The Nineteenth Byte

Veskah

00:00

The coin with the Queen's head plus the coin with running legs combine into a pretty great picture

Adám

Maybe that's the 25p one and a few of them are slightly less than they say (we know they do partial p) so rounding gives the visible values, while summing the rounded values gives 102.

Veskah

i.imgur.com/OhUkcmT.jpg This one

Adám

@Veskah Hidden message?

Veskah

@Adám Deepest Lore

New Main Posts

0

Q: Densely packed decimal (DPD) to decimal

Background Densely packed decimal (DPD) is a way to efficiently store decimal digits in binary. It stores three decimal digits (000 to 999) in 10 bits, which is much more efficient than naive BCD (which stores one digit in 4 bits). Notations The lowercase alphabets a to i are the bits that ar...

Neil

01:21

@Adám only 3 months? huh, my bank will only show 3 months on a page, but you can choose those three months from a range of several years

@flawr nice, I've got several sets of coins, I've always wondered whether there's a diagram I can print out and then stick them to for posterity

Thomas Ward

01:56

@flawr they're fine they're just normally like that :p

Veskah

02:49

I'm crap at JS, codegolf.stackexchange.com/a/176372/78849 Is this answer actually able to produce all patterns? I've ran it several times and only got permutations of !!!..

Which is what it seems to do

ETHproductions

03:10

@Veskah That's exactly what the [...'!.!.!'].sort part is written to do. You'll need to generate 5 random characters instead, e.g. [...Array(5)].map(x=>'!.'[Math.random()*2|0])

other than that, the answer looks very good :-)

Veskah

@ETHproductions Alright, glad I'm not losing my mind

ETHproductions

Actually [0,1,2,3,4] would be shorter than [...Array(5)]. I'm just so used to the trick that I forgot you could do anything else, haha

[...Array(5)].map(x=>'!.'[Math.random()*2|0])
[0,1,2,3,4].map(x=>'!.'[Math.random()*2|0])
[...s='!.!.!'].map(x=>s[Math.random()*2|0])

Downgoat

03:38

do you prefer global sin(x) or a math object? Math.sin(x)

lirtosiast

@Downgoat I'm so used to golfing languages that I expect way too many things to be global

DJMcMayhem

04:27

@Downgoat The latter

quartata

Sin/cos can be in a module but sqrt must be global

i think python having sqrt in a module but abs be global is one of the most annoying things about it

Veskah

@quartata Pretty much all nice math in Powershell is behind .net's [math]

Downgoat

@quartata what about log functions then?

quartata

That can go in math too

Downgoat

I mean technically you can always do x ** .5

should Int ** Double be a thing?

on one hand lets you do what I said above but then again kinda weird

Veskah

04:48

Feels like it should promote the int to double

However, allowing Int ** Double while keeping it as int might be neat in a handful of applications. Otherwise, seems like it'd be a gotcha waiting for the unaware

Bubbler

The result of ** is complex in worst cases, be the base int or double

New Main Posts

05:46

0

Q: Flipkart coding question hackerrank

You are working in Samara, Russia for a few days. Each day has a new pay per unit of work and a new cost per unit of food. Working 1 unit costs 1 unit of energy, and eating 1 unit of food adds 1 unit of energy. Here are some specifications of your employment: You arrive with no money, but with e...

code-challenge programming-puzzle

feersum

06:37

People keep talking about this "nandgame", but I assume it would be boring to just make circuits for a desired function because NAND isn't esoteric enough.

Is there anything more to it?

quartata

i figured it was someone trying to trick me into doing their digital logic design HW And closed the tab immediately and disinfected my computer with a blowtorch

feersum

Yeah see, that's why I'm asking before opening it.

quartata

they cant stump me. i am Protected from help vampires in all walks of life

feersum

What about nerd snipes?

quartata

no but i have someone on retainer to snipe back upon the event of my demise

For real though didnt we all do NAND gate to an ALU at least. i was only slightly joking when i said this was homework

if theyre really just logging all the answers for their homework I Admire their Creativity

this is how frontend people solve hardware problems

Adám

06:46

@feersum At least it requires more thinking than clicking out paper clips or feeding ants.

feersum

'Tis so much more noble to solve a problem with 10 hours of coding, than 10 minutes of pencil and paper.

DLosc

2-3 of the levels I completed, I didn't get it in the smallest number of components. I'm still a bit irritated

quartata

and 10 hours of mood board construction

@DLosc wait theres golf

ok now its good. its good again

Zacharý

However, if you try it with fewest NAND gates ...

... which it doesn't track

quartata

no point though

and & or are still straight up CMOS

nand is just cool to have one primitive in theory

Downgoat

06:53

I think after like 3 years I have finally got dynamic memory in VSL to work :DDDD:

define i32 @main(i32, i8**) {
entry:
  %2 = call i8* @FSTCPointerNallocAcountTCUInt64.generic.TCUInt8(i64 1)
  %3 = alloca i8*
  store i8* %2, i8** %3
  %4 = load i8*, i8** %3
  call void @free(i8* %4)
  ret i32 0
}

now with generics you can do a Pointer<T>.alloc(count: UInt64) and it'll generate a malloc() & free() call where it expects

what's cool though is unless an object has a possibility of being assigned somewhere we can generally predict where to place deallocation calls

feersum

What other kind of memory did you have?

quartata

this is glibc malloc ?

