The Nineteenth Byte

Nobody Needs Names

00:02

???

emanresu A

Looks like it's Flobnar @JoKing

Nobody Needs Names

@mods

Jo King

sounds good

"@Jok" actually did ping me lol

Nobody Needs Names

who's in charge of lyal?

aketon

wait whats happening

Nobody Needs Names

00:05

lyal

who's in charge of it

Is lyal now?

Jo King

i mean, i'll probably be teaching flobnar this time round, i dunno who usually announces it

emanresu A

You can :P

aketon

what

Nobody Needs Names

I just saw this code: ᵖʳⅰₙｔ(*ﬁˡｔₑʳ(ﬆʳ.ⅰˢªˡₙｕᵐ,ᵐªｐ(ᶜʰʳ,ʳªⁿｇₑ(ºʳｄ('{')))),ˢᵉᵖ='')

(python) lol

Jo King

Welcome to the twenty-first Learn You A Lang for Great Good! Today, we'll be learning Flobnar, a 2D lang described as what you'd get if Befunge got drunk. During the event, feel free to post CMCs to practice Flobnar (FMCs), ask questions about the language, and so on.

4

emanresu A

00:11

> Pip

Jo King

shhh

some useful links, the TIO interpreter and the documentation

Essentially it is a function style version of Befunge, with instructions having similar effects and with 2D navigation

emanresu A

This is so cursed

FMC: Factorial

Jo King

Each single character instruction either returns a value or calls other instructions in cardinal directions

Programs start at @ going left, and will print the final value that is returned (unless you use the -r flag)

emanresu A

Oh wait, there's a factorial example on the page :P

Jo King

well, try golfing it I guess

The directional commands ^v<> are all as you'd expect, simply calling whatever is in that direction. The # instruction will skip over the next instruction (note that blank spaces don't count as instructions, so this is slightly different to how Befunge's # works)

Dyadic instructions like arithmetic ones (+-*/%) will evaluate the first parameter from the north and the second from the south. On division by zero, the modulo and division operators will continue straight to get a value instead

 7
 -@
 5

will evaluate as 7-5 and output 2

There is a stack in Flobnar, but it's very different to Befunge. \ is used to push the value south of it to the stack, : can be used to retrieve the current top value, and $ to pop the top value before continuing. Note that this is only per evaluation, you can't push a value in one branch of the program and then use it in another.

:
$
+\<<\@
:7  9

This program outputs 16, where the bottom : evaluates to the 7 and the top one to the 9

@emanresuA I suppose i should also mention my answer at 19 bytes

Bubbler

00:38

@emanresuA 24 bytes without looking at joking's

Jo King

the -d flag enables decimal input with &

Bubbler

ooh nice

Jo King

If a function would fail, e.g input with no input left or the aforementioned division by zero, the evaluation will continue in some direction to get a default value

Nobody Needs Names

how do i split an input intio two?

emanresu A

FMC: Print 1-100

Bubbler

00:45

@NobodyNeedsNames push it in call stack with \ and use the "function argument" : to peek it

Nobody Needs Names

actually, i take that back

that's stupid

Bubbler

@emanresuA 17 bytes!

emanresu A

Woaj, you beat joking!

Jo King

very nice!

emanresu A

Why doesn't this work?

Jo King

00:50

@emanresuA they go up/down, not left/right

i thought the same thing, it's more intuitive

Nobody Needs Names

just went ahead and golfed it

+<
5*<
+<^
5 ^@

emanresu A

Oh, because it wraps. Neat

In fact...

Jo King

yeah, you can usually translate code up or down, left or right without changing the behaviour in order to reduce whitespace

@Bubbler i was golfing yours myself and came across 16 but outputs the negative unfortuntely

Nobody Needs Names

3oh no

OldSandboxPosts

Sandbox posts last active a week ago: Decompress a Sparse Matrix (WIP), Find the best character arrangement, Photosynthesis [king-of-the-hill] [python]

