The Nineteenth Byte

12.001k rep. It must be a floating point error.

@miles Now I see what you meant with that test case. Nice catch

miles

@LuisMendo sure, no problem

@Quill there's also git-fire

There has to be a shorter way to do 0,0j...

flawr

12:02 AM

@LuisMendo hehe, I think it is just so cool what you can do with convolution! I remember before I started my degree I read about convolution on WP and it looked soo damn cool but I had no idea what it was all about, and now during my degree I saw applications pop up everywhere. And at some point I noticed that it is the same as polynomial multiplication, and deconvolution is nothing but polynomial long division (in the discrete case). This blew my mind=)

@flawr In my area (telecommunications / signal processing) it pops up all the time. Any linear, time-invariant operation is a convolution (except for pathological cases in continuous time). You learn to think in terms of convolution and yes, it's nice to apply it when it's not initially called for :-)

Adnan

@Dennis Wx4?

Nevermind, that doesn't work

why not?

forking?

Adnan

It only works when there is no argument

It defaults to 0

Maybe something like ⁴BḊ?

A: Convert xxd output to shellcode

Still three bytes.

I need to prepend and append a zero to the list.

Oh dear, this is a mess.

0

Jelly, 51 bytes ɠ,ɠi@€€Øh’ḅ¥®L’¤Ạ?€⁴ḤĖS€r/ ṣ⁷ṫ€⁵ḣ€40Fḟ©⁶¢ịs2“\x”;Ð€ Expects the hexdump as command-line, and start/end on STDIN, separated by a linefeed. Try it online!

12:15 AM

Longest Jelly answer ever?

No, but the longer ones had compressed strings, so they don't really count.

Why is 1 - bool_to_i(x) valid ruby, but not 1 - bool_to_i x? is there some possible ambiguity I don't see?

I'm bummed about the second and third arguments on that challenge since otherwise I'd have an extremely competitive answer

v.tryitonline.net/…

Yeah. At least for Jelly, that's definitely the hardest part.

Hexadecmal to boot.

I'm guessing handling the second and third args are like 3/4 bytes tops?

12:23 AM

Without those, I'd have a 21 byte solution.

Really? That doesn't seem right.

It's string manipulation. Jelly's bad at that.

Taking a substring given indices is 30 bytes?

The indices are in hexadecimal (no built-in) and there's the special case -1 as well.

Don't you effectively want string[arg1*4:max ((arg1+arg2)*4),-1]

12:26 AM

Plus, Jelly uses 1-based indexing.

@DJMcMayhem Something like that. arg2 is end though, not length.

Oh, OK. Why is it so complicated then?

ɠ,ɠi@€€Øh’ḅ⁴ (convert from hexadecimal to integer) is already 11 bytes.

And that doesn't work for -1.

I predict that 05AB1E will end up winning

probably.

12:44 AM

@Dennis When's the code page change coming

@Quill holy smokes

A: Convert xxd output to shellcode

1

05AB1E, 39 bytes Input in the form: arg2 arg3 arg1 Code: ²\|vy9F¦}40£ðK}J2ô"\x"ì²H>²®Qi²}£¹HF¦}J Uses the CP-1252 encoding. Try it online!.

Greetings again

I've made a few changes to the leaderboard

I hope it works now

At 100%

jsfiddle.net/akc5f15a/5

I've changed it to only use 1 API call for everything and the answers are only classified heuristically invalid based on a score of the comments

The OP has an higher power and changes the score by double

why not use an input box instead of an option box for the amount to show?

It's automatic

oh, cool

BTW, what does green highlight mean?

12:48 AM

I can try to fiddle with it to allow a "show all" option

That's the accepted answer

ah

I though it was a good idea to mark it

And I found a bug

which is?

The score wasn't really being taking part on the condition to show when an answer was considered invalid

jsfiddle.net/akc5f15a/6

Should be fixed now

I got it down to 44. I should stop now.

12:51 AM

Now, I need a question that has comments saying that stuff doesn't work

@quartata It probably won't. I'm considering moving to a assembly/binary approach.

@IsmaelMiguel look at hello, world

v
\ {"+":[01],"-":[01],"*":[02],"/":[02],"^":[13]} "prec" =
\ {"+":[+],"-":[-],"*":[*],"/":[%],"^":[J]} "func" =
\ ["prec"`~F~$~$]kKp ["prec"`~F$~$]kKa [dnNEQ]kKn [2@dHNE~$]kKh
\ 0b "queue" = 0b "stack" =
            v                  =q@"queue"&~ <j*44             <
                      >>:c=Knv
> "> "oiSBq > l0E!v_Kn|^+~c. _ .cn "queue"` ^
                      >d" "E?v$>"stack"`V?v$"stack"`~&"stack"=^
            ^              $$<            > :2= :1= .1Kp .2Kp   .1Ka [Y][y]2b~H~$  g  ?v$"stack"`.1&"stack"=

