by Unrelated String
the tags feels weirdly fitting :P
This one is now also without context
We'll never know what the fitting tags where
14:40
@mousetail let me dig in tnb
I mean I know the context since I was there
@mousetail not in tnb
I'm in tnb right now
2
i mean message isn't in tnb
but i found messages of posts bullying sklogic
Because it was moved
14:49
@mousetail what was moved?
Because it was spam
@mousetail the posts bullying sklogic were worse than spam.
That's why they where moved
@mousetail shouldn't they be deleted
?
There is no delete option
That's why we always move content that is not appropriate for whatever reason
14:52
really?
Can't mods delete such posts?
:63977233 Self-deletion is different
Mods can I think, not ROs
The New Posts bot won't self-delete
@mousetail what was the context?
14:54
If you read the convo it's pretty obvious
The 3 tags told a story
MORE BUTTER, MORE ROBUX
For some reason when I saw that from TGC, it showed up as being posted by caird rather than NP
15:26
0
Q: Reverse Conway's Game of Life

c--Conway's Game of Life is a cellular automaton "played" on an infinite grid, filled with cells that are either alive or dead. Once given an initial state, the board evolves indefinitely according to the following rules: Any live cell with 2 or 3 living neighbours lives to the next state Any dead ...

@mousetail That's really neat, but unfortunately, it causes underlining of the code characters, which is problematic for APL that has both and and both and and both and .
It's also not helpful for mobile users
Ah, true.
@Bbrk24 At least it shouldn't harm mobile users. Graceful degradation
@Adám Wish there was a way to add a title without a link :/
@Bbrk24 Well, actually, each could be a link to the docs for the built-in.
15:37
I imagine Bbrk means the letters are too small to effectively click an individual one on mobile
Long press will display the title if you manage to click one
You can zoom in by putting two fingers on the phone and dragging them apart from each other.
You can't usually zoom in quite far enough to reliably click one individual letter, and even when you can it's pretty tedious
@Adám It is, but what mousetail said
I'm still experimenting with techniques, if anyone has ideas how to trick SE markdown to behave more like what I want I'd be interested
Can you do <span title="...">code here</span>?
You'd need to use <pre><code> rather than triple backtick
15:42
Span doesn't seem to survive
Even on elements that do survive only <a> seems to preserve the title attribute, and then only if href is a valid link
Oh nice
This page seems to confirm only img and a are allowed to have titles
I could make every letter a img...
I think I'm a genius, you can just use images with broken URLs. They will just display their alt text and be invisible. But the title still works
Disadvantage is it takes a moment to pop in
16:28
Why not have one image ready-made for every built-in, the image showing the symbol in a good font?
copy-paste difficulties, perhaps?
That's true of the URL too, though
@mousetail Long press normally selects the text; on mobile, it's impossible to select a link except by selecting some non-link text near it and then moving the cursor.
@Bbrk24 There'll be a TIO link that you can open and copy the code. Also, how many really need to copy the code?
What if TIO is down and/or I want to test it somewhere else?
Perhaps not a common case, but still something to consider
Code could appear twice; as text and as images. It is anyway common to repeat for explanatory purposes.
CMC: Shortest href attribute that SE posts render as a clickable link.
@Bbrk24 Invalid images with only alt tags like I'm doing now have the same issue
@Adám https://. was the shortest I could find
16:37
I have significantly shorter.
ftp://.
yup, that's mine.
ftp doesn't seem to work in images unfortunatly, http doesn't work either
Will it accept ws://.? I imagine not but
Nope :/
16:40
"Here, get the image via websocket."
Can you do paths within SE? like /tags for example
Doesn't seem to work either
17:29
0
A: Sandbox for Proposed Challenges

bigyihsuanDiagonalize a vector code-golfarraymathmatrixlinear-algebra Input A vector, list, array, etc. of integers \$\mathbf{v}\$ of length \$n\$. Output A \$n \times n\$ matrix, 2D array, etc. \$A\$ such that for each element \$a_i \in \mathbf{v}\$, $$ A = \left( \begin{array}{ccc} a_1 & 0 & \cdots & 0 ...

17:57
i wanna do a poll, how many people here have learned to solve a rubik’s cube at some point
@Bbrk24 i use those sometimes for linking to user pages, it definitely renders them as links
@noodleman I have 3x3 through 7x7, pyraminx 3x3 and 4x4, pyramorphix, megaminx, meffert's gear ball, sudoku cube, mole cube, and void cube
I also have the master kilominx but I never learned the parity algorithm for it
oh yeah and the siamese cube and the skewb
I don't own a mirror cube but I know how to solve one. I'd love to try the ghost cube sometime
18:15
@noodleman i used to solve 3x3s, but now i can only do the top and the first 2 layers
0
A: Sandbox for Proposed Challenges

l4m2https://chat.stackexchange.com/transcript/message/63012181#63012181 Make a 0-byte metagolfscript solution that outputs its name. Shortest name wins.

