« first day (4117 days earlier)      last day (741 days later) » 

12:06 AM
Golfing in this language may actually be surprisingly fun though. I've discovered a few neat tricks already.
E.g., to increment every item in an array, you can horrifically abuse the parsing by doing array{+{1}} (which y'all have no context for how that works but once I publish the spec it'll be concerning)
 
Does it count as golfing if a code execution vulnerability in the interpreter or compiler is exploited?
 
Yes
We do that all the time at vyxal
 
those are the best golfs ;)
 
Do you have to use the latest version of the compiler/interpreter or any official release?
 
12:13 AM
Nope
 
hell no
 
@forest We actually have a bounty for answers that do that
5
A: List of bounties with no deadline

EasyasPi100-500 rep for solving challenges using an ACE exploit in an esoteric language I've reached 2300 rep, so here's a bounty for you RE/CTF enthusiasts for an answer I really want to see. CVE-2018-6849: Code execution results in code execution — TheZZAZZGlitch, "A friendly reminder: Lua scripts are...

 
@RadvylfPrograms Does it have to be a 0day or does an n-day suffice?
 
Anything counts
Example:
5
A: Create and write to a file

Aaroneous MillerVyxal 2.5.3, 32 bytes kh→@|;@+open(*"ww").write(VAR_)# Vyxal can't really do anything with external files, which would normally render this challenge impossible. However, in Vyxal 2.5.3 and prior, there was an ACE exploit allowing for arbitrary Python execution. Yep, I found another one. The AC...

That used 2.5.3 when the latest version was 2.6
 
Neato.
 
12:15 AM
we define languages by their interpreters, up to even allowing platform-dependent behavior on unstable nightly builds should you so choose
 
Let's see what happens when I throw some esolangs at a $50,000 fuzzing cluster. >:D
@UnrelatedString Does it have to bypass common exploit mitigations, or can I assume executable stacks and the like?
 
So long as you aren't breaking any standard loopholes you can do whatever the hell you want. codegolf.meta.stackexchange.com/questions/1061/…
 
It can rely on the computer/OS/... it's running on lacking various security things, although it might be less interesting in that case
 
@RadvylfPrograms It's just that it's easier to use an executable stack than to find ROP gadgets.
Likewise bypassing stack cookies is a pain.
 
spectre in vyxal
 
12:19 AM
> Custom compiler flags for building interpreters follow the same loophole rules.
Oh interesting... so you can use custom flags as long as the content of the flags adds to the byte count?
There are quite a lot of compiler flags that can really screw with security.
 
Flags are free now afaik
 
oh huh
 
We rely on "answers that abuse flags for free and unintenteresing golfs will get downvoted", but if the flags make it more interesting/actually answerable they're perfectly fine
 
They just define a separate language
 
12:22 AM
the only real restriction on doing Wack Shit is you can't both rely on and ignore a technical limitation like floating point imprecision or limited memory at the same time
 
@UnrelatedString What if you rely on lack of a technical limitation? E.g. an integer overflow that requires allocating a massive buffer? Or something that takes a very long time to succeed (even if it succeeds on the first try)?
 
that should be fine so long as the integer overflow can happen if you do in fact have enough memory for the buffer (or having insufficient memory for the buffer causes the overflow)
and yeah unlimited time is usually acceptable too so long as there isn't a nonzero chance of nontermination
 
@UnrelatedString Usually such bugs rely on the process allocating more memory for itself, not that it's necessarily fed all the data. E.g. an int overflow on len that occurs in exploitable code after void *ptr = malloc(len * 512) or something.
 
don't entirely get what you're talking about at this point so all i can say is virtually everything assumes sufficient memory for the desired behavior, so the only thing you can't do is assume unbounded memory at the same time as your program's behavior requires it to run out of memory
like counting on a stack overflow to terminate with an error or some such
 
But it's fine if it relies on the system not running out of memory, even if it requires, say, a thousand petabytes?
 
