The Nineteenth Byte

Ginger Industries

20:00

why do so many Apl symbols look like people

Examples:
⍢⍤⍥

@cairdcoinheringaahing Extra interesting, the case of the word doesn't matter, nor does the cases of each letter in that word

Neil

@pxeger late to the party, but I think -[1, 2] should be [2, 1] and -[1, 2] * -2 is therefore [2, 1, 2, 1] i.e. same as [1, 2] * 2.

hyper-neutrino

do you mean [1, 2, 1, 2]?

pxeger

@Neil oh that's a good point, it keeps the algebra working while also adding a nice logical feature

Ginger Industries

idea: programming language where math works like it should

pxeger

20:05

the only problem being the backwards compatibility

Ginger Industries

i.e. "(" * -1 = ")"

caird coinheringaahing

@GingerIndustries isn't that what APL was supposed to be?

Ginger Industries

@cairdcoinheringaahing whoops

les-citrons

idea: programming language where math works like this: xkcd.com/2313

Ginger Industries

or this: xkcd.com/2295

caird coinheringaahing

20:10

That's just any programming language :P

Ginger Industries

1 + 1 = 2
1 + "1" = "2"
"1" + 1 = "11"
"1.0" + "1.0" = "2.00"
= + "1" = "2.001"
0 / 2 = "("
[1, 2] / 2 = [[1], [2]]
[1, 2, 3] / 2 = [[1, 0.5], [0.5, 3]]

math

caird coinheringaahing

I love that there's a question on the HNQ which is just "Who is Kevin Bacon?"

pxeger

I wonder why I titled it that?

although this remains the best clickbait I've ever seen probably anywhere on the internet

Ginger Industries

[1, 2, 3] * 2 = [1, 2, 3, 1, 2, 3]

caird coinheringaahing

@pxeger I bet its just a coincidence :P

Ginger Industries

20:19

2 * [1, 2, 3] = [1, 1, 2, 2, 3, 3]

caird coinheringaahing

@pxeger "Please delete the Nineteenth Byte" would've been good :P

Ginger Industries

[1, 2, 3][*2] = [2, 4, 6]

[1, 2, 3][*"2"] = ["11", "22", "33"]

pxeger

@GingerIndustries TypeError: int is not iterable

@GingerIndustries TypeError: list indices must be int, not tuple

Ginger Industries

4 ** 2 = 16
4 *** 2 = 32

Presenting Mathyon: Where math works like it should

DLosc

Welp, time to go change my name to Backets reamyappenzell — pxeger Mar 23 '21 at 17:37

Ginger Industries

20:25

[1, 2, 3][[1, 1, 2]] = [1, 2, 6]

@DLosc excuse me

caird coinheringaahing

@DLosc @pxeger you still haven't followed up on this :P

Ginger Industries

can someone get @lyxal to do the GPT-3 thing again?

"Mathyon" is pronounced "May-yun"

caird coinheringaahing

And in England, its "Mathsyon" :P

Ginger Industries

@cairdcoinheringaahing I'd say this would be a good challenge idea but lookup tables so :/

brb

pxeger

@cairdcoinheringaahing not to be confused with the surname Mathyson

caird coinheringaahing

20:30

CMC: Given a string that contains math as a substring, replace math with maths if the code is being run while in the UK :P

pxeger

My proprietary programming language which legally can only be run from my computer in the UK: s->s.string::replace_all("math","$1s")

ladies, gentlemen, and people who don't give a fuck, may I introduce you to the worst piece of code I have ever written in my life: github.com/pxeger/whython/commit/…

@cairdcoinheringaahing codegolf.meta.stackexchange.com/posts/24225/revisions ?

caird coinheringaahing

Wezl for +500 rep — Wezl - a cautionary tale 18 hours ago

user

Are you...looking at the file in which a function is?