Nobody Needs Names

01:03

FMC, random number between 1 and 100 snicker

emanresu A

That's impossible

Jo King

flobnar has a random instruction (?), so not really

Nobody Needs Names

why

@JoKing almost as if you are questioning "flobnar has a random instruction"

emanresu A

@JoKing Oh

Jo King

uniform randomness would be very annoying though

Nobody Needs Names

01:10

IT

S

possible: tio.run/##S8vJT8pLLPr/…

and i did it: tio.run/…

Jo King

@NobodyNeedsNames i was thinking more along the lines of 16 bytes

assuming 1 to 100 inclusive

Nobody Needs Names

is that uniform

Jo King

no lol

Nobody Needs Names

i keep getting really small numbers

Jo King

1/3 chance you'll get a 1, 1/3 you'll get a 4, otherwise increment

Nobody Needs Names

01:23

increment what

Jo King

the number

Bubbler

Is @ functionally the same as < during execution?

Jo King

yup!

@NobodyNeedsNames on the next iteration it'll be 1/3 chance to get 2, 1/3 to get 5. etc

Nobody Needs Names

oh....

can it get 101

i think you can ;-(

@JoK There?

Jo King

01:39

it moduloes by 100 each iteration, so no

Nobody Needs Names

oh

but can it get 100

@JoK ?

Jo King

eventually, sure

Nobody Needs Names

but it mods 100

or is it 101?

Jo King

approximately ~1 in 3^96 odds

Nobody Needs Names

and will it get 0?

Jo King

01:49

it moduloes and then increments, so it is one indexed

Nobody Needs Names

oh

what's the next lyal?

Steffan

02:15

Jellyfish probably

Jo King

02:36

FMC: Cat program

Nobody Needs Names

@JoKing ~,@?

~,@

Jo King

this only prints one character

emanresu A

FMC: Hello, World!

A get-based approach is probably fastest

Nobody Needs Names

03:00

<+,@
~
v<

perhaps

@emanresuA ~~a bunch of ~ and + and numbers and a @~~

Nobody Needs Names

03:15

@jok

Jo King

doesn't seem to terminate? you know you can test these on TIO right?

it doesn't even reach the ~

Nobody Needs Names

bad formatting

but it doesn't work, why? tio.run/##S8vJT8pLLPr/30Zbx0GBS6FOgavM5v//kozMYgA

Jo King

it never terminates, so it never calls the print. additionally, this would print backwards lol

Bubbler

I'm not getting how this is different in semantics from my previous 17 (factorial)

Nobody Needs Names

what about this, why doesn't this work... does it only print after termination?

tio.run/##S8vJT8pLLPr/30bbQYGrzIZLx6bu//@SjMxiAA

Jo King

03:26

@Bubbler ah, the default for EOF is below the & for some reason

Bubbler

oh

Jo King

i wish flobnar was more directionally ambiguous

i might make a equivalent ><> analog with updated syntax

someone remind me next lang workshop

emanresu A

@JoKing Ooh

Gotta love compilerPotions

Nobody Needs Names

03:42

why doesn't this work

<+@
v<
,<~

and why does this output nothing

+@
,
1
+

but this does:

+@
>,1
+

And why does the example cat program

~,<
  +<@
  >^

does nothing?

emanresu A

@NobodyNeedsNames This loops forever because the two + are calling each other

(I think)

Nobody Needs Names

isn't it evaluating north and then south?

wait

good point, what if...

and it still doesn't work:

+@
1
,
+

Jo King

it outputs infinite sohs if that's what you want

Bubbler

damn, flobnar is way too vertical

Nobody Needs Names

oh

@Bubbler yes indeed

Why isn't this working?

+@
~
,
+

Bubbler

03:55

this is interesting

Jo King

FMQ: Is there a way to terminate by error?

emanresu A

@JoKing Yes

Jo King

yeah, that's the only way I could think of

emanresu A

It might be possiblw to recursion error

Bubbler

@JoKing 16b, terminating

wow, this is clever

Jo King

04:11

oh lol, i guess invalid instructions are also errors

emanresu A

Waiiit... Is Esolanging Fruit Chris Pressey?