12:29 AM
yep
completely absurd solutions are a time-honored tradition around here
 
haha nice
Does the shellcode have to be x86 or can it be in any ISA (assuming a compiler for the esolang compiler/interpreter/runtime/whatever with a target for that ISA exists)?
 
anything you can run it on
 
Neat. Shellcode can be made so much smaller if it's an ISA with very high code density. :P
 
1:13 AM
In fact, we recently changed some rules that allow "anything you can run it on" to include soviet-era ternary computers too without a byte penalty :p
So if you can find an ACE exploit in Vyxal on a Сетунь it's fair game :p
 
@UnrelatedString *glances over at my O(2^{n!}) answer* nah :P
 
@RadvylfPrograms Tbf you don't need a Сетунь for that
even a potato battery computer would suffice
 
i hate how i can't right click the links because it brings up mathjax settings lmao
but i'm not sure i've ever written a brachylog answer that inefficient holy shit
 
I ctrl-click :P
> it times out on TIO for n≥3.
That's from my O((n^2)!) answer :P
 
Ooh if I get a 5 on all four AP tests I'm taking this year, they'll count for a full semester of college classes (at the college I'm hoping to go to)
 
1:17 AM
Imagine only being ablt to run an answer on n = 1 and n = 2 :P
 
okay looking at the problem i can see how
i should try to find my worst brachylog answer given it's practically a language purpose built for the most awful brute force possible
 
Yeah, all of the answers listed there are brute force approaches that rely on generating all possible "things", then filtering on "is it a special type of thing?"
 
unfortunately i didn't actually get complexity for most of my active career with it
 
1:21 AM
But still, the complexities are horrifyingly inefficient :P
They all have Œ!, or ŒP in them as well :P
> This doesn't finish on TIO for any valid input
 
Wait, I think I got the complexity of one of those answers wrong, it should be \$O(n^m(nm)!)\$ instead of \$O(nm(nm!))\$
 
Might be \$m^n\$, not \$n^m\$ actually
 
1:49 AM
dragonair desmos graph i made for math class :P (totally did not use regressions lol): desmos.com/calculator/tele8g1fmm
ok why is it so zoomed in
 
beautiful
 
the ears (or whatever theyre called) look a bit off but im too lazy to fix :P
@UnrelatedString thanks! spent a while on it even tho half the equations are literally just regressions
 
can't believe someone made a 3d maze game in desmos
what's next? Fully functioning sans fight using only graphing? /s
 
ugh, I just wasted two hours on a single game of Backpack Hero
 
2:16 AM
@lyxal well, someone has made desmos in desmos before, its fucking crazy what ppl can do on desmos :P
 
3:01 AM
for string i/o, can u always use list of character codepoints instead? is that some default i/o method?
 
cool thanks
 
Idea: I or someone else should make a gist that lists all the default I/O methods as like, short bullet points arranged by category, so it's easier to figure out if something's allowed
You know what I'm going to do that
Monday though, after I've taken the AP test :p
 
how many ap tests yall taking this year
 
3:08 AM
nice, im only doing 3
 
3:31 AM
@AidenChow 0
 
@lyxal oh... u in college already?
 
I'm in Australia
We don't have AP tests lol
Can't do AP tests if there aren't any to begin with
 
oh ye facepalm
what type of standardized testing u guys have at australia?
 
The HSC
But that's only for NSW
 
Not Safe Work™
 
3:34 AM
Other states just don't standardised test
@RadvylfPrograms programming can get quite dangerous with kids like you making horribly cursed esolangs
 
I prefer to think of it as a horribly cursed praclang, because the cursedness it subtle enough you could go a few hours without really finding any particularly egregious stuff
If it screams "this is cursed" from the second you see it, your brain can go into self-defense mode and block it out. It's when you're slowly steeped in it, that the horror can sink in, like a slightly sharper version of what JS does to you.
 
att
4:31 AM
wait are you all in high school
 
uni
 
@att one year before high school
 
att
wow im a boomer
 
I accidentally spelled abstract wrong and I find it funny
 
5:04 AM
@att uni
Generally, I've found that TNB, mostly, ranges from 13y/o to 22y/o
With the odd proper adult around to how young we are :P
@cairdcoinheringaahing "tell us" should be between "to" and "how", idk why my laptop is being weird :/
 
Good grammar very yes excellent done
 
5:21 AM
@att 10th grade.. (in a month)
 
6:03 AM
freshman in uni here as well but took a gap year
or should i say in uni as well, freshman but after a gap year
actually how does year get reckoned because the last day of finals was like today but there's also two summer terms so yeah still freshman
freshman but losing sleep on a 500 level machine learning class 🙃
 