I've finally got shunting yard parsing working

λ ruby reticular.rb examples\shunt.ret
> 3 + 4 * 5
[3, 4, 5, "*", "+"]

for the interested

@ConorO'Brien Lets see

@ConorO'Brien Not good enough. The only thing it found was a false positive :/

Which is something

odd

Also

jsfiddle.net/akc5f15a/7

I've updated since I left some console.log() around

It took 40 seconds on a StackSnippet

Should take 2 seconds now

12:58 AM

your script says R - 39 is an invalid header :/

I copied that piece from the original leaderboard

Also, R - 39 is invalid

Should be R, 39 bytes

(The last word seems very important)

In fact, that is the only piece I've copied from the other leaderboard

is bytes optional?

I don't think so

I feel like you should include entries like that, but flag them as potentially incorrect.

I can't include them since the regular expression fails

How else would I classify them?

I mean, they are included, but WAY down to the bottom

1:03 AM

not included in the ordering, however.

They are

But always at the bottom

incorrectly included.

Why "incorrectly"?

because they are <X> bytes, not <INFINITY> bytes

If this were me doing it, I'd remove all tags & their contents, and select the last number that appears in the string.

I tried it

I went insane after 5 hours of it not working

Also

It won't work for R 5 bytes

Or similar forms, like R (5 bytes)

I do, however, agree 200% with you

I just can't get it to work

1:06 AM

lemme try.

@Dennis Dada gave a good reason to add [50 0] as test case, which I just did

Wait 1 minute

I've almost got it ;)

I found a nasty bug

jsfiddle.net/akc5f15a/8

@IsmaelMiguel which string does the regex perform on? The HTML inside the header?

1:09 AM

There, "fixed" it

@ConorO'Brien Yes

Actually

The HTML of the body of the answer

@MartinEnder did it, he should know

so, say I was to write a function called extractNumber that gets the number. it would be called with the string representation of the answer's HTML?

Yes

But I can try to extract the first node

From there, should be easier

I can do that. Could you perhaps give me an example of an answer HTML? too lazy to find one myself

@LuisMendo Already added.

Go to the version 2 or 3 of the jsfiddle

There, you got plenty of data

But I remind you: that data is incompatible with the new version

But you can get a few headers from there

1:12 AM

:/

You have no idea of the headers some people use

Some scary headers show around

yes, I do.

I tried doing this before :p

I salute your bravery

Crazy idea:

Trying to detect the language based on the detected language by the syntax highlighter

Nah, doesn't work

function extractNumber(string){
    var dummy = document.createElement("div");
    dummy.innerHTML = string;
    var header = dummy.children[0];
    for(var i = 0; i < header.children.length; i++){
        var child = header.children[i];
        header.removeChild(child);
    }
    return header.innerHTML.replace(/.+(\d+)\D+$/, "$1");
}

this should work @IsmaelMiguel

haven't tested rigorously

works on lots of inputs

I'm fixing a rather annoying regression

All invalid answers are "stuck"

They don't sort

1:25 AM

i could try, but no laptop rn :(

invalid answers?

Answers with bad headers

example?

x86/x86_64 on Linux 34 bytes (codegolf.stackexchange.com/a/89117/14732)

how is that a bad header

You can't exactly know the language

Have you ever heard about the language x86/x86_64?

Is on Linux part of the language name?

1:31 AM

oh

just take the on linux as part of the name

^

jsfiddle.net/akc5f15a/11

Here

Fixed

Now, time to try your function

After I get a nice cold beer

ΛεγίωνΜάμμαλϠΨΠʹ

What's the name of the theorem that states that the first 30% of any video is unimportant?

@ΛεγίωνΜάμμαλϠΨΠʹ the incorrect theorem

ΛεγίωνΜάμμαλϠΨΠʹ

@ConorO'Brien ?

1:38 AM

:P

bad joke

@ΛεγίωνΜάμμαλϠΨΠʹ When you find out, make a video and mention it in the first 30%.

@ConorO'Brien The regular expression at the end should be /.*?(\d+)\D+$/ and not /.+(\d+)\D+$/

@IsmaelMiguel why?

For an input like <h1>aaaaa efef - dsd 45 bytes (non-competing)</h1>, yours returns 5 and mine returns 45

Also

.+ is greedy and will match as many as it can

While .*? will match as few as possible

Nathan Merrill

My compiler works!

1:52 AM

Hai.

Greetings

@IsmaelMiguel ah, true.