I have a 4x4 but never looked up the proper way to solve it, so it takes me like 3 hours to solve, and only if the parity gods smile upon me
I own a 3x3 but have not learned to solve it, so I put it in a configuration I found aesthetically pleasing and set it on a shelf :P
lol
@noodleman i can (slowly) solve any nxn cube you can throw at me, i've got {2,3,4,5,6,7,10} cubes
not particularly fast though, best 3x3 is about 45 seconds? not much faster than that
i have timed myself solving the 10x10, took me an hour
i also have pyraminx, master pyraminx, and a megaminx, all don't know how to solve
(by memory)
(megaminx is the closest one, i don't have memorized the last layer yet)
18:34
Megaminx is actually pretty easy to solve, the first and last layers are very similar to 3x3
At least, as far as I remember
Yes it's just a 3x3 with four middle layers
@bigyihsuan That's pretty impressive, have you ever tried 11x11?
it's not difficult, it just takes forever to find all the pieces
@user Once you get past 6x6 they don't get harder, it just takes longer
Ah, same techniques apply?
@user yep
pretty much everything past 5x5 uses the same algorithms, generalized to higher order cubes
18:37
6x6 is only harder than 5x5 because you can catch and prevent parity errors a lot more easily on odd-sized cubes
beyond that is just novelty
there's a 21x21 out there for sale, but it's $1500, and as i said, just a novelty
@Bbrk24 yep, a dodecahedron is just a 3x8x3 cuboid
i used to be super into cubing and still have a bunch of weird ones (my favorites are the megaminx and dino cube) but i’m getting back into it
but yeah i had a feeling that a lot of people who frequent a chatroom about code golf would be good at rubik’s cubes lol
@DLosc rotating each middle layer of a solved cube twice? :P
18:52
Ooh those look really good
Even better on a mirror cube if you get the long part to stick out
19:18
@Seggan-OnStrike I don't even remember, lol
But yes, it's about five or six moves away from a solved cube
@user just got a megaminx, it's just 3x3x3 but twice
my only problem is all colors look the same to me
especially white/tan/yellow I confuse a lot
is it blasphemous to sticker a gan megaminx with better colors?
Yes. Do it.
Would be pretty easy to get out a pen/permanent marker and use dots/slashes/whatever to distinguish colors
Or get a mirror megaminx :P
@RubenVerg Mine has two shades of yellow, two shades of blue, and two shades of green
mine has lime green and standard cube green, which I easily distinguish, and azure and standard blue, which I somewhat distinguish
I think I might replace either yellow or tan with black, a la sq1
The number of times I've done "now I get the yellow-green-blue corner... no, the other one"
oh yeah I do have a square 1. but I don't know the algs off the top of my head
19:33
I never fully learned sq1, just like the back and white opposite sides many have
I wonder if stickering my gan megaminx is gonna be hard with those indented corner pieces
@Bbrk24 I never remember that there are five corners per color so when I do bear paws I see three wrong ones and then when I see the fourth with the required color I insert it before remembering it is not the right one :)
though I'm, like, a couple weeks in, I'm sure I'll get used to it
still doing beginner method + some f2l and s2l for simple cases
I think I'm in the three minutes range, sub 3:30 ao5s
20:22
I can solve 1x1 cubes in less than a second.
20:37
@noodleman I can do the 1x1 through 3x3 including the void cube
@Bbrk24 which method do you use for the void cube?
0
A: Sandbox for Proposed Challenges

bsoelchNumbers you can negate by reading backwards Balanced tenary is a modified version of tenary (base 3), using the three digits 1,0 and -1 (written as T) to represent numbers. To negate a balanced tenary number you simple have to replace 1 by T and T by 1. Examples: 1 // decimal: 1 T // deci...

21:07
@Neil Solve it like a 3x3. I don't know the parity algorithm, so if I get it wrong, I make a D move and start over from the middle layer
21:25
That's how I used to solve 4x4s lol
@Bbrk24 ah, because there's two methods for 3x3, if you do the method of corners then you don't have a parity problem
I just do beginner's method for 3x3 :P
I have a Rubik's brand cube, so the algorithm was never the limiting factor. I can solve it twice as fast on someone else's cube
22:10
@JoKing I feel the same way, FWIW

« first day (4548 days earlier)      last day (592 days later) »