Jo King

i'm pretty sure they aren't

stackoverflow.com/users/6996080/user2110704 is chris i believe

> Last seen more than 6 years ago

Nobody Needs Names

@JoKing yah

what does the flag do

tio.run/##S8vJT8pLLPr/305Bm6smxsGKy6DOikvBQOf//8Sk5JTU/7pFAA

@emanresuA Example program given:

Hello, World!
:
g<
0+<   6
\< 0 0v
 :,` |<\@6
>-^:#<6
 1    +<

Jo King

04:28

-r supresses the final returned value from being outputted

emanresu A

@DLosc idea: Acc!!!, which is like Acc!! except there are no variables

Only Count while <expr> {

Bubbler

fun 67b HW

emanresu A

Trying to figure out how / why that works

Jo King

the return result of , is zero

emanresu A

Oh

OHHHH, the execution sorta zigzags

Bubbler

04:41

yes, ignoring invalid commands using -i

Nobody Needs Names

what's -i again

is there even a documentation?

Bubbler

See here for flags

TIO uses Esolanging Fruit's impl

Nobody Needs Names

thnx

o/

Bubbler

05:58

0

A: "Hello, World!"

Flobnar, 34 bytes Hello, World!� :> v: g|\@>!, 00> + Try it online! (Previous record: 41) The last character at the end of the first line is the null character \0. @ Start the program : : \ >!,g Initially n (top of stack) = 0. > + 0 Print the char at (n, 0), push n+1, and evaluate ...

Hello, World!

06:14

1

A: "Hello, World!"

Flobnar, 34 bytes Hello, World!� :> v: g|\@>!, 00> + Try it online! (Previous record: 41) The last character at the end of the first line is the null character \0. @ Start the program : : \ >!,g Initially n (top of stack) = 0. > + 0 Print the char at (n, 0), push n+1, and evaluate ...

Nobody Needs Names

06:28

good

Bubbler

07:21

Somehow 1 to 10 is harder than factorial

New Posts

07:51

0

Q: Add in the blanks

In this challenge you will be given a list of positive integers which represents some range of integers which has been truncated for display. Your job is to find the missing bits and insert ellipses to show that that part has been truncated. The input will always be strictly ascending and if two...

Sandbox Posts

08:49

0

A: Sandbox for Proposed Challenges

Build a file/directory tree Task In this challenge, your task is to write a program or function which takes an array of paths with an additional boolean indicating it is a file or directory and outputs a file/directory tree in any reasonable format. Remarks Some directories can end with a forwar...

seshoumara

09:27

hi, long time

I'm making a programming game. Can I use some of the questions asked here, or parts of them, as tasks for my levels? I won't use the answers.

by here I mean the codegolf.SE

mousetail

@seshoumara stackoverflow.com/help/licensing

creativecommons.org/licenses/by-sa/4.0

seshoumara

thanks

hmm, seems complicated, back to making my own. Although programming-related SE sites have all the serious and non-serious tasks under the sun :))

mousetail

09:44

It's not that hard, just give credit and re-share your modifications

seshoumara

the sharing back is hard, I want to sell the game on Steam

giving credit is the easy part

Deadcode

10:46

@pxeger Please don't mess with my posts like that. There's a reason I chose that font, and I have many other posts that use it in the same way. And using backticks doesn't give that font; I don't understand why you would try to make me think backticks give the same font as Mathjax.

I didn't use the regular header font for that because it could be confused with a language name.

There's plenty of compromises we have to do to make text look the way we want on StackExchange sites. For example, we also have to use <sub><sup>...</sup></sub> to get a small font. It sucks, but that's the way it is. So for this, using the Mathjax font to get a distinct-looking header is just what I'm stuck with.

pxeger

11:02

@Deadcode I meant you should write \$\text{Modification}\$ instead of \$Modification\$ if you specifically wanted a different font. Otherwise the kerning is all wrong

emanresu A

I have never been as self-conscious about my handwriting as when I was inking in the caption for this comic.

Someone hated you

pxeger

@pxeger I used the backticks to contain the formatting I was suggesting; they weren't supposed to be part of it

Deadcode

@pxeger That gives a different font, that basically looks the same as a standard header.

And why did you edit my post again. I asked you kindly to please stop messing with my posts like that.