2:07 AM

Urgh, this is really not working

Your function works, don't worry

It does what it should

I'm sorry, but I will leave the code as-is and bad headers should be ignored or fixed

This is too much work for irrelevant data

2:23 AM

kk

I'm sorry, but the effort is not worth the results

It's a lot easier for me to fix the answers than to change my code

Also, there's a warning saying: "Your answers must have a header with the following format"

true

They don't respect it because they didn't read it/legacy answers

Your function is really cool, I just got fed up with the failure

it's okay, I feel ya :)

I will make a git and you can clone/make a pull request/something with it

2:25 AM

oh, cool!

github.com/ismael-miguel/ppcg-leaderboard

I hope I didn't pick the wrong license

mınxomaτ

MIT is never wrong

You sure?

mınxomaτ

@IsmaelMiguel Yes. Though for such a short snippet, I'd personally go full public domain.

^

2:38 AM

Why's that?

@EᴀsᴛᴇʀʟʏIʀᴋ brew update and that should be fixed.

mınxomaτ

@AlexA. :O

You're alive.

Sort of

What happened?

I died

2:42 AM

Elaborate.

I stopped being alive.

Doorknob

ALEX HYPE

8

lol yo, the hype is real

eeeee

Whoa, I haven't seen you in an entire millennium

2:45 AM

I have been very busy

Although I did see you furiously star Julia things on Github so I was vaguely aware that you hadn't been eaten by something

\o/ my calculator is almost done! there's just one bug

Haha. One of the ways I've been busy is in Julia development.

> 3 ^ 3
27
> 5 / 3
1.6666666666666667
> 3 + 4
7
> 2342
2342
> 234 * 234234
54810756
> 3 - 1
(hangs)

Who needs subtraction anyway

2:49 AM

Pretty impressive considering it's written in ><>

what

Or Reng or something. One of those 2D languages

What's one more D than I'm used to

He was showing some incomprehensible snippets of the parser a while back

Once you've seen Adnan write 05AB1E's compressor in 05AB1E you've seen it all

O_O

2:53 AM

@quartata reticular.

v [part I.a: constructing the objects]
\ {"+":[01],"-":[01],"*":[02],"/":[02],"^":[13]} "prec" =
\ {"+":[+],"-":[-],"*":[*],"/":[%],"^":[J]} "func" =
v [part I.b: defining helper functions]
\ ["prec"`~F~$~$]kKp ["prec"`~F$~$]kKa [dnNEQ]kKn [2@dHNE~$]kKh ["func"`~H~$]kKf [dT"String"E]kKs
v   < [part II: repl stage]
\ 0b "queue" = 0b "stack" =
            v                  =q@"queue"&~ <j*44             <
                      >>:c=Knv
> "> "oiSBq > l0E!v_Kn|^+~c. _ .cn "queue"` ^
                      >d" "E?v$>"stack"`V?v$"stack"`~&"stack"=^

here is the code.

Chat mini-challenge: paste an ASCII table into the chat (containing just the characters, no character codes, names, or anything).

@AlexA. Anyways you missed this room reach a new record of one instance of a mod moving messages to trash a day

@feersum Umm... what about unprintables?

@feersum can we have unprintables being printed, since SE removes them?

@Geobits It's OK if you skip those.

2:54 AM

@feersum You can use asciiflow.com to make those things

You can start at 32.

I used it before

I of course blame this depression on the absence of your soothing presence, feel free to feel guilty about it

6

moi aussi

._.

2:56 AM

So like this?:

for(char i=0;i<256;)System.out.print(i++);

Like that except with more rows.

VTC: Not specific enough

VTC: Not enough rows.

I was hoping someone would use the J builtin.

@feersum jolf, 6 bytes: ₯I₯o~ (trailing space). assumes jolf is loaded correctly on the page.

2:59 AM

Moar rows: for(char i=0;i<256;)System.out.print(i+++(i%8<1?"\n":""));

anyway, good night now

Good night. See you again in a few minutes.

2

yeah, probably >_<

Oh, I almost forgot. We managed to cajole Conor into making his name ASCII only with the overwhelming power of stars. 30 of them to be precise.

I think that's pretty much it.

\o/ \o/

This is truly cause for celebration!

3:05 AM

Next target is the Irk of the East.

Let small caps go back to the hell from whence they came

3

Yeah, unfortunately it hasn't gotten anyone else to change theirs. And we have a lot of Unicoders now it seems

:/

hai

@ConorO'Brien wat

i missed you

@Geobits NO.

Unicoders are the 𝔴𝔬𝔯𝔰𝔱.

3:07 AM

@AlexA. fuck

speak of the devil

@AlexA. nope

⒪⒣ ⒢⒪⒟ ⒲⒣⒜⒯ ⒣⒜⒱⒠ ⒤ ⒟⒪⒩⒠??

amazingness

𝕀 𝕓𝕖𝕘 𝕥𝕠 𝕕𝕚𝕗𝕗𝕖𝕣

3:09 AM

i don't

OH and Martin reached 100K. I think that's everything for real this time

sᴍᴀʟʟ ᴄᴀᴘs ᴀʀᴇ ʟɪғᴇ

@EᴀsᴛᴇʀʟʏIʀᴋ How recently have you checked? I guess that stuff was fixed about 8 days ago.

