@Adám does it use primitives introduced in the latest version?

@ngn No. Works in 12.1 (I don't have an earlier version installed).

i can match rak1507's 15 bytes but no better :|

≢¨({⊣⌿⍵}\16⍴⊂)~⊃ (16 but i know there's a way to save 1 byte from {⊣⌿⍵})

           0*0j1    0⍟0j1
ngn/apl    DOMAIN   DOMAIN
Dyalog       0      DOMAIN
GNU-APL      1        0
NARS2000  NaNjNaN    ¯0
J            0        0
Wolfram    undef      0

> NaNjNaN

@Adám what is this research for

@Razetime I was reporting a few GNU APL bugs and noticed this discrepancy. I am suggesting internally at Dyalog that we return 0 for 0⍟0j1 but I'm not 100% certain about 0*0j1.

@ngn fails on non-simple arrays

@dzaima ⊃ → ⊂¨ fixes that

@dzaima (also fails on shapes containing 0)

@dzaima ah, damn.. you're right

@Bubbler Naja naja

@dzaima bit shorter, even though voidhawk's latest beats it anyways - {¯1↓{0::∇0,¨⍵⋄⍳⍣¯1⊢⍵}↑⍬⍨¨⍵}

another random solution, this one failing on scalars for Some Reason™: ,/(⍳16){0::⍬⋄≢⊃↓[⍺]⍵}¨⊂

@dzaima working: ,/(⍳16){0::⍬⋄≢⊃⊂[⍺]⍵}¨⊂

whats the double : for

@Razetime 0:: catches errors

was trying this: {⎕←≢⍵⋄⊃⍵}⍣{≢⍵=0}

@Razetime ≢⍵=0?

forgot the parentheses

codegolf.stackexchange.com/questions/215436/… anyone got a clever idea to construct the matrix used in the sagemath solution?

@rak1507 one obvious but inefficient way is ⍉↑+⌿¨∘.=∘⎕a¨'ALPHA' 'BRAVO' ..

I think I'm going to cheat and do something devious...

ngn's with a golfed string vector

I reckon I can get a max 50 char solution

@rak1507 dfns doesn't appear to have NATO spelling from what i can see so i have no clue what that is :P

I assume the lack of precision is fine?

@rak1507 the question specifies either mod 2^32 or bigints

Ah ok

I wonder if that makes the matrix solution not work

Probably

I think I'll port the R solution

Getting a working solution isn't the fun bit anyway ;)

@rak1507 why wouldn't it work?

Well I'm not entirely sure if it would or wouldn't either way, but if you took the matrix mod 2^32 at each stage of multiplication, would it not result in something different to what it should be?

@rak1507 with ⎕FR←1287 you can get >64 bits of precision so mod32 after each multiplication should be fine (or something like that idk)

Yeah, I'm just wondering if it results in the calculations not working

@rak1507 it's as if you're using a 32 bit int type and it overflows - the last 32 bits are still correct

(as usual: floating point issues aside)

@dzaima would this work? {0::⍬⋄(≢⍵),∇⊂[0]⍵}

@ngn seems so :D

it feels like there should be ≢[K]Y :)

well I got a working 188, which is decent, but time to get the 50!

I should have a working ±147 but it's not working

(also TIL that newlines are fine in strings in ⍎ed code)

interesting, I wonder what you're doing

Is there a nicer way to find the last occurrence of some element x in a vector xs than doing 1 + (≢xs) - (⌽ xs) ⍳ x?

⊃⌽⍸x=xs

Cool, thanks!

np

@dzaima ah ಠ_ಠ

What is TIO's compression in the URL format?

@rak1507 pako zlib of base64 with a couple replaced characters

Thanks

Is there find and replace in TIO?

@rak1507 no. the fields on it are just regular textareas

That's what I thought

using dfns.base64 or some other compression tool in there might be useful

@Razetime using base64 instead of full 8-bit bytes means 33% more length though

oh well

(which in my case would be 35 lost bytes)

dfns.dyalog.com/n_Data_compression.htm

@dzaima oh, i had ⎕IO←0 in the generator but not in the final solution.. ಠ____ಠ

i have 147 in Extended (classic is being extra annoying and giving some translation errors)

Yay I have a 31.... technically not breaking the rules

@rak1507 technically? :D

I would send it here but the TIO link is too long!

hastebin.com/guwoviciqu.apache

I mean, it does say whitespace doesn't count

@rak1507 that's for the letter counting, not the source

@rak1507 multiline messages allow for longer links

that's hilarious but it's probably under one of the loopholes

Oh oops I read the comment 'Whitespaces do not count. Only letters do. that means someone could implement it in Whitespace at a grand total of "0 bytes", right?' and assumed they were right

tio.run/…

I should've checked

tio.run/… this is my 188

implementing in whitespace sounds like victorian era torture

There are probably transpilers

Or you'd do it with different characters and then sed it

as a hint for my 147, it has precisely 104 full 8-bit bytes of data

unicode

@dzaima oops

@Razetime Turns out it was a user error:

@dzaima Does dzaima/APL not have exponential notation?

@Adám it does, but not on TIO

Ah. I always forget to check that.