Correcting typos is fine, but messing with my design decisions is not.

pxeger

The effect of the kerning is not as bad on "Variations" as it was on "Modifications" at least

But I don't see any need to get in such a tizz about it. This is a collective site, and I'm aiming to clean it up where it helps. If you insist on having it one way or the other, you can just rollback.

Deadcode

@pxeger It feels like a personal attack to have that kind of edit done on one of my posts. I'm a perfectionist and I carefully consider every design decision. To have one casually changed like that without my permission feels like an insult.

pxeger

11:12

Well I can't help if you feel that way about an SE feature which I'm using for its intended purpose

Deadcode

I would disagree that that's its intended purpose.

You wouldn't edit someone's post to include a new golf, would you?

Nobody does that. They suggest them as comments.

emanresu A

pxeger, stop caring.

Deadcode

Messing with the posters' design decisions is not within the scope of what edits to other people's posts are intended for, I would say.

pxeger

No, because we have a policy against that

Fixing formatting is definitely an accepted use for edits

Deadcode

It's one thing if a poster casually used the wrong formatting. But if they intentionally decided to do that, it's a whole other matter.

pxeger

11:14

If you don't consider it "fixing" because that's what you want, then fine

How was I supposed to know you intentionally decided to do that? I just assumed you didn't know about \text

Deadcode

But you did the same edit again even after I asked you not to. That's not okay.

pxeger

It was a completely different edit

Deadcode

It looked the same.

pxeger

Not remotely, for me

Deadcode

You changed an italic serif font to a non-italic sans serif font.

pxeger

11:16

@emanresuA I think I'm not the one who cares. I'm defending myself against an accusation of being rude, when I had no intention of doing anything other than helping tidy some formatting

Deadcode

And don't stop caring. Caring is good.

pxeger

1 and 4 are mine. They look completely different to me

Deadcode

1 and 4 are both non-italic fonts. The whole point was that I was going for a font that looked very different, and italic serif accomplished that.

Now I'll admit that 1 is serif and 4 is sans serif. But 1 is not the look I was going for.

And furthermore you did it without even asking me, even when you knew I was here in the channel ready to chat.

You could have just told me here what your suggestion was.

And let me decide if it fit my specifications, and whether I wanted to implement it.

pxeger

@Deadcode You can add the italics without going back to the bad kerning. But I think 1 and 4 look plenty different (not that I even think they need to be); the italic difference is much less than the whole font appearance, and serifs. But that's besides the point.

Deadcode

@pxeger And "you can just rollback" is not a valid argument. I am not notified when someone does an edit to my post unless I am actually looking at that post, or I happen to look at the index and see that my post got bumped and I wasn't the one who did the editing.

It's easy to miss someone else doing an edit to one of your posts.

pxeger

11:21

@Deadcode I thought it unnecessary to ask you, because 1. my edit effectively already was me asking: if you didn't like it, you can just roll it back. 2. my second change brought it so close to your original while still fixing the kerning that I thought you couldn't possibly object

@Deadcode I'm pretty sure that you do get a notification when someone else edits your post

Deadcode

@pxeger But I had just asked you to please not do that kind of thing to my posts, and then you went and did the same thing immediately.

mousetail

You do get a notification

pxeger

To be honest, I completely ignored "don't mess with my posts like that" because SE encourages me to edit posts when it improves them. It's not "messing with".

Deadcode

## \$Variations\$
## \$\textit{Variations}\$

so your suggestion would have me do the second one instead of the first. But the second one is indented ever so slightly, and it's ever so slightly smaller. I don't like it as much...

If I do decide to do this, BTW, there will be at least half a dozen posts I have to edit. All will get bumped

pxeger

If you seriously care that much about the amount of spacing around the header, you can change it to what you like. I just think the readability gain from fixing kerning on the letters was more important

AncientSwordRage

11:28

Thanks @pxeger

pxeger

AncientSwordRage

That table looks better for me

pxeger

@starrers: I understand if you want to show which side you support in an argument. But I think starring those messages is not a good idea, for several reasons:

4

AncientSwordRage

and I keep forgetting the preview of tables is messed up

pxeger

- it may encourage people to star opposing messages, and then the starboard would fill messages all about the same one topic

AncientSwordRage

11:31

I have a strong urge to star these reasons but I will resist

pxeger

- it may appear to turn a useful debate into a popularity contest

- it brings attention to a argument that both original sides have more-or-less finished, which is likely to cause unnecessary annoyance

Deadcode

I see what you mean about the kerning. But I just have to decide if this is acceptable, and then go and ahead the at least half a dozen posts if so.