@AlexA. 1min ago

He should be receiving his swag soon

3:10 AM

@quartata That I knew. That would certainly be hard to miss.

rebooting computer

@quartata you actually get swag for 100k?

@EᴀsᴛᴇʀʟʏIʀᴋ ɃɇȼȺᵾsɇ ɏøᵾ ħȺvɇ ƀɇɇn føøłɇđ ƀɏ ŧħɇ ɇvɨł ønɇs

Have you done brew upgrade?

@EᴀsᴛᴇʀʟʏIʀᴋ Yes

I should be getting Puzzling swag soon :D

cool

@AlexA. yes, does it every night automatically

3:10 AM

Also, Tio! is begging for money now.

@Dennis So it comes to this... from the one who complained about how Vim's startup screen mentions a charity for children in Uganda. ;P

@AlexA. thanks, brew update + rebooting fixed it

@EᴀsᴛᴇʀʟʏIʀᴋ Closed as duplicate of something you said 2 minutes ago

@Geobits i doubt that

@EᴀsᴛᴇʀʟʏIʀᴋ Okay great

3:12 AM

@AlexA. nuked by owner

@EᴀsᴛᴇʀʟʏIʀᴋ Well... yeah, you would. That's part of the whole "being fooled" thing.

Unrelated, I just got a replacement battery for my 2.5yo phone, and the difference is ridiculous. Stupid batteries degrading over time.

@AlexA. That still seems a tad unrelated, but whatever. 512 MiB of RAM only gets you this far, and Tio! is becoming increasingly popular.

@Dennis What seems unrelated? The Uganda thing?

It's part of Vim's license terms IIRC

Yeah, not sure how it's related to a text editor.

Rather than giving them money, you give money to the charity

3:19 AM

halp does SE markdown keep HTML comments

@Geobits So true

@ASCII-only Yes

I think a TIO church really could work. "Tired of your normal prayers going unanswered? Try it Online! Looking for a group of like-minded individuals you can confide in? Try it Online!"

Hmm

brb, registering trademark

xD

@AlexA. so there will be the comment in the source of the html right?

@Geobits Oh, so it's you they were talking about...

3:20 AM

@ASCII-only I think so

5 hours ago, by Dennis

> Dear President & CEO,

This email is from China Intellectual Property Office, which mainly deal with trademark and domain name registration internationally. We received an application from "Ecolume Investment Ltd" on August 8, 2016. They want to register "tryitonline" as their Network Brand in China,Hong Kong,Taiwan and some Domain(.asia/.cn/.com.cn/.hk/.com.hk/.tw/.com.tw/.in/ETC).

Now we are handling this registration,We checked and found the keyword is your company's registered trademark. But have not registered in China,Hong Kong,Taiwan. So we inform you to confirm whether this regis…

@ASCII-only Yes. People sometimes do it on Puzzling to include hints.

@Dennis Yeah. Dunno how they got your email, but I responded to them appropriately when I saw it.

@Dennis you have only yourself to blame, putting your email plain on your website

I didn't.

Well, I did, but not that one.

hmm, if TIO is mostly used on PPCG I should be able to make a caching userscript for TIO links, especially for the design userscript

@Dennis btw how is the SSD starting to fill up? interpreters?

3:34 AM

Mostly interpreters and the required packages for those, yeah. Tio! has over 70 interpreters right now. It's not that the SSD is close to being full, but if I'm going store user code on the server, I can't afford to suddenly run out of space...

20 GB just isn't a lot of space...

The RAM is much more urgent though.

but 20GB should be plenty for code and output

@AlexA. why do people hate small caps so much

I have 20 GB in total, not free.

yeah, but even 100MB should be more than enough for quite a while

@EᴀsᴛᴇʀʟʏIʀᴋ because it's not ASCII-only

miles

until somebody decides to spam it?

3:41 AM

@ASCII-only and whats wrong with that?

That's not the reason

waffles

sorry

@EᴀsᴛᴇʀʟʏIʀᴋ it's more that it's easy to type, you don't have to click at all

you don't for me name

3:43 AM

@EᴀsᴛᴇʀʟʏIʀᴋ I hate small caps and other unicodey weirdnesses because I can't post comment replies on my phone without switching to the full site or breaking my fingers with copy+pasta, and because it returns the user page's search effectively useless.

user page's search?

codegolf.stackexchange.com/users

oh

serch by username in most places becomes useless

Type to find users: I can't.