The Nineteenth Byte - 2021-12-07 (page 2 of 4)

« first day (3965 days earlier) ← previous day next day → last day (881 days later) »

9:01 AM

Properly implementing Haskell sequences is whole new level

If you want that, you basically need to model your language around lambda calculus in some form

what's hard about them? The fact every single computation is thunked? Or the fact they can recurse on themselves?

uh jelly + lambda calculus = ????

Husk?

no; helly

> hell

Unrelated String

i've unironically ocnsidered the name helly

9:04 AM

so helly -> (h)askell + j(ellly) -> hell-y

@UnrelatedString why the hell[y]?

Unrelated String

oh yeah wait i was actually considering using that name for literally porting jelly to haskell

@pxeger Letting nilads recurse on themselves

@Bubbler whats haskell sequences?

Something like fib=0:1:(zipWith (+) (drop 1 fib) fib)

9:06 AM

Haskell is hard to understand

@Bubbler can't they just be turned into monads with an ignored argument?

I guess that basically ends up turning everything into a thunk though

@pxeger Why:

yes that was/is my sock

ZORP THE SURVEYOR, shoes

1 1

Unrelated String

shoes

9:08 AM

Oct 15 at 12:17, by pxeger

I only have two socks: ZORP THE SURVEYOR and ATO

@UnrelatedString yeah, socks ≈ shoes

Unrelated String

socks are logically located within shoes

I renamed it to ZORP for a joke in chat; I can't remember the exact context

Btw I love the fact that AoCG day 6 and 7 are at the top of HNQ at the same time

and I haven't used it since so yeah

9:09 AM

Sep 1 at 19:50, by user

PXEGER HAS BECOME ONE WITH ZORP

Sep 1 at 19:44, by Dude coinheringaahing

> Zorp the Surveyor is the 28-foot-tall lizard-god savior of the cult known as the Reasonabilists that took over Pawnee, Indiana in the 1970's.

The thing is, that account has 1 rep - how did it chat?

I think it had an answer on a question which has since been deleted

look at doorknob's rep

@pxeger No, it's not that simple. Even OCaml (a functional but strict language) has to do this abomination

@Bubbler huh lol

can someone golf this even more?

CMC: ^

9:21 AM

Nope

Unrelated String

yes lol

Oh neat

Even shorter

wat lol

Python map can be used as zipwith

9:26 AM

Slightly shorter, output as generator

(Blatant ripoff of bubbler's)

Or if rearranging args is acceptable then map is a valid 3-byter

True lool

wow

MISC: Try it online!

^ make it short challenge

Unrelated String

you can start by using a generator comprehension instead of a list comprehension

but then you can replace that with map

9:52 AM

Try it online!

I guess it's pretty much you can do in Python

i am getting MAD AT android BROWSERS

none of them allow to change fonts

classic

??

10:14 AM

blog.xkcd.com/2008/09/09/…

11:01 AM

@PyGamer0 fun thing to add to flax: vectorised repeat / repeat iterable n times

so like ["Fizz", "Buzz"] <op> [1, 0] -> ["Fizz", []]

(pretend that those strings are lists of character codes)

or at least ["Fizz", "Buzz"] <op><zipwith modifier> [1, 0] -> ["Fizz", []]

11:13 AM

i have an atom replicate, and a modifier each

so i might add that as a single command

11:26 AM

although

@lyxal how useful is that?

idk

uh

eh

then do i really want that

you'll only know what's more useful with experience.

yeah

1 hour later…

12:44 PM

@Bubbler I mean, Scala can do that and it’s not based around lambda calculus at all

I think it can do what you mean, at least

12:55 PM

New Posts - Challenges - Code Golf: Find the IP address class

posted on December 07, 2021 by Razetime‭

Task Given an IP address as a string, find its IP address class. For reference, here is a table:

how do i search only answers in vyxal, 05ab1e and jelly that have explanations?

You can't really.

@PyGamer0 SEDE probably

1:35 PM

select id as [Post, Link], body from posts where body like '%Jelly</a>,%' or '%Vyxal</a>,%'

^ would that work?

it worked

its like python's or

my first query: data.stackexchange.com/codegolf/query/1527006/explained-answers

GingerIndustries

1:53 PM

good morning

challenge: Given a string that contains only superscript numbers 0-9, return that string as an integer. (example: ¹² -> 12)

python, 61 bytes: lambda n:int("".join([str("⁰¹²³⁴⁵⁶⁷⁸⁹".find(c)) for c in n]))

Probably just eval in Perl6.