No, the folder

Wow, that definitely won't be brittle at all

pxeger

yep; in fact this is so hilariously horrible that I've decided to write a blog post about it lmao

user

May I ask why you need this? (or I guess the blog post will explain it anyway)

I'd appreciate it if you linked the blog post here if and when you write it, I'd like to cringe at and look down haughtily on your Python hacks

Ginger Industries

20:45

back

pxeger

Because a[b] is slightly backwards-incompatible: because of PEP 585 type generics, when a and b are types, a[b] already does something

So overriding it to do function composition will break stuff

Ginger Industries

one second, if we're looking at stupid python code I need to go find something

pxeger

@cairdcoinheringaahing oh, how did I not see that comment on my own post?

lol

caird coinheringaahing

I'm not just nominating people to give their rep regardless of what they want :P

pxeger

I was assuming there'd be some context, but I wasn't expecting to have missed something that obvious lol

Ginger Industries

20:47

Observe the most ridiculous piece of Python code I have ever made:

return (list(_sources.keys())[list(_sources.values()).index(_sources[widget.get_parent().get_parent().get_parent().label.get_text()])], _sources[widget.get_parent().get_parent().get_parent().label.get_text()])

That is all one line

user

Clearly you have either not been coding for very long or you write very good code :P

Because if that's the most ridiculous piece of Python code you've ever written, you'd be horrified to see everyone else's

Ginger Industries

This was from around 2020

I tried to make it one line

@user it's up there

i'm going to look for others

user

If they're really bad, please be sure to also provide us with eyebleach afterwards :P

Ginger Industries

We've also got this function, from the same program (which is now defunct): github.com/GingerIndustries/newsstand/blob/…

user

A good for loop could make it a little better, but...wow

Ginger Industries

20:51

yea :/

and I'm not done yet

    if random.choice(range((monster.get("difficulty") + 1))) != 0:`
        damageTaken = random.randint(((monster.get("difficulty") ^ 2 ) + 1), (1 + (round(random.random() * 10 * monster.get("difficulty")))))`

user

Btw @pxeger are you accepting PRs for Python? I'd like to work on the assigning to constants "feature"

Ginger Industries

only had to edit it 6 times

caird coinheringaahing

assign = all_equal([name,strength,speed,skill,age]) and name == None
self.assigned = not assign

if assign:
    return None

Lovely bit of my old code :P

Ginger Industries

@pxeger speaking of which: can you maybe add some of those math functions from earlier?

pxeger

@user sorry, but I don't think that feature really fits within my ahem Artistic Vision for Whython

user

20:57

:(

What about mvirt's automatic partials/currying one?

pxeger

I want to add things which sound like a good idea, but end up being significantly less usable than expected

user

Ah, assigning to constants is kinda cursed from the start :(

Ginger Industries

@pxeger so my math functions

pxeger

@user I'm not too sure on that either - it's too close to an all-round good feature

wizzwizz4

@pxeger 070 being an octal number.

And 081 also being an octal number (for backwards-compatibility with C).

pxeger

20:58

@wizzwizz4 that was already the case in Python 2

@wizzwizz4 that might have too small an impact to be cursed; I'll consider it

wizzwizz4

lists have a __dict__ and assigning a non-numeric index assigns to the dictionary?

user

@pxeger Hmm, what about if it does the application in reverse? So if bar is a two-argument function, bar(1) would be the same as baz => bar(baz, 1) isntead of baz => bar(1, baz)

emanresu A

@GingerIndustries You do realise ^ is xor?

Ginger Industries

@emanresuA hm

I did not

that explains some things

emanresu A

Lol

pxeger

21:01

@user maybe a bit too random

emanresu A

(So it's flipping the second bit of the number)

pxeger

I'm looking for things that a new user with little language-design experience might suggest to Python

user

oic

pxeger

so not totally obvious trash

but things that really are not a good fit for the language

user

Maybe allowing monkey-patching on builtin types?

I myself really wanted that when I was new to Python :P

pxeger

21:03

Imagine if you took Raku's "just add everything" approach, and filtered it through a ISO C++ style committee but composed of high-school students

4

DLosc

@pxeger TBH, I think your ? operator is the exact opposite: it doesn't sound like a good idea, but it's ended up significantly more usable than I expected

pxeger

@user now that's more like it, although I'm not sure how easy it will be

wizzwizz4

Add <> as a synonym of !=

user

Depends on how it's compiled, probably not too easy

emanresu A

Unicode operators

user

21:04

But if it's simply a check somewhere that you can disable, very easy

pxeger

@DLosc it was directly based on Ruby's inline rescue operator, which is definitely not recommend for real use

user

@emanresuA User-defined operators are a whole 'nother can of worms

pxeger

@wizzwizz4 because Python 2 also had that, it's probably not interesting enough

user

But yeah, I could totally see user-defined operators being suggested for Python, but they're actually useful too :(

New Posts

0

Q: Make an ed clone

ed is the standard text editor on Unix systems. Your goal is to write an ed clone. Task Write a programs that reads an input stream and after receiving a end of line character, print ? followed by a end of line character. The input has to be a stream. Such as stdin, a socket, UART / Serial input...

code-golf

wizzwizz4

21:05

print was a useful keyword.

Why can't we use all functions as keywords?

pxeger

A whole custom operator system would be interesting, and I have been looking into things like this for potential approaches

user

Like Nim?

pxeger

or Ruby

wizzwizz4

You'd have to overhaul the Python parser for that.

user

Yeah :(

Ooh, UFCS like Nim would be nice. Basically, foo(bar, baz) = bar.foo(baz), foo.bar = bar(foo)

All these features are too hard to implement, though :(

wizzwizz4

21:07

Strict typing.

emanresu A

How easy/hard would it be to implement reversed indexing (index[array])

Wait, you can do that in vanilla python I think

wizzwizz4

No, that's annotations.

I tried to implement strict typing in pure Python once, and it crashed Python repeatedly.

emanresu A

I meant ^^^^

wizzwizz4

You can't do that in vanilla Python without subclassing int.

3[list] is impossible.

There isn't a __rindex__

emanresu A

Oh

user

21:10

But if you were able to do monkey patching on builtin types, then all sorts of wonderful things could be possible

hyper-neutrino

i wish that were possible

wizzwizz4

You can do that already, with ctypes.

Let me find the code.

emanresu A

@pxeger Actual ranges (.. or -> operator or something)

wizzwizz4

26

A: Set a Read-Only Attribute in Python?

Despite its dynamicity, Python does not allow monkey-patching built-in types such as file. It even prevents you to do so by modifying the __dict__ of such a type — the __dict__ property returns the dict wrapped in a read-only proxy, so both assignment to file.write and to file.__dict__['write'] f...

user

Also github.com/clarete/forbiddenfruit

hyper-neutrino

21:19

oh, i c

wizzwizz4

Personally, I prefer making the number 43 a string.

You have to avoid 44 if you do this, though.

DLosc

> from forbiddenfruit import curse

Oh dear X^D

wizzwizz4

@DLosc \x1c

Actually – undefined variables are treated as strings.

Rationale: tkinter.

Instead of a NameError, you should just get the name you used as a string.

DLosc

@wizzwizz4 I've definitely thought of making a language like that

wizzwizz4

One already exists.

caird coinheringaahing

21:27

Perl

DLosc

You can assign to barewords in Perl?

caird coinheringaahing

Not sure, wouldn't surprise me tho

DLosc

@wizzwizz4 File separator? Not sure what you mean

wizzwizz4

@DLosc bytes(map(operator.xor, b"X", b"D"))

DLosc

@wizzwizz4 Ah, lol

Though, coming from a QBasic background, I'd say X^D really ought to be 1, unless X or D was already defined.

wizzwizz4

21:32

Why 1?

caird coinheringaahing

xor is essentially does not equal

X != D, so 1

wizzwizz4

Yeah, but it's bitwise.

DLosc

@cairdcoinheringaahing No, ^ is exponentiation

(obviously :P)

wizzwizz4

Maybe ^ should be exponentiation for floats.

