The Nineteenth Byte

there isn't too much more you can do with it other than maybe run-length encode things for hyper-optimization and that'll maybe save you a couple nanoseconds here and there

caird coinheringaahing

HN's interpreter with a -d flag

hyper-neutrino

nice

how did you get that message to send

caird coinheringaahing

magic

(it's a multiline message, shift-enter at the end)

hyper-neutrino

it even says the message is too long and won't let me update if i try to edit your message

@cairdcoinheringaahing oh ಠ_ಠ doesn't appear like that when i open the edit box on it tho

lol

caird coinheringaahing

9:49 PM

I'm more powerful than a mod :P

user

@dzaima Not a problem for me :P

caird coinheringaahing

@dzaima I use Python over C, because although C is faster to run, it's a lot slower for me to code :P

dzaima

@user I highly doubt it.

user

@dzaima It was mostly a joke. If I managed to get a C program to run, it'd probably faster than the Python one, but it's unlikely I can actually get it to work without segfaulting or something

hyper-neutrino

well, an instant segfault would make it finish running pretty quickly

AviFS

9:52 PM

@hyper-neutrino @cairdcoinheringaahing Hyper's interpreter showing the current element is a great addition. Here's the final Bash interpreter with that implemented!

dzaima

well, yeah, you'd have some hiccups while writing it, but you'd have to be doing absolutely crazy things to end up slower than python

AviFS

HERE

caird coinheringaahing

If this is a rickroll, I commend the dedication :P

@AviFS https://

AviFS

Hahaha, I swear it isn't

@cairdcoinheringaahing Beatcha!

Ah, shoot

dzaima

(if you want a fast BF interpreter, just transpile the characters to some C expressions and clang -O3 or whatever)

AviFS

9:53 PM

Wrong link, smh

hyper-neutrino

@dzaima yeah, that's exactly what the one on TIO does :p

idk about the clang -O3 tho

dzaima

@hyper-neutrino oh right, i thought i saw it somewhere. Didn't even read the code posted here ಠ_ಠ

AviFS

FIXED

In case that expires and we want to use it:

#define LIM 10
...
\#) echo 'for(int i=0; i<LIM; i++) {(p==i)? printf("[%d]", t[i]) : printf(" %d ", t[i]);} putchar(10);';;

hyper-neutrino

it's golfier and easier to understand if you inline the ternary into the format string like printf(p == i ? "[%d]" : " %d ", t[i]);, at least IMHO

dzaima

@hyper-neutrino it's gcc -O3, close enough (and there's precisely zero way a python program is beating that)

hyper-neutrino

9:56 PM

ah

AviFS

WOWOOWOW

I've never coded in C tbh

Good to know, thanks!

EliteDaMyth

@dzaima github.com/prplz/brainfuck-jit

converts brainfuck to machine code in memory

and executes it

hyper-neutrino

lol

AviFS

Ouchy...

hyper-neutrino

x=>x - compiles brainfuck code to brainfuck

dzaima

9:58 PM

@EliteDaMyth gcc -O3 will do that, but also apply a ton of regular optimizations (i've seen it optimize the fancy BF constant calculations to just mov rax, 85 or whatever)

AviFS

I imagine the C is faster though, since I'm sure it optimizes a lot before converting to machine code

EliteDaMyth

iirc, code.golf uses the one i linked

hyper-neutrino

you could probably manually do a crap ton of pre-processing and optimizing on the brainfuck code itself

caird coinheringaahing

@hyper-neutrino bf, 0 bytes. Takes input via concatenation to the source code, transpiles bf to bash, then C then bf :P

AviFS

10:12 PM

(Last change for now, this pads to make sure 3 digit numbers [0..255] don't ruin the spacing)

\#) echo 'for(int i=0; i<LIM; i++) {printf((p==i)? " %3d*": " %3d ", t[i]);} putchar(10);';;

user

11:06 PM

Searching for burn yields an awful lot of messages posted by me (and that search doesn't seem to include a couple more with the characters "burn" in them)

DLosc

@pxeger The more interesting thing in that document to me: they've finally caved and are adding (a more powerful version of) a case statement! :^0 I always understood that was something Python was never ever going to do.

user

iirc the reason it took so long was because they wanted to do it right

DLosc

[still reading] I'll be the judge of that! (jk)

user

Well, it does seem to lack custom extractors, which is another reason Python is bad and everyone should use Scala, which totally doesn't have any weird features

DLosc

I haven't heard of custom extractors. Got an example?

user

11:11 PM

Here's an article from Tour of Scala

DLosc

val customer2ID = CustomerID("Nico")
val CustomerID(name) = customer2ID
println(name)  // prints Nico

Okay, that's pretty cool.

Well, maybe Python will see fit to ~~stea~~ to adapt that feature in a future version.

user

@DLosc Here

AviFS

11:37 PM

I'm going to post it when I get back, but I just finished a BF answer to this question

The shortest Python and JS as of now are 24 & 22 bytes, respectively. With some answers in those langs going up into the 40s. Does anyone have a guess for how long the BF answer is?

caird coinheringaahing

200 ish?

There's going to be some way to isolate the second input byte, then there's various different hashes from that that'll be fairly long in bf

totallyhuman

second byte would just be ,,

caird coinheringaahing

..oh yeah, I'm dumb

AviFS

Haha, you've correctly guessed the first two bytes of my program!

@cairdcoinheringaahing 200 chars or 200 bytes?

caird coinheringaahing

Aren't they the same for bf?

AviFS

11:45 PM

(I'm dividing by 3 to get bytes)

caird coinheringaahing

Do you have an interpreter that accepts octal-encoded bf?

AviFS

Given the confusion, I'm guessing you meant chars

caird coinheringaahing

Yes

AviFS

I read about this in the meta, I thought there was, but now I got myself confused bc 8^3=/256

Wait, or are there 6

caird coinheringaahing

But if you don't have an interpreter which can accept the specific byte stream you're claiming as the length, then you can't claim that length

Bubbler

11:47 PM

FYI, the optimally (uniformly) compressed BF variant is this

AviFS

Yeah, nevermind that piece

caird coinheringaahing

@Bubbler Is that proven optimal, or just the current best?

AviFS

@Bubbler Interesting thanks

But sorry, back to chars. Curious what you guys would expect for a BF program

Remember you can't just calculate the number and print it

You have to print each digit

Bubbler

Each BF command can be encoded in 3 bits, which can make exactly 3/8 length of original BF code

AviFS

Oh yeah, that's how it works, thanks!

I knew it was something to do with 8 & 3, but it was 2^3==8, haha

Not 8^3==256, which it doesn't

caird coinheringaahing

11:50 PM

@AviFS You can output integers as characters by default

Bubbler

@AviFS You can print single number in BF, we allow it in the default rules

AviFS

.......................

I swear to ..........

caird coinheringaahing

Does that save ~50% of your chars? :P

AviFS

Darn, haha

I spent over a day on this, not proud of it

Bubbler

Of course the number you want to print should be within the bound of the BF cell

but there's always a BF interpreter that supports 32-bit cells and such

AviFS

11:51 PM

I searched 'output brainfuck' in meta in different ways and couldn't find anything

caird coinheringaahing

We're too general for that :P

Bubbler

Because it's not specific to brainfuck

AviFS

Oh...

Well now I know!

But I'm still pretty happy with it

caird coinheringaahing

yesterday, by caird coinheringaahing

@xash You may be thinking of this, which allows numbers to be input/output as characters, not the other way around

AviFS

The hash function is optimized for BF

Bubbler

11:52 PM

I recommend to dig through the entirety of Default I/O post at least once

AviFS

It calculates the digits of the number

Rather than the number itself

Bubbler

to get the hang of how we're generous in I/O formats

AviFS

@cairdcoinheringaahing So it doesn't really save a single bit, haha

But it's pretty cool; it's case insensitive

Bubbler

@AviFS That definitely makes your code interesting

AviFS

@Bubbler That's a really good idea, thanks

And the case-insensitive part too, I think

caird coinheringaahing

11:54 PM

@AviFS Yeah, that's a very interesting approach, I think that's unique of the answers

AviFS

It said we could take titlecase, lowercase or uppercase, but it can do anything

Phew!!

@cairdcoinheringaahing Thanks for thinking so, haha

caird coinheringaahing

I#m guessing it doesn't rely on a hashing algorithm then? :P

AviFS

It still does...

But two of them

Here it is

The first digit is always 1

And it does two iterative hash functions on the input

To get the second two digits

Bubbler

Cool

@cairdcoinheringaahing More well-compressed BF variant would probably be Sesos

Transcript for