GingerIndustries

well sure

Much easier to solve for subscript than superscript.

GingerIndustries

why?

Because they are consecutive in UCS.

GingerIndustries

2:04 PM

good point

i still think the above method works pretty well

@GingerIndustries Try it online!

argh its more than yours

GingerIndustries

you could probably make mine smaller with regex

but I think w/o regex it'd be hard to get it smaller

g'morning folks

GingerIndustries

hi

@GingerIndustries dzaima/APL: 10⊥'⁰¹²³⁴⁵⁶⁷⁸⁹'⍳ Try it online!

2:13 PM

i really wish python had function composition :/

GingerIndustries

ah well

not having function composition builds character

Try it online!

^ if function composition worked

GingerIndustries

could you imagine

yes

@pxeger add function composition to whython pls

and / or add it to python using that gc module

Can't you simply implement composition?

GingerIndustries

2:18 PM

pom pom pom

working on another language

like i want f . g == f(g(*args))

would this be a dupe? [this challenge](https://codegolf.stackexchange.com/questions/92168/slowly-turn-a-string-into-another) but
1. you cant remove any characters from the first string that will end up being in the second string
2. you dont have to do the like, removing from the end of the string always thing

oh right, no formatting for multiline messages

thankfully it didnt embed

8 hours ago, by PyGamer0

flax: codepage feedback?

^ @RedwolfPrograms

argh i cant remember which atom i was supposed to add

ah square root

Redwolf Programs

2:43 PM

I'd drop ⓑ

It's inconsistent with the rest

Personally I'm also not a fan of unicode super/subscripts, but those probably aren't going to go

⤒ renders as too narrow for me, as does ∇ and ⍳

Lots of your characters are inconsistent widths, actually

That's not a massive issue, but it is going to be annoying

Font issues, though. Not codepage issues.

Redwolf Programs

I'd argue that is a code page issue

If most fonts consider it the wrong width

Including the ones used on CGCC :p

SE refused to fix our font stack issues. Easy to solve with user style sheets.

Redwolf Programs

You know they don't make the fonts right :p

All fonts I've seen have that issue

yesterday, by PyGamer0

so i will force everyone to use Iosevka Term :P

2:54 PM

@RedwolfPrograms That's not true. I've seen TIO, so you've seen DejaVu Sans Mono. It has almost all the characters in that codepage.

tio.run/##fZJZT5NBFIbv51c87gtW2Rf3fd/3FS0UcKmCO@hoSom0/…

so what should i use instead of ⓑ

What does it do?

a nilad that outputs bits [0,1]

Redwolf Programs

I wouldn't try to make the glyph relevant to the action it does

@PyGamer0 Maybe ⚆

2:57 PM

@Razetime idea: i might try to solve some AoC with flax

Redwolf Programs

Pick a character from the US-INTL layout and use that

¦?

Or how about ƀ?

@Adám no, its emoji

@PyGamer0 Wait, really? Screenshot? Oh, only Samsung. Odd.

@Adám ƀetter :p

user image

@Adám how does it look like for you?

3:00 PM

Like ○ with a ∙ right off centre.

Better like this? ⚆︎

I wonder what Samsung thinks it looks like.

@Adám still looks the same

Yay, non-conforming rendering.

i used ƀ

for [0,1]

Fun alternative: Ꮊ

@RedwolfPrograms iosevka term has all the glyphs from the codepage

3:03 PM

What does ⊰ do?

Redwolf Programs

That's cool! Consolas doesn't, Roboto Mono doesn't, Cousine doesn't, Courier doesn't, Source Code Pro doesn't, should I go on?

@Adám suffixes

prefixes is his brother

@RedwolfPrograms then blame their creators :p

I don't see the mirrored one.

Redwolf Programs

Sure, but your job as a code page designer is to work around that

7 hours ago, by PyGamer0

user image

Redwolf Programs

3:05 PM

Otherwise, regardless of who's at fault, your code doesn't line up right for most people

argh code page stuff is hard

...I have learned that I am not very good at language design

@Mayube Don't worry. Everyone is presumed "not very good at language design" until proven otherwise.

Redwolf Programs

That's not a very encouraging thing to say...