AncientSwordRage

Changing the topic then: Is there a challenge for randomly picking from a list, and if so would it make a good challenge?

Deadcode

@pxeger That's another reason it was not acceptable for you to do the edit, BTW. I have done the same thing on many posts, and if just one were edited, that would make my posts inconsistent.

pxeger

@AncientSwordRage almost certainly yes to both

AncientSwordRage

11:35

@Deadcode I would prefer half a set of posts to have 'good' kerning than none of them

@pxeger I couldn't find the former

Deadcode

And I agree with pxeger about starring the chat messages. Whoever did that, it was in extremely poor taste.

AncientSwordRage

@Deadcode Bear in mind, that different chat rooms have different tastes. Someone who's rarely here won't know the customs and norms perhaps

the reasoning holds

Deadcode

@AncientSwordRage What reasoning?

pxeger

It's not a punishable offence, and we can just clear the stars. Don't get too worked up about it

AncientSwordRage

it's just that it's not always an issue in some places

7 mins ago, by pxeger

@starrers: I understand if you want to show which side you support in an argument. But I think starring those messages is not a good idea, for several reasons:

@pxeger yup, I cleared some myself

@pxeger if you do come across a 'random pick from list' then let me know

pxeger

11:39

@AncientSwordRage If we don't already have one, feel free to write up a sandbox post. Just keep in mind these guidelines about randomness

AncientSwordRage

@pxeger I knew there would be some sort of sticking point about randomness, thanks

Deadcode

@pxeger On a different note, could you please install python-chess on ATO? It's pretty small; shouldn't be a problem. I'd also appreciate having Boost (C++) installed, though that's much larger, so I'd understand if you don't want to.

pxeger

11:50

@Deadcode Done python-chess

Boost is 180MiB though, so I probably won't be able to add that in the near future

Deadcode

@pxeger Thanks!

emanresu A

@pxeger Imagine having hard disk space restrictions. I cheat

mousetail

If you just deployed why is the fix for syntax highlighting not live yet

pxeger

It was only a partial deploy, just to update the Python image

mousetail

Ah ok, how often do you do a full deploy?

emanresu A

11:53

@pxeger Yet it's not on staging...

pxeger

@mousetail Every few weeks, but it depends on how much has changed, and how much of it I've already deployed in bits and pieces

@emanresuA staging is pretty broken at the moment

emanresu A

WHy?

pxeger

@emanresuA because I'm testing some things

user

@emanresuA do you just query ato for dso?

pxeger

I'm working on cancellation

emanresu A

11:54

@user ~~good idea~~

Perhaps you need a staging.staging.ato.pxeger.com

pxeger

@mousetail done that now

user

It’s okay for staging to be broken

mousetail

Need more levels of staging

user

Would be nice to have a staging for every branch tho

pxeger

really, I shouldn't be training my users to rely on staging

if I deployed to prod more often, that wouldn't be such a problem

mousetail

11:56

Another feature request would be the ability to resize the answer box. It's not really big enough for verbose languages like rust.

emanresu A

@pxeger Have you got incremental deployment?

Jo King

@pxeger you would just have more, exciting problems

pxeger

@user if someone wants to sponsor me the £££s worth of disk space, I'll gladly implement that ;)

mousetail

Abuse discord for infinite storage

pxeger

@emanresuA sort of, but it's non-zero downtime, and only partly automated

mousetail

11:57

files saved there stay forever

pxeger

I'm already kinda abusing Docker Hub for infinite storage

user

Ooh how?