6:46 AM
literally just did my final uni exam yesterday
i get to dunk on all you young'uns
 
7:10 AM
@UnrelatedString Is freshman first year?
@JoKing gosh that's an early exam period
Those don't start until June for me
(exams in general that is)
 
trimesters
 
Ah
 
semesters + wacky extra terms over the summer that exist if you want them
looked up another school in my state and their exams are actually a week earlier than ours
 
there's a summer term as well as the other three
 
oh wait you meant jo's exams
yeah makes sense
except i also had my last final yesterday
(of the term, not my degree lol)
 
7:15 AM
australia time is a little different
it kinda weirds me out that northern hemisphere school years don't match with actual years
 
7:45 AM
oh yeah wait y'all's actually do what the fuck
that never occurred to me
 
we usually start around late february, end mid-december
 
because the new year occurs over your summer
 
our financial year is still borked though
 
spent about the past 2 hours handdrawing my dragonair desmos graph, apparently i cant turn it in digitally, it has to be a physical, handdrawn copy -_-
 
7:54 AM
F
 
just hand in the equation, let them figure it out
 
"let them figure it out" i wish that could happen lolol
 
@JoKing Same lol
@JoKing IME for primary/high school, it's usually end of january/start of february
 
8:21 AM
Cursed idea: Language with a 32-char codepage (so 5/8 byte), but some of those are digraph characters that become single bytes when certain characters (3/8 byte) are appended.
 
8:53 AM
not a bad idea
 
9:19 AM
@emanresuA i don't think that really works in practice, it'll become amiguous if you mix between the two
unless every character has to be digraphed? in which case it's really a single byte language with a strange representation
 
 
1 hour later…
11:04 AM
YES
i got the scanfixedpoint and foldfixedpoint working in flax
 
@emanresuA VTD please?
 
^^ sierpinski triangle in 9 bytes
 
@pxeger I'm not EmanresuA but I'm happy to oblige
 
actually 8 bytes ;b¬=`ʲW₅
 
 
2 hours later…
1:21 PM
 
¯\_(ツ)_/¯
 
it has 3 stars
i think github's trending system is broken or something
 
1:54 PM
somehow my github follower count has finally managed to break 179.
I swear it was stuck there for a month
just 179
not 180, not 178, just 179
 
lol
 
0
Q: Nominations for new TNB Room Owners

caird coinheringaahingIn response to the increase in off-topic discussion in TNB, we've decided to appoint more ROs. Here's how we'll be deciding the new room owners: For one week (until 14:00UTC, on the 15th of May 2022), nominations will be accepted as answers on this meta post. You may nominate any user, including...

6
 
That was too quick @lyxal :P
 
I may or may not have been working on that for a few days now ...
 
wow lol
@lyxal when did you reach 20k rep?
 
2:04 PM
I totally haven't spent several hours preparing to FGITW my nomination if an election were to happen
@PyGamer0 last year
or this year
let me check
10th Feburary this year
 
Can we nominate NP? Pretty much active 24/7, and only occasionally posts off-topic things. Never engages in long off-topic conversations.
8
 
VTC as dupe of Radvylf's joke
Apr 30 at 2:44, by Radvylf Programs
Actually, y'know what, NP/SP for RO. That's my vote :p
 
@Adám true...
 
hate to say it but you've been ninjad by 8 days
 
Shows how active I've been… Travelling a lot lately.
 
2:08 PM
@Adám i finished travelling yesterday..
 
I just came home Friday afternoon, and I'm flying out again tomorrow.
 
Where've you been?
if you don't mind telling us
 
I was Wednesday–Friday at APL Germany's spring meeting in Berlin. Now going Monday–Thursday to Dyalog's all-hands planning meeting in Brockenhurst, Hants.
I had to take Berlin because the bosses were attending FinnAPL's spring meeting in southern Finland.
 
 
2 hours later…
4:28 PM
if I want to print a float to two decimal places I can do round(x, 2). If have a list of floats, how can I show them all to two decimal places using map()?
 
which language?
 
python
(sorry, should have said)
 
map(lambda x: round(x, 2), numbers)
 
oh yes!
thanks
 
@graffe assuming python, i think something like map(round, x, [2]*len(x)) would work too
if x is your list
 
4:32 PM
@des54321 oh wow.. I had no idea you could do that
in fact I don't fully understand it
 
im only like 80% certain that works though, lemme test it
 
map(round, x, [2,2]) say
 
you could also do map(lambda x: f"{x:.2f}", numbers)
 
@pxeger very nice
@des54321 it does work
 
@graffe from the python docs: " If additional iterable arguments are passed, function must take that many arguments and is applied to the items from all iterables in parallel."
 
4:35 PM
@des54321 but I didn't know map could take three arguments
wow.. thanks!
 
it can take any number of arguments >= 2
 
map takes a function, plus a number of iteratables equal to the number of arguments of that function
 
I learned something new. Thank you
 
4:50 PM
@graffe if this is for code golf though, a list comprehension like [round(i,2)for i in x] is actually i think the shortest answer by a byte or two
 
 
1 hour later…
6:03 PM
@des54321 cool
 
 
2 hours later…
7:49 PM
@pxeger Done
 
cool
 
just how long did lyxal spend writing their RO nomination?
 
too long
 
@emanresuA well, it was posted 65 seconds after nominations opened, so he's a quick typer :P
 
696 wpm, definitely
 
7:54 PM
6 hours ago, by lyxal
I may or may not have been working on that for a few days now ...
 
8:10 PM
I do love when buses cease to exist
2
 
 
1 hour later…
9:18 PM
I swear if there's lateral surface area of a parametric equation or some similar thing on the AP calculus BC test tomorrow I'm getting up halfway through, buying a plane ticket to Denver, and jumping out halfway through the flight
 
D.B. Cooper's origin story
 
And if someone even thinks the word "polar" I'm making it a round trip ticket and doing it twice
 
> Fun fact: this is my second time running for Room Owner of The Nineteenth Byte (my first run obviously didn't go so well).
What happened the first time?
 
It was almost a tie but I barely managed to steal it from lyxal
They were ahead for a while, IIRC
 
@forest lyxal came 5th, with 4 positions open
 
9:26 PM
I'll vote for whoever has the same taste in anime as me. :^)
 
Glad I'm not running (unless your taste is "I don't like it") :P
 
Heathen.
 
> You may nominate any user, including yourself
Oh, really‽
 
Wow an actual interrobang.
 
I was a sentence into writing a self-nomination when I remembered I'm already an RO
 
9:28 PM
he uses them sometimes
 
@Adám I thought it was implied that existing ROs and mods need not apply :P
 
What about mods who already have a superset of RO powers? :P
Maybe if I get RO my name will be so italicized that each letter will be laying down.
 
That said, if someone does want to post a "No new ROs" 'candidate', anyone is welcome to do so
 
(Renames Sprow to "No new ROs")
 
i remember when i first discovered interrobangs i was like "wow why don't people use these"
and then i realized they just look weird in terms of weight or whatever
but now i've gotten used to adám using them enough that i've looped back around to thinking they're cool
 
9:31 PM
@forest 𝘧𝘰𝘳𝘦𝘴𝘵
 
What unicode trickery is that?
 
Probably the mathematics letters
 
seems like italic codepoints that are then italicized
^
 
I keep forgetting that italic codepoints are even a thing.
 
it is kinda weird that they are
 
9:34 PM
Every once in a while I remember that there are unicode superscripts and I get all excited before re-discovering that there's only a few numbers and very few letters.
 
ᵃᵇᶜᵈᵉᶠᵍʰⁱʲᵏˡᵐⁿᵒᵖʳˢᵗᵘᵛʷˣʸᶻ render a bit inconsistently but the only letter missing is q
 
@forest Uh, all digits exist as both super- and subscript.
 
Plus, they're weirdly spaced/sized
 
That's a font issue.
 
@Adám Really? hm, maybe I'm just thinking about alphabetical characters?
 
9:36 PM
and i'm not sure if that's for lack of a superscript q in unicode or not because i just got that from the brachylog codepage bookmarklet :P
and yheah all digits are there
the codepoints for the digit superscripts aren't contiguous though
 
Honestly, I'm shocked that the alphabet is contiguous in unicode
 
That's for historical reasons; Latin-1.
@cairdcoinheringaahing "the alphabet"?
 
@cairdcoinheringaahing Well it kind of has to be since it's compatible with 7-bit ASCII.
 
@UnrelatedString 𐞥
 
For UTF-8 at least.
 
9:39 PM
@Adám Latin, sorry
 
@forest That's an encoding. Nothing to do with the Universal Character Set.
 
There's too many character encodings. I say we design a universal encoding that makes more sense, and covers all use cases to prevent these competing standards :P
 
oh
@cairdcoinheringaahing UTF-32! Turn everyone's life into hell!
 
@cairdcoinheringaahing Why do we need an encoding anyway? Just store the code points!
 
Let each person create their own personal mapping from code points to symbols :P
 
9:42 PM
just store all text as image files
 
@forest Why is UTF-32 even a thing? UTF-24 is plenty.
 
I guess someone thought there wasn't enough kanji to go around.
 
Fun fact: Many implementations that go by the name UTF-8 are actually WTF-8.
 
9:56 PM
 
Yup.
 
 
1 hour later…
11:26 PM
@emanresuA 4 days
Intermittently of course
 
11:44 PM
I was gonna propose UTF-12 as a compromise between UTF-8 and UTF-16 but turns out it already exists and is quite cool
It's also more efficient than UTF-16 for everything except a small number of codepoints for some West African languages
 
Efficient as in less overall bytes?
 
No, fewer overall bytes
 
In fact, excluding ASCII/Latin-1 and those 64 N'Ko characters, it's also more efficient than UTF-8
 
smh speak the President's English properly lyxal
 
11:50 PM
@RadvylfPrograms new golflang using utf12 when? :p
 
@RadvylfPrograms Like, for the average piece of text?
 
@user Joe Biden is not my president.
 
I expected UTF-8 to be better for most things because people usually just use ASCII
 
@user No, for literally every character in Unicode. It's a pretty good compromise.
 
Whoa
 
11:50 PM
like yoda speak you must
 
Speak like Yoda, you should
 
wait, how is that even possible
 
UTF-8 loses a ton of efficiency since proportionally more of each byte is for overhead
So outside of Latin-1, UTF-12 can easily beat it
 
it can't fit two ascii characters into one 12-bit block so how would it be more efficient if utf-8 only takes 8 bits per ascii character?
 
@hyper-neutrino Ever since learning about the Monty Hall thing, I don't question it when people tell me a mind-blowing thing involving math :P
 
11:52 PM
@hyper-neutrino All Unicode characters except Latin-1 and N'Ko
2 mins ago, by Radvylf Programs
In fact, excluding ASCII/Latin-1 and those 64 N'Ko characters, it's also more efficient than UTF-8
 
oh
so wouldn't it not be more efficient for the average (english) text then
 
Yeah. Languages that mostly use the latin alphabet would be about 50% less efficient, but UTF-16 was used all over the place until recently so I don't think that's much of an issue
UTF-16 is just really bad in general
 
ah
is utf-16 better than utf-8 for like, say, the average block of chinese text
 
Fun fact: Windows internally uses UTF-16 for everything.
 
11:54 PM
._.
 
IIRC places that primarily use CJK have been slow to adopt UTF-8 because it's way less efficient for that
UTF-12 would actually help there, though
 
@hyper-neutrino Not ⛔ if we all start 🟢 writing ✍️ like this 😄😂😄😂😆
 
ಠ_ಠ
 
Proposal: UTF-24-2, which uses blocks of 24 bits which can either encode a single 21-bit code point or two 12-bit code points (or possibly chunks of a single code-point)
 
Only problem with UTF-12 is that it doesn't fit into bytes evenly but it's really not that huge a deal
 
11:56 PM
Well ^^ addresses that :p
 
Yeah
You should make a new golflang using this :P
 
Unfortunately I don't have the time to implement 0x10FFFF operators :p
 
Just use copilot and tell it to generate all of those operators however it wants
 
Is it possible to write code solely using copilot? A golflang made by just rolling with what copilot does would be pretty fun to see
 
You'd have to give it prompts and some boilerplate here and there to keep in on track
But it could be done
 

« first day (4117 days earlier)      last day (741 days later) »