It's something that takes practice, like anything else

why cant there be a single monospace font wihh all glyphs :/

Redwolf Programs

3:08 PM

I'd recommend sticking to the US-INTL keyboard pretty strictly

Most stuff there renders fine in most fonts, and is faster to type for people

@Adám That only matters as far as other peoples' opinions on my languages, which I don't particularly care about. I'm more concerned with the fact that I have cool ideas but not the skill to realize them.

@RedwolfPrograms who are in usa

Redwolf Programs

@PyGamer0 Not necessarily

I switch between my preferred layout and US-INTL whenever I do Jelly stuff

It's just Ctrl + Space for me

@PyGamer0 Because the font formats have a cap on how many glyphs they can have. Also, it is a massive amount of work to design all the 140-something thousand glyphs.

Redwolf Programs

@Adám I doubt that's the reason

The first one, I mean

3:10 PM

107

A: Why isn't there a font that contains all Unicode glyphs?

"Why would you even want that?" questions aside, from a programming perspective there's a very simple reason: the OpenType spec only affords an addressable glyph index space of one USHORT, so one font can only support 16 bits worth of glyphs identifiers, or 65,536 glyphs max. (And note the termin...

@Adám i have heard unifont has around 10k glyphs

Redwolf Programs

@Adám 65k is so, so much higher than any commonly used fonts I've seen

@PyGamer0 I don't think you understand what the gc module is for

Redwolf Programs

And I'm pretty sure other formats lack that restriction

@pxeger what it is for?

3:12 PM

it's an interface to Python's garbage collector

it wouldn't be able to implement a new operator, which seems to be what you're suggesting

oh

but you can do it in whython

right?

I could add it to whython, yes

@RedwolfPrograms to be fair, if it's impossible to make a font with everything, it probably doesn't make sense to bother with that many glyphs anyway

and I think it's actually the kind of good-but-actually-bad idea I want for whython

But... why(thon)?

3:13 PM

so are you gonna make it?

if it were possible to make a truly universal font someone 100% would've done it

Redwolf Programs

I'm talking about fonts in general

Not ones made specifically to have a lot of characters

ok so i might change the codepage tmrw with better characters

Redwolf Programs

@RedwolfPrograms I don't care if RedwolfIsSuperAwesomeMonoProPlus has every character in Unicode, if Consolas and Roboto Mono and Cousine are what people actually use, those are what matter

@hyper-neutrino GNU Unifont does cover the whole BMP, which is a lot of characters

3:15 PM

> RedwolfIsSuperAwesomeMonoProPlus

^ someone should make that

Google's Noto has 65,535 glyphs. Hm, I recognise that number…

There's also Unicode Last Resort: github.com/unicode-org/last-resort-font

@Adám so its already done?

Redwolf Programs

@Adám Maximum for older formats, yeah

No reason you couldn't use woff or woff2 and add all of Unicode, fill the PUAs, and make up some others just for fun

@pxeger wait, that's not actually what I thought it was

@RedwolfPrograms but would the client-side renderers support a font with that many glyphs, even if the library reading the font file supports it? (e.g. libxft)

Redwolf Programs

3:19 PM

I don't see why not

> WOFF files are OpenType or TrueType fonts, with format-specific compression applied and additional XML metadata added.

Redwolf Programs

Oh lol

Yeah the spec's kinda dense

It mentions something like a glyph count that's a Uint32

That must refer to something else

But like I said, this is totally irrelevant to why most fonts don't have support for that many characters

Designing fonts takes a lot of work, and people rarely care about weird characters lining up correctly

Most don't because of the work involved, but none do because of the technical limit.

Redwolf Programs

But the BMP fits in 65536

So?

Redwolf Programs

3:21 PM

And everything outside of that's basically irrelevant for code page design

Unicode has already defined twice that.

Redwolf Programs

We don't need emoji, PUAs, or whatever's in the SMP

The point of this conversation is code pages, right?

Mathematical alphanumerics are in SMP.

But right, it is possible to make a font that covers all of BMP. Google has done it.

Unfortunately, afaik, nobody has done it in a monospace font.

Redwolf Programs

Don't see why it's unfortunate

When I was experimenting with making an APL font, I tried merging missing glyphs from Noto Sans into Noto Sans Mono, but I didn't have the skills to do it right.

3:58 PM

Whoops, AWS us-east-1 is down...