@pxeger Dyalog perhaps?

emanresu A

If I end up compiling any wasm binaries myself, I'd probably externally host them

pxeger

@user by putting an entire new Arch Linux installation there once a week

Sandbox Posts

0

A: Sandbox for Proposed Challenges

Generate a Tiefling's Traits A Tiefling is a character race from Dungeons and Dragons, that have a list of possible traits: small horns fangs or sharp teeth a forked tongue catlike eyes six fingers on each hand goat-like legs cloven hoofs a forked tail leathery or scaly skin red or dark blue ski...

pxeger

11:59

I suppose Docker can't complain that much about me though, given that Arch Linux themselves are putting an entire new Arch Linux installation there several times a day

user

lol that’s an interesting way to do it

@pxeger ato is probably a drop in the ocean for them anyway

AncientSwordRage

Also while I'm here, do my latest edits here make sense and/or work for people?

mousetail

Storage is extremly cheap in bulk, it's just us pebs that have to pay for individual petabytes

pxeger

The problem is that ATO's storage needs to be accessible speedily

I can't just put it all on S3 glacier

emanresu A

Checked, and DSO has about 70KB of code

Deadcode

12:03

@pxeger In any case, thank you for pointing out the kerning issue, even if the way you did wasn't to my liking. As it turns out, I found a way to get basically the exact same font size but with proper formatting:
old version: ## \$Text\ goes\ here\$
new version, fixed: ### \$\large\textit{Text goes here}\$
The new version is still ever so slightly indented more, but I can accept that.

emanresu A

+ 330KB of DVSM

o/

pxeger

@user At some point I'll get around to setting up GitHub Sponsors, and then hopefully all my worldly problems will be solved

mousetail

How is ATO financed?

pxeger

My own pockets

mousetail

Ouch

pxeger

12:04

It costs about £8/month in total

mousetail

Oh that's suprisingly cheap

Ever get issues with people abusing it for crypto mining or similar?

pxeger

I think that would be pretty hard to do

It's max 60 seconds execution time per request, and max 20 requests per minute, per IP address

Deadcode

@pxeger That's amazing! All the same, thank you very much for doing it.

PyGamer0

^

mathcat

@mousetail isn't that against the TOS?

pxeger

12:11

well yeah that too

mousetail

Yes, that's why I asked if he had problems. If it was allowed it wouldn't really be a problem but just a design decision

AncientSwordRage

@pxeger thanks for funding it!

graffe

12:43

In python, is it faster to compute the union of sets of integers than sets of shortish strings?

pxeger

How big are the integers?

graffe

Up to 2**32

pxeger

I would guess integers are faster then

shouldn't be too hard to benchmark though

graffe

A benchmark is a great idea

mousetail

Python has the builtin timeit module for benchmarking

pxeger

12:47

The difference is pretty insignficant

ato.pxeger.com/…

int: [0.6171469429973513, 0.6378249230328947, 0.6157839560182765, 0.6052223299629986, 0.6150264090392739]
str: [0.6953656200785190, 0.6999090739991516, 0.6987536610104144, 0.6986689930781722, 0.6942863309523091]

mousetail

Strings seems consistently slighly slower, order of 2-5%

thejonymyster

i'd call that significant

graffe

Hmm

That's a shame

thejonymyster

~~they would be the same speed if you used string ints~~

mousetail

Strings would probably still be slighly slower because of the system that prevents collisions from being predictable

thejonymyster

12:52

is that standard jargon?

mousetail

no

at least IDK

thejonymyster

fair lol :P i just dont know what you mean

what's colliding? and why dont we want to be able to predict it?

mousetail

Do you know what a hash collision is?

thejonymyster

nope, but now i know what to google :)

mousetail

Basically when 2 values hash to the same item

There was a issue in python that since there where strings known to collide you could pass these as query parameters to servers running python

This would massively slow them down and make them unresponsive

thejonymyster

12:55

oh wow

mousetail

So they changed it so the exact hashes for strings are unpredictable so you can't pre-compute hash collision payloads to perform denial of service attacks

hash seed is randomized on startup

thejonymyster

and so this (possibly?) slows down string comparison because of random?

mousetail

The algorithm is more complex

For integers the algorithm is literally number%hashsize

there is no real way to make this unpredictable

If you XOR the number you would still get the same collisions

Strings use a much more complex algorithm that allows the seed to effect different strings differently and actually break collisions

graffe

But a 5% slowdown seems surprisingly small

Transcript for