The Nineteenth Byte

10:00 AM

@lyxal Idea: What if you create more Vyxal servers (lyxal1.pythonanywhere.com,lyxal2.pythonanywhere.com,lyxal3.pythonanywhere.com etc) and route the processing requests to the next available one?

Wasif

Yeah then we can accommodate more Vyxal users

But what I personally do is clone Vyxal repo and run locally

for fastest speed

Bubbler

The owner of pythonanywhere won't like that (or will want to charge you for that)

A username

Make them look different then - ihbefn.pythonanywhere.com, hiropviefui.pythonanywhere.com, ueihrgojf.pythonanywhere.com etc

Bubbler

That's even more sus :P

Anush

dear coding gurus, is it possible to replace qsort with the C++ sort routine in C code somehow?

Wasif

10:03 AM

If you want I can donate my pythonanywhere server

A username

I can donate one as well

I don't have one but it's free right?

Wasif

@Ausername yeah

I have one at 2ddeadfish.pythonanywhere.com

A username

Ok

We can ask lyxal when they get back

Bubbler

@Anush I think this is the answer

basically write an extern "C" function that wraps the sorting routine on the C++ side

Wasif

what is the most well documented 2D language you know?

Bubbler

10:08 AM

@Anush Protip: when you have a general programming question, google it and check out the relevant SO question

xash

@Anush If this is for performance reasons, this could backfire: the cmp-function cannot be inlined by the compiler then.

Bubbler

@Wasif Gol><> is pretty good, despite having lots of commands

Wasif

thx i Will check that

StackMeter

@Wasif what's 2ddeadfish

Anush

@Bubbler thanks!

Wasif

10:12 AM

@StackMeter my own 2D language: codegolf.stackexchange.com/questions/200306/scream-very-loudly/…

Anush

@xash it is for performance. The problem is that qsort can't inline the comparison

Bubbler

@Wasif and all of Martin Ender's 2D langs come with quite descriptive READMEs, which is a different kind of "well documented" from Gol><>'s

rak1507

hexagony lol

Bubbler

(which includes Alice, Wumpus, Labyrinth, Hexagony, and maybe more)

xash

@Anush Have you actually checked the assembler output? A proper compiler/stdlib will detect qsort and inline the function with -O3 or something, despite it being declared elsewhere.

Anush

10:19 AM

@xash I haven't but the web is full of articles explaining why qsort is slow

Jo King

@Wasif This isn't Turing Complete right?

Wasif

@JoKing yeah not turing complete (intentionally)

But I was surprised Bubbler wrote a parity test code using it

I was able to write a truth machine too

Jo King

not too hard, you just decrement by two until zero or one right?

Wasif

@JoKing it has only checks if accumulator is 0

Still I don't know how it works

Jo King

well yes, decrement the one before doing that check i assume

not sure why
```
>(oh)d v
^d)hoi(<
```
doesn't work

StackMeter

10:27 AM

ok

Jo King

oh, brackets only work left to right? that's kinda weird

Wasif

@JoKing yes AFAIK

Jo King

ouch, they don't even work up to down either

Wasif

its just a rip off Deadfish~

nothing serious

Jo King

no nested brackets??

i was thinking

>(d(v)i)o
^   <

for a parity check, but unfortunately doesn't work

um wait, {} doesn't execute a block ten times, it executes each commands it finds ten times??

And if you do {{}} it executes it 100 times?

i like that doing } at any point without first doing { basically dooms your solution

A username

10:45 AM

I like challenges with funny reasons for short code, e.g. hyper's "You have to execute your code by hand because your computer was stolen"

Jo King

@Bubbler i see bubbler made the same mistake with brackets as i did

best i can manage is 23 bytes

(oh)d(ioh)v
^d        <

rak1507

it's always frustrating having so much empty space in 2d langs, found that in ><> as well

Jo King

well this one is basically because ( always sets the direction to right

><> is much better about direction-agnostic commands

A username

Just edit it out'

Of the interpreter

Adám

@Wasif You may want to frequent apl.chat but either way, let me know if I can help.

Jo King

10:51 AM

that's kinda beside the point. i'm actually enjoying programming in this language because of how bad it is

rak1507

lol, harsh

lyxal

@Ausername I could do that, yes. But the question is whether or not the process of doing so is actually worth it.

You see, in order to make a new site, I'd have to create a new account

Each new account would have to be updated with each commit

Sure, there's a way to have it autoupdated, but there's still another problem

Every 90 days I have to click a button that says "keep the site hosted", otherwise it gets deactivated

Now if I have multiple accounts, that means I need to keep track of all their expiry dates

Bubbler

@Anush Just don't tell us you want to use Rust's sort in C (which uses a different, faster algorithm than C++'s std::sort) :P

lyxal

@Ausername of course, that also isn't a major issue. It's just something to consider

Anush

@Bubbler ooh.. good idea :)

Jo King

11:05 AM

@Wasif should i report issues here or in the Github?

A username

@lyxal True. Another idea could be to make the default timeout 10 seconds and allow that to extend up to 60 via flags. The only time a program takes longer than a second for me is when I've entered an infinite loop :p

Also is there a builtin for [[1,2,3],[4,5,6]] => [[1,4],[2,5],[3,6]]?

Transcript for