Redwolf Programs

Maybe someone drove a train into the datacenter

4

Of course, status.aws.amazon.com shows everything to be A-OK, but then again, look at which domain hosts that page…

Redwolf Programs

user image

That's some interesting justification

I have a coding problem that is stretching my abilities. I was hoping at expert might be able to help

Aaroneous Miller

@RedwolfPrograms relevant xkcd

4:17 PM

@RedwolfPrograms oh, text justification

I was trying to work out how they were justifying (as in explaining) the course level changes lol

@PyGamer0 an incomplete language is very inconvenient

caird coinheringaahing

If anyone wants to know what kind of a mess SE's JavaScript is in, take a read :P

2

ugh, cygwin doesn't trust any CAs and it's a real pain in my ass :(

caird coinheringaahing

@RedwolfPrograms R E Q U E S T E D

Redwolf Programs

503 on collegeboard's website lol

4:30 PM

ah, it wasn't trusting CAs because for some reason curl was using a CAFile from my windows install of Neovim...

4:44 PM

anyone very kind know how to do this? stackoverflow.com/questions/70263808/…

@pxeger thanks very much. I will delete this question and replace it with one that just asks how to aggregate by day. Do you think that is better?

@pxeger actually can you tell me what is not clear please?

@Anush you didn't specify in the question that you wanted it grouped by both user id and by day. It looked like you were going to assume the day was always the same for each user id

@Anush that would probably just be a duplicate of "how to get day from datetime object" and/or "how to group items in a list by value"

@pxeger let me clarify this in the question

@pxeger I am googling....

stackoverflow.com/a/5695268/1473517 is an example of group items in a list by value but I am not sure how to use it for my problem

@Anush I would just do it with a dictionary (pseudocode: ato.pxeger.com/…)

O(n^2) is also no good so I should use groupby as you say

Actually groupby only works for consecutive elements, so that's not what you need I realise

5:01 PM

let me see if I can work out what get_day_of() should be

@Anush so

you want two separate groups or one group split in two ways?

because the first is possible, the second is impossible.

@ThomasWard I want for each day to have all the texts for a given user to be aggregated

there will be a lot of days and a lot of users

oh so 'user' is going to be specified then.

yes there is a user_id in each triple at stackoverflow.com/questions/70263808/…

just as an FYI

#IHateUnnamedTuples

because i have to do work to remember what's what :p

but i digress.

sounds like an interesting problem for me to stab my brain at

5:06 PM

@ThomasWard named tuples are just as bad, because you still have to remember what order the things go in

Python needs proper object destructuring

but I digress...

true

@ThomasWard yes please!!

that would be awesome, thank you

but they're better because you can do namedtupleobject.specificitem - more like a class with specific methods, etc.

@Anush no guarantee of a nice solution though

i do have an FT job after all :P

unrelated, guess who hasn't updated his pycharm IDE in over a year xD

@ThomasWard :)

a positive vote would be awesome too :)

please don't ask for upvotes

5:12 PM

0

Q: Write an interpreter for a Turing complete ZISC

A ZISC is a hypothetical computer where the only instruction in that language has no parameters. Therefore, the output is entirely dependent on the input since there are no "programs" in a ZISC. Therefore, your goal is to design a Turing-complete ZISC and code an interpreter for said language wit...

code-golf code-challenge

(and especially on questions which don't give all the necessary details and ask too many things at once)

@Anush yeah, asking for upvotes is... not cool

and i agree that you're asking a bunch of things at once in your question and it's slightly unclear what your goal is here

@ThomasWard what about asking for downvotes?

@AlanBagel i mean

you can

but it just hurts you as a user overall :P

ive seen it here and its cringe anyway

5:14 PM

@ThomasWard then how about asking for flags or close votes

@ThomasWard point taken

@pxeger bpa.st/Y2XA aggregates by day which is a good start

@AlanBagel sounds like a good way to get banned tbh

'cause you'd clog up the mods' queues

@pxeger makes sense

@AlanBagel we have a special tool for that. watches the orbital weapons array fire at you

I hate the fact that you can't define functions in a loop without the functions closing over the loop variables

And I also hate the fact that generator comprehensions implicitly define functions which have closures

But you don't even notice that's a problem until you start persisting those generator comprehensions outside the loop

5:23 PM

@pxeger yeah it is annoying sometimes (although i never ran into that problem)