DLosc

And undefined numeric variables default to 0

wizzwizz4

21:34

Or maybe it should XOR the floats' raw representations.

caird coinheringaahing

You should be able to do bitwise on floats

DLosc

@cairdcoinheringaahing Doesn't look like it

wizzwizz4

Note to anyone overriding magic methods on Python types; you need to be a bit fancier than just altering the __dict__.

caird coinheringaahing

@DLosc No, I want it to be able :P

DLosc

@wizzwizz4 Are you referring to Tcl?

wizzwizz4

21:43

@DLosc I was actually thinking of PHP. I should've said “many”, rather than “one”.

DLosc

Oh, really? I've only ever seen $ variables in PHP, and I get an error when I try it without. Is that one of those php.ini settings?

Redwolf Programs

@wizzwizz4 Barewords?

user

Kinda like strings but without quotes

Remember wat in Ruby where they had bare strings with method_missing or something?

@DLosc Don't worry, you can reverse() it ⍳]

DLosc

@RedwolfPrograms Barewords, but assignable. Or, alternately, every variable is preinitialized to its own name as a string.

Redwolf Programs

Oh huh

Fmbalbuena

21:55

Challenge: print regex that checks the divisibility of input number

caird coinheringaahing

80

Q: Hard code golf: Regex for divisibility by 7

Matthias Goergens has a 25,604-character (down from the original 63,993-character) regex to match numbers divisible by 7, but that includes a lot of fluff: redundant parentheses, distribution (xx|xy|yx|yy rather than [xy]{2}) and other issues, though I'm sure a fresh start would be helpful in sav...

code-golf math regular-expression

Redwolf Programs

/^(.*)\1$/ works with unary input, right? (with the number of \1s being the number to check divisibility by)

Fmbalbuena

> Challenge: print regex that checks the divisibility of input number

caird coinheringaahing

If you mean in decimal, I linked that challenge to show how difficult it is for 7, let alone higher primes

DLosc

@pxeger This inspired me to search "butter" in the TNB transcript, and I discovered a huge conversation about peanut butter back in July 2015 that's really funny

Fmbalbuena

21:59

decimal or binary.

Dom Hastings

@DLosc you kinda can set bare word vars in Perl and it supports barewords for anything undefined: Try it online!

caird coinheringaahing

@Fmbalbuena This might work as a challenge, where your score is how high you can go such that it works for all numbers below that (i.e. if you can do divisibility by 1,2,3,4 and 5, but not 6, your score is 5)

Not sure how much of it would just be lookup tables tho

Fmbalbuena

@cairdcoinheringaahing ok

don't make golfed

DLosc

22:14

@DomHastings Ah, interesting

Dom Hastings

It's a bit convoluted, but I was playing around with it ages ago trying to make a golfer Perl! Not much use if you need to predefine usable vars for every eventuality though I imagine...

Redwolf Programs

22:50

It was below freezing today, and due to some misconfigured sprinklers at a city park, I saw ice for the first time this winter!

user

We've had 3 snow days this week, winter break basically got extended by a week here :)

Redwolf Programs

Y'all are lucky, probably no snow for us this winter :/

caird coinheringaahing

Literally the entire UK got snow except for the city I was in a few weeks ago

DLosc

No snow here today, just cold

(Okay, not very cold. I haven't been outside since morning, but it looks like it warmed up a fair amount this afternoon. Still below freezing, tho.)

lyxal

23:22

@cairdcoinheringaahing you can thank my strategic vote capping for that

I've been going for the full 40 vote cap recently instead of the standard 30

@DLosc that's the period of time that I was active on esolangs.org and when I moved away from exclusively answering on Stack Overflow lol

Neil

23:50

@cairdcoinheringaahing not seen snow recently here

Unrelated String

@pxeger maybe adding ?= such that x ?= [expression] is equivalent to x = ([expression]) ? x but i can't help but feel like any language designer would hesitate to switch the order like that

...that's another idea
backward versions of every augmented assignment operator

not sure if it would be worse to create an __ri[op]__ dunder for each of them or to refuse to

user

Ugh, I've been trying to route all assignments foo.bar = baz to my custom assignment function in Python, but there's no easy way to call it from C :(

Hacking the CPython interpreter is not as easy as pxeger made it out to be

Transcript for