Downgoat

@feersum stack allocations

technically dynamic memory kinda sorta existed but no direct pointer-access

and also it wasn't actually managed so it wouldn't be cleaned until program exists

@quartata yeah

feersum

Or whatever C runtime you're using.

quartata

i love having the os reclaim all my memory. true userspace power move

@Downgoat are you going to try to make your own allocator

Downgoat

06:58

@quartata in that case you can just disable the memory optimizer pass and have no free calls but also no slow ref counts :P

@quartata i dont think my allocator could be better than what the smart people who make glibc can make

quartata

actually youd be surprised

the problem is threads

feersum

The glibc malloc creates per-thread arenas, so supposedly it shouldn't be locking on every call...

ngn

@Downgoat glibc's is crap. creating your own is easier than you think

quartata

@feersum that got duct-taped on

real ancient malloc hours

Downgoat

what syscall do I do to ask OS for memory?

ngn

07:02

@Downgoat mmap

feersum

or brk

or is it sbrk, I don't remember

quartata

sbrk is relative

brk is absolute

feersum

It does become unportable if you do this stuff

quartata

rust uses jemalloc now yeah?

ngn

"sbrk() is implemented as a library function that uses the brk() system call" --man

@feersum who cares about windows :)

Downgoat

07:08

@ngn do you have link to any numbers?

ngn

i heard once that windows can run linux elfs. i don't know if it's (still) true

@Downgoat no, but i can feel it :)

feersum

Like WSL or something else?

ngn

in my language i got rid of libc altogether

i mmap a lot of memory in the beginning and manage it as a buddy system

it's very simple and efficient

@feersum ah, so that's what it's called...

feersum

I still have never tried WSL.

Pavel

@feersum If you've used Linux ever you really won't be surprised by anything. The terminal is kinda crappy and the fact that your Windows home isn't the same as your WSL $HOME aren't the same is annoying.

feersum

07:19

Hmm personally I think it would be more annoying if WSL home was the same.

Pavel

If you use not-WSL it's annoying when ~ refers to different things

feersum

Oh, mixing WSL and other unixy shells?

Downgoat

is it just me or does LLVM not have bitwise not O_o

feersum

Seems not very important since you can -x -1

quartata

@Downgoat you mean logical? it has bitwise

@feersum assumes twos complement

Downgoat

07:33

@quartata looks like doesn't natively have bitwise but the IRBuilder/BinaryExpression class expose helper function to create

feersum

@quartata Which LLVM does.

quartata

@Downgoat google says BinaryOperator::CreateNeg(nbits, "bitwiseNEG", insertBefore);

(is this not how youd normally make it, I never used LLVM frontend)

Downgoat

@quartata O_o your compilers directly write IR?

feersum

In my LLVM-based language I just generated textual IR with Python string format operations.

Downgoat

> Is binary operator so must have 3 args — source

this is why it took me a day to unbreak operators :P

New Sandboxed Posts

07:45

0

A: Sandbox for Proposed Challenges

Densely packed decimal (DPD) to decimal, with logic gates atomic-code-golf logic-gates Background Densely packed decimal (DPD) is a way to efficiently store decimal digits in binary. It stores three decimal digits (000 to 999) in 10 bits, which is much more efficient than naive BCD (which store...

Bubbler

^ inspired by nandgame

Downgoat

08:07

for types that haven't overloaded == should the default behavior be to check addresses or just error. I'm thinking of adding a === anyway to check address equality

flawr

08:28

@ThomasWard so it is just the photo? you know that photoshop and various other image editing softwares have red eye removal tools? :D

Cows quack

09:30

oh cool, anagol has a one less byte for modular multiplicative inverse spoiler

New Main Posts

-1

Q: Increment a number without using ++ or +

My Question is to Increment a number without using ++ and + operators. Input : 5 Output : 6 Same logic should execute with character data type as well, Input : a Output : b

code-golf c programming-puzzle

Cows quack

oh wait, that's because they don't have to check for invalid ones hmm

Erik the Outgolfer

12:44

@flawr are you sure Thomas doesn't want to embrace just how he actually is? :P

@Downgoat better put functions that aren't going to be used outside of certain contexts in built-in libraries for those contexts

New Sandboxed Posts

14:05

0

A: Sandbox for Proposed Challenges

Exploring generalzations of Cheryl's birthday problem In this challenge we will generalize the Cheryl's birthday problem to arbitrary finite subsets of pairs of integers Challenge Cheryl gives Albert and Bernard a finite collection of pairs of integers S. Cheryl chooses an element (x, y) from ...

user202729

15:31

I'm trying to find out what can be done ini 38% off Javascript.

So far I found a way to assign/retrieve variable: spoiler

but it takes 5 bytes, unfortuantey.

Zacharý

Happy thanksgiving to all the people in the U.S.A.

Skidsdev

15:48

Yeah happy turkey day guys

ngn

happy no-fuсk-giving to the rest of the world :)

user202729

Anyone have any idea to modify any global variable (in 38% off JS)?

I tried calling functions that modifies this like let y;new Set().add.bind(y)() but the types are not compatible.

Why do both new Object() and Object() works?

dzaima

16:04

@user202729 because js. Also new Object

New Sandboxed Posts

0

A: Sandbox for Proposed Challenges

Output The End Poem. Outputting as array, or with only one \n or three \n to separate sentences, etc. are fine. Upvote to prefer: You can take an input containing all letters in lowercase or uppercase Downvote to prefer: You can take an input containing all letters. You can decide each letter b...

ngn

@dzaima and {}

dzaima

@ngn but } is not allowed in the challenge

user202729

Ok, got it.

Basically there are some functions that modifies this

and I am thinking that let y; somefunction.bind(y)() may help modifies y.

user202729

16:26

However, the only possible values that can be assigned is undefined or NaN, which are both immutable.

tsh

16:40

If you don't mind to switch Spidermonkey to browser environment (e.g. Firefox) Use parent.b would make the variable name shorter.

And print should be switched to alert if so.

@user202729 So, maybe parent.__defineGetter__('b',''.replace.bind(8>>98));alert(b) in browser?

user202729