@pxeger Sounds like more things to "fix" in Whython :P

5:51 PM

@Anush I have a very dirty solution, but I kind of expanded so that i have a TextRecord class for object oriented programming. I have to adapt it for your needs, but there's a lot of loops in this, so it's not nice in terms of O-notation

0

Q: An Scalable Event Scheduler That Can Handle Billions of Events

The event scheduler is a recurrent theme in event-driven architectures: something triggers an event that needs to be checked in the future one or more times. I has particular use in trading where you want to correlate one market event with price movements in the future at several time deltas (1ms...

sorting fastest-code restricted-complexity data-structures c++

but since you're trying to do multi-faceted aggregation you don't really have an O(n) solution here since you need to first group by user ID

but i digress this isn't a CGCC question ;)

@cairdcoinheringaahing ... Is it bad that I don't know what the better solution is to most of these problems? Evidently my JavaScript knowledge hasn't progressed much past 2008 either. :P

@NewPosts slightly sad this got closed

like, objectively, yeah

but i kinda wanted to try it haha

6:06 PM

I think it might be a duplicate of this--they're not exactly the same, but the ZISC challenge is a proper subset of the general TC language challenge, and I think the shortest answers to the TC challenge can be seen as ZISCs. Thoughts?

6:20 PM

oh ok, thanks :) i think youre probably right, but more importantly to me, i can now go try that challenge :D

Redwolf Programs

Ooh stuff with infinite series is fun

Aaroneous Miller

I can't believe GitHub accepts this as a table: gist.github.com/AMiller42/5103949447b54f0af620d89c2febb999

Redwolf Programs

Like how 1 / 2 + 1 / 4 + 1 / 8 + ... is the same as 1 / 2 + (1 / 2) * (1 / 2 + (1 / 2) * (1 / 2 + (1 / 2) * ..., so you can change that to x = 1 / 2 + (1 / 2) * x, and then solve it

Aaroneous Miller

@AaroneousMiller Raw markdown: gist.githubusercontent.com/AMiller42/…

Redwolf Programs

Indeed smol, I didn't know the |s on the sides were optional

Aaroneous Miller

6:25 PM

me neither

SE Markdown accepts this for tables:

x|
-
a|
b|
c|

@RedwolfPrograms I see you're one of today's lucky 10,000 :D

Redwolf Programs

I had to figure that out all on my own lol

But with a more complicated, and thus more fun, one :p

6

Q: Why do the humans become sleepy after meals?

I don't know about all the mankind, but I know enough people, who becomes sleepy after their meals. Also, I'm not sure, what kind of food do they consume, but I personally get sleepy almost from any food: sweet, fat, spicy, salty, liquid, tasty, not tasty, etc. Is this phenomenon well known? And...

human-biology biochemistry physiology nutrition sleep

How to sound like a robot :p

@RedwolfPrograms I am super sleepy after lunch.

@RedwolfPrograms Well, they claim to consume food, so maybe they're not a robot, just some other sentient life form – def. not human though.

Redwolf Programs

@AlanBagel You must be one of those "humans" then

6:39 PM

@RedwolfPrograms I am one of those "humans" then (robot style)

@PyGamer0 Only with ones that are links to Jelly

@NewPosts This should probably be vtc's as a general programming question, not needing details or clarity

Redwolf Programs

It's fastest code, right?

It just needs details

An "event scheduler" is not something you can implement in 90% of languages.

why does that matter

@hyper-neutrino because

6:51 PM

challenges don't need to cater to every language they just need to not impose unnecessary restrictions that add nothing

I reread it and it really looks like "Help, we need fast code to do this, plz bring da codes"

nah dude, it really doesn't - OP specifically provides sample code that's a (n unoptimized) solution

> Help, we need fast code to do this, plz bring da codes

if it's up to our standard as a challenge, i don't think we care about that

Redwolf Programs

^

6:55 PM

"We also don't care about what language it's in"

caird coinheringaahing

@emanresuA this is absolutely a challenge. one way you can tell: it was Sandboxed

4

oh also, i checked out op's github and they have their website linked & all of that lmao. including their work history at places like JP morgan for low-latency code

caird coinheringaahing

And, just for reference, not every language must be able to compete in every challenge

i remember clicking on it yesterday when i saw the cool sandbox post but couldn't find it again immediately

« first day (3965 days earlier) ← previous day next day → last day (881 days later) »

Transcript for

Dec '217

The Nineteenth Byte

The Nineteenth Byte: General discussion for codegolf.stackexc...

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile