The Nineteenth Byte

user

19:00

@ngn Pretty sure destructuring assignment's an example of pattern matching, but that'd be something like [1, 2 | A] = ... instead (not totally sure, though)

Peilonrayz

@ngn Python's destructuring assignment only works on iterables, and can't have defaults. Python's now (last I checked like a year ago) added a dunder method to allow pattern matching from most objects.

ngn

@Peilonrayz so, when it's with an iterable, it's called destructuring assignment, and when it's with an arbitrary object, it's called pattern matching?

user

@ngn When it's an assignment where you look at the structure of an iterable or an object, it's destructuring assignment

Pattern matching can also be a switch-style thing where you have multiple branches, like if + assignment instead of just assignment

Peilonrayz

@ngn Python's analogue to destructuring assignment is *args. But I don't think it has a name... or at least I forget it every time I learn it.

user

@Peilonrayz Python can't do destructuring assignment on objects?

Peilonrayz

19:13

No

user

Oh, that's weird

Peilonrayz

Yeah. Every time I use JS I always leave thinking "it'd be nice if Python had {foo, bar} = obj"

hyper-neutrino

@Peilonrayz splat?

i think that's what it's called under some contexts

Anush

what does (~0 << 3) mean in C?

caird coinheringaahing

-1 << 3 = -1 * 2 ^ 3 = -8

I think. I can never remember the precedence of <<

hyper-neutrino

19:17

unary always takes precedence i believe

user

~ takes precedence over << (which is pretty low on the list anyway)

Peilonrayz

@hyper-neutrino Splat comes back with unpacking, which AFAIK is different to destructuring assignment. head, *tail = value vs values = head, *tail.

hyper-neutrino

@Peilonrayz ah, never mind then. i've just seen the term thrown around when it comes to functions like lambda *x: ... and f(a, *b)

Anush

@cairdcoinheringaahing this is in code optimized for speed. Why would you do that?

hyper-neutrino

:58266186 tio.run/##Sy4o0E1PTv7/XzkzLzmnNCVVwaa4JCUzXy/…

user

19:18

@cairdcoinheringaahing It's -8

@hyper-neutrino lol imagine using good formatting

hyper-neutrino

@Anush context?

Anush

@hyper-neutrino github.com/AwardOfSky/Fast-Radix-Sort/blob/master/radix.c

hyper-neutrino

@user better?

user

@hyper-neutrino Nah, get rid of those spaces :P

Anush

I thought shifting a negative sign value was undefined

Peilonrayz

19:20

@hyper-neutrino The joys of every language seeming to having different terminologies ;)

user

In my language, "splat" is when a pigeon poops on your head

hyper-neutrino

@Anush why would it be? you just need to move the bits over

0 is 00000000000000000000000000000000

Anush

@hyper-neutrino it's what gcc says :)

user

@Anush Shifting by a negative sign value probably is

hyper-neutrino

so ~0 is 11111111111111111111111111111111 or -1

caird coinheringaahing

19:21

@Anush Shifting by a negative value is undefined IIRC

hyper-neutrino

so ~0 << 3 is 11111111111111111111111111111000

ngn

@Peilonrayz try this :) locals().update({"foo":123,"bar":45});print(foo)

Anush

why would it be in this code? I mean why not just write -8?

hyper-neutrino

which is -2147483648 + 2147483640 which is -8

@Anush probably clarity

user

@hyper-neutrino ???

Peilonrayz

19:22

@ngn Are you trying to trigger me, being a CR mod and all ;)

hyper-neutrino

it's a bitmask

Anush

@hyper-neutrino how is it clearer?

the whole line is confusing

for(m = (unsigned char *)(&vector[size & (~0 << 3)]); n < m;) {

hyper-neutrino

@Peilonrayz being a CR mod in this room is a risky endeavor ;)

@Anush true

Anush

can anyone translate it?

user

@Peilonrayz Here, get triggered >:)

@Anush Guess it's to make them look smarter :P

Anush

19:24

@user :)

@hyper-neutrino oh!

hyper-neutrino

size & (~0 << 3) is just "truncate the size to a multiple of 8"

user

@Anush That was an odd edit

Anush

I replied before your message by mistake

:)

size & -8 truncates to a multiple of 8....?

doesn't size * size%8 do the same?

user

@hyper-neutrino That's smart (and actually looks clearer now)

@Anush % and * isn't fast enough maybe?

Anush

ah maybe

user

19:26

Shifting bits is probably trivial

Anush

yes

I get 62 warnings from clang :)

mostly warning: shifting a negative signed value is undefined [-Wshift-negative-value]

but it is very fast

hyper-neutrino

@Anush wait what how would that work

caird coinheringaahing

Doesn't C optimise bitshifts? Maybe it actually is faster to do ~0 << 3 than use negative numbers?

hyper-neutrino

also (size >> 3) << 3 would be clearer IMO

user

@cairdcoinheringaahing Probably won't be faster, but it'll be done at compile time

hyper-neutrino

19:28

@Anush yes because -8 is 11111111111111111111111111111000

and so & -8 will just set the last three bits to 0

thus truncating to a multiple of 0b1000 or 8

Anush

@hyper-neutrino is it wrong?8*(63%8) = 56

ah I made a typo above

user

@Anush I think what HN meant is that it rounds to a multiple of 8 by clearing the last 3 bits

Anush

I mean 8*size%8

@user thanks

hyper-neutrino

okay and 8*(62%8) isn't 56

Anush

@hyper-neutrino oops

user

19:30

Did you mean 62 - 62%8?

hyper-neutrino

(x/y)*y works (assuming int division) in general

^^

Anush

8*(62//8) in python

@user yes.. sorry all

hyper-neutrino

if you reverse the multiplication you can avoid brackets

Anush

@Peilonrayz I meant size - size%8

thanks for the deciphering all

hyper-neutrino

but again & -8 is way faster lol. because bit operations are really quick

multiplication takes 4 ops i think?

division takes like 32 or 60? i don't know details

idk if the compiler or CPU will optimize division by a power of 2

Anush

19:34

agner.org/optimize/instruction_tables.pdf

that code is designed to avoid the compiler needing to do any optimization

user

@Anush May I ask why?

Anush

it makes no difference if you use -Ofast or no optimization it seems

user

Isn't that the compiler's job?

Anush

@user I don't know. Maybe they wanted to make sure it was equally fast on all compilers?

github.com/AwardOfSky/Fast-Radix-Sort is all I know about it

user

¯\_(ツ)_/¯

Anush

19:35

it is about 10 times faster than qsort

maybe they just didn't think the compiler could optimize as well as they could

I am amazed someone who makes that many typos in their English can write correct code!

user

English and C are completely different languages

Anush

yes but sloppiness is the same in all languages :)

user

@Anush IDEs catch typos in C, but not English (some do, but somehow I feel this person uses butterflies to code)

Anush

20:13

@user that's a great expression :)

hyper-neutrino

i believe it's a reference to

xkcd.com/378 (onebox is too large)

user

Why does SE chat even bother oneboxing xkcd? There's nothing special about it

hyper-neutrino

because there is always a relevant xkcd (disclaimer: not an xkcd)

(if you visit that page, click the link in the comic panel. it's cool)

user

AAHHHHH RECURSION

Adám

@hyper-neutrino That's brilliant. Also, why didn't Randall draw such a cartoon?

user

20:23

@Adám Too busy doing this, I guess

hyper-neutrino

I've been wondering too. With how prominent the meme is (or seems to be) having a relevant xkcd for "relevant xkcd" would be absolutely great

Adám

People regularly submit questions to him. Hint hint.

user

Has xkcd ever been mentioned in an xkcd?

caird coinheringaahing

I don't think so

user

Well maybe there was an xkcd about a relevant xkcd, but it's on meta.xkcd.com :P

I wish I could express thoughts directly without resorting to such imprecise methods as typing with my stupid fingers in this stupid language

caird coinheringaahing

20:32

just use Jelly's "this is what I mean" builtin

user

I half suspect Mathematica has a ReadMindAndTransmitThoughtsDirectlyToSomeoneElsesMind builtin, but it's just that no one's found its true name yet :P

hyper-neutrino

Its true name keeps changing; it's always what thoughts you are trying to transmit written in whatever language or medium you are trying to use.

Anush

@hyper-neutrino that's great!

Adám

@user YES! I found one mention.

caird coinheringaahing

@hyper-neutrino Are we sure it isn't some kind of fae or demon?

No one eat any food offered to you by Mathematica, you might be trapped in the world of Stephen Wolfram

Anush

20:36

@hyper-neutrino in think Seinfeld was like that to a lot of older people too

hyper-neutrino

It's too late, the TrapCairdCoinheringaahingIfTheyDiscoverStephenWolframsSecretPlanInStephenWolframsWorld builtin is already implemented

Adám

Appears in xkcd.com/1416

user

@Adám Interesting

I love the interactive ones, Monroe seems to put a lot of work into those

caird coinheringaahing

I do love it when Jelly absolutely demolishes 05AB1E :P 53 vs 36 bytes

hyper-neutrino

apparently chat feeds pull from meta half as often as main, which is why they are noticeably slower. TIL... presumably some of y'all already knew? i didn't :P

rak1507

20:56

@Wezl codegolf.codidact.com/posts/281956#comment-10576 sorry only just saw this, go ahead

rak1507

21:09

I am once again asking for solutions on my new challenge codegolf.codidact.com/posts/282075

@hyper-neutrino that was fast!

hyper-neutrino

i am speed

New Posts

0

Q: Evaluation order of an APL n-train

From Codidact with permission. Description APL trains are a series of functions, that get applied to an argument in this way: (f g) x = f g x (f g h) x = (f x) g (h x) (a b c d e f) x = (a (b c (d e f))) x = a (b x) c (d x) e (f x) Trains evaluate from the right to the left, so in the last examp...

caird coinheringaahing

@NewPosts I'd steal HN's Codidact answer, but that's cheap :P

rak1507

hey that looks familiar :P

Adám

I am Lobachevski.

caird coinheringaahing

21:17

Out of interest @rak1507, why didn't you post it?

rak1507

because I'm lazy and Adám said he'd do it :P

hyper-neutrino

you already wrote up the post, it would've taken you like 20 seconds :p

rak1507

lol, true

caird coinheringaahing

@Adám Lobachevski?

Adám

@cairdcoinheringaahing Yes.

caird coinheringaahing

21:20

@Adám Ah, Lobachevski

Adám

Yes.

caird coinheringaahing

Any feedback?

Redwolf Programs

I think I'll do some more work on my chatbot library

I'm going to make NP/SP use it eventually

caird coinheringaahing

Imagine needing a diamond to edit mod tags. Made by the Ryan Donovan gang :P

Redwolf Programs

Oh nice, a new user's started posting lots of answers so I can get Late Answers reviews :p

hyper-neutrino

21:23

a staff without a diamond? o.O

caird coinheringaahing

They usually only give staff a diamond if it's likely they'll need it (e.g. the CMs, devs and higher ups)

hyper-neutrino

ah, I see

caird coinheringaahing

Besides, they can usually do stuff directly to the database if they want to :P

hyper-neutrino

true

Redwolf Programs

I'd be absolutely terrified I'd forget part of the UPDATE and ruin everything

caird coinheringaahing

21:25

What can non-diamond staff do?

hyper-neutrino

docs.microsoft.com/en-us/sql/t-sql/language-elements/…

that's what BEGIN TRANSACTION is for :)

Redwolf Programs

I have a habit of managing to destroy production code, development and testing copies, and backups with a single mistake, so I'm sure my superpowers could evolve to find workarounds for transactions too :p

hyper-neutrino

or you could pull a tom scott and just forget to use BEGIN TRANSACTION and destroy a ton of prod data :)

caird coinheringaahing

I love that Tom Scott could be 25 or 45 and I'd have no idea

Redwolf Programs

16 or 60 is a more accurate range honestly

caird coinheringaahing

21:30

> When working on a database, you don't work on the live version. So I was working on the live version, ...

Redwolf Programs

You know how sometimes golfing languages can be painful to use? Well I decided to make one. I got the pain part right, but I might have added too much of that to compensate for reduced golfiness. Can't share any details yet, but it should be released within a day.

As far as I can tell it includes a major feature never before seen in a golfing language.

(Or any languages (that I know of))

caird coinheringaahing

Is it a golfy feature?

rak1507

exciting

Adám

How do you keep a code golfer in suspense?

caird coinheringaahing

21:45

Hmm, there are ~28716 bugs/FRs on MM with no [status] tag

Sounds like a fun cleanup for the staff :P

Wezl

@Adám Hw do u ⌸p a cdglfr∊suspense? ftfy

further golfing can be accomplished through string compression

caird coinheringaahing

This looks like a minor issue that isn't impacting how things are working which is why we status-declined this. However, if there are more instances of users being awarded the same badge twice please let us know and we may revisit if this is a larger bug. — Rosie ♦ 4 hours ago

Idk, letting lyxal have two silver code golf badges severely affects his ability to brag about having two silver code golf badges :P

Redwolf Programs

The condition that caused that to happen is known, right?

caird coinheringaahing

I guess so, otherwise it'd be [status-reproduced]?

Redwolf Programs

Probably has something to do with one of the conditions getting fulfilled, then unfulfilled, then fulfilled again

Adám

21:49

So status-wtf?

Redwolf Programs

Since badges can't get revoked

caird coinheringaahing

Tag badges can

Redwolf Programs

Oh they can?

caird coinheringaahing

Yep

If you delete a bunch of answers, or get a heavily downvoted one, you can lose your tag badge

caird coinheringaahing

22:29

Apparently, I had this answer in my SE app drafts for the past month (converting my comment to an answer) and only realised just now :/

Neil

@hyper-neutrino well, (set 'a b) is syntactic sugar for (set (quote a) b), where quote is what the q of setq stands for. setq apparently also takes an even number of arguments, while set always takes two; I don't know why that is.

@hyper-neutrino did you mean the link in the 4th panel of the comic, the link in the 4th panel of the comic in the 2nd panel of the comic, or the link in the 4th panel of the comic in the 2nd panel of the comic in the 2nd panel of the comic?

@hyper-neutrino I think the first client-server database that I used turned transactions on my default, so if you typed "delete from foo;" then all that would happen is that everyone's application would wait for you to ~~commit~~ rollback. I don't know why that setting isn't more popular.

caird coinheringaahing

22:49

For anyone who wants to play it: Unikong - SE themed Donkey Kong style game from 2016's April Fools

Wezl

@Neil (setq a 1 b 2)=(set 'a 1)(set 'b 2) so in the quoting and multiple assignment cases, it's a feature set' adds

and set is the lower-level work-doer

however picolisp allows direct access to the dynamic environment, so set may actually be a function implemented in terms of replacd

lyxal

@user this is an epic victory royale

Wezl

@Neil btw 'a is the same as (quote . a) in picolisp

hyper-neutrino

@Neil oh, okay, that makes more sense. from mathematica's naming Q seems like a predicate lol

@Neil yeah that sounds like... a very wise choice

Wezl

in lisp (not scheme), predicates usually end with a p (listp, evenp), however picolisp chooses to have no suffix (atom, pair, not) or have a ? (num?, sym?, flg?)

caird coinheringaahing

23:01

2 hours ago, by caird coinheringaahing

Any feedback?

hyper-neutrino

i think ? is what racket uses so that makes enough sense to me

lyxal

Picolisp? What inspired your choice to use that language?

caird coinheringaahing

I'd like to post that soon, feedback would be appreciated :)

Wezl

@ hyper-neutrino same

Adám

@Downgoat Wow, welcome back!

Downgoat

23:04

@Wezl O.o huh how in tarnation did this ping me

hello!

Wezl

oh sorry for that

Downgoat

haha all good, just curious

caird coinheringaahing

Hey Downgoat, been a while!

How's things?

Wezl

in Sandbox, 7 hours ago, by Wezl

@person-it-looks-like-i'm-pinging I present to you...

...A new generation of stealth pings!

caird coinheringaahing

Oh please no

Wezl

23:06

only if you get on my bad side :P

Redwolf Programs

If they won't fix what's practically an XSS, they probably won't fix that :/

Wezl

> fix

caird coinheringaahing

Hang on, let me rephrase that: Don't do that. Stealth pings got banned for a reason. Pinging can be disruptive, and you should only use it to ping people you're trying to talk to

Redwolf Programs

It's worse than a stealth ping because it works regardless of the time delay

caird coinheringaahing

If you want to "double ping" someone by reply and @ing them, go ahead. But don't ping someone who isn't a part of the conversation

Wezl

23:09

@Wezl it's probably status-bydesign because placing the name of the user messes up alignment, example:

@Wezl .--.
'--'

Redwolf Programs

I can't find the exact message, but the policy here will stay that stealth pings result in a kick-mute.

gtg for now o/

Wezl

.--.
'--'

Adám

@Wezl Except of course that it doesn't apply when alignment is most important; multi-line code blocks.

\○/

Wezl

and people probably shouldn't be replying with ascii-art usually anyway

Adám

No, but I've often wanted to reply with a code block as an answer to a "how do I" question in chat.

Instead, I end up writing
@user Like this:

And then the block.

caird coinheringaahing

23:12

@Downgoat Jelly, 6 bytes: ḢḤ;ṖḅƓ

Downgoat

@cairdcoinheringaahing good! I'm at work now, so I suppose I shouldn't be here :P but I see a lot has changed

caird coinheringaahing

Given that you're around, I figure the old ping won't be too disruptive :P

@Downgoat Yeah, not all for good unfortunately :(

Downgoat

Now working at the Swift team, and w/ WWDC here, it's a big week!

@cairdcoinheringaahing did the community ever migrate to another platform? Axtell, was never finished, but I recall other people working on other sites

caird coinheringaahing

A few people moved to Codidact, a few to Top Answers

Wezl

codidact has seen some activity

caird coinheringaahing

23:15

Codidact really took off after Monica's firing (actually, wasn't it created because of that?)

@cairdcoinheringaahing I am idiot

Jelly, 5 bytes Ḥ1¦Ṗḅ

Adám

Top Answers Code Golf is basically dead.

Downgoat

@Wezl woah, it looks pretty well fleshed out

caird coinheringaahing

It's pretty good tbh. The biggest issue is activity - it gets maybe 1 question every few days, if that

Plus, I can't speak for others, but I'm definitely "sunken-cost fallacy"-ed for CGCC

@hyper-neutrino the chaining in this answer really confuses me. Why does ) not create a new monadic chain so that it's parsed as (×BS)€(ċ>1)# instead of (×BS)ċ>1)#?

hyper-neutrino

aw, did something happen while I was at dinner again :P

hey there Downgoat :D

@cairdcoinheringaahing I believe µ creates a new chain and groups the old chain into a subchain which if you apply a quick to, loses the "new chain" meaning

don't quote me on that

user

@Adám Not sure if you meant to ping me there

Unrelated String

23:29

yeah I think the quick subsumes the chain into the next chain

hyper-neutrino

but how I see it is ((×BS -- €)ċ>1)#

Adám

@user That's what you get for having such a user name ;-)

user

>(

Unrelated String

like I'm pretty sure the way applying quicks after separators works is if the current chain is empty it just substitutes the last chain

hyper-neutrino

yeah that's how i believe it works

caird coinheringaahing

23:30

Yeah, it's got a (chain or chains) condition in the parsing

Unrelated String

yeah the exact line is something like chain.pop() or chains.pop() except I think popping from an empty list in Python is an error so it's probably more like chain.pop() if chain else chains.pop()

caird coinheringaahing

Interestingly, I think µ doesn't actually create a new monadic chain, it creates a new variadic chain when followed by a quick: Try it online!

Unrelated String

:|

caird coinheringaahing

But I could be wrong, chains and quicks are weird

Shouldn't ^ be [[[1, [1, 2, 3]], [2, [1, 2, 3]], [3, [1, 2, 3]]]?

Unrelated String

speaking of ) i'm trying to golf hyper's train evaluation order solution and it may or may not be biting me in the ass that ) just isn't allowed at the start of a line

caird coinheringaahing

23:33

That's what ,¶ç€ does anyway

Unrelated String

i say may or may not because i'm not sure it would actually work either way

caird coinheringaahing

@UnrelatedString Why would you start a line with )?

Unrelated String

as some kind of jank substitute for Rµ

except for also the whole

subsuming it into the next chain bit

caird coinheringaahing

I don't even think that would work as ) retains the left argument, so the bit after the ) would be dyadic

hyper-neutrino

@cairdcoinheringaahing i don't see anything wrong with that

it could be starting a new monadic chain

it's just not ending one

oh. wait. we might have different interpretations as to what "create" a new monadic chain means

Unrelated String

23:35

yeah come to think of it the "retains the left argument" caveat is also probably just a weird way of explaining how it doesn't really act like a separator to begin with

caird coinheringaahing

Speaking of golfing hyper's answer: -1 byte

Unrelated String

nice

hyper-neutrino

does it end the current chain and create it as a monad, or does it end the current chain variadically and create a new one being a monad?

Unrelated String

...of course

hyper-neutrino

@cairdcoinheringaahing thank

caird coinheringaahing

23:36

If you ever have R and increment/decrement next to each other, there's almost always a one byte version :P

Unrelated String

R in general is usually suspect

Bubbler

@hyper-neutrino Actually the comic has three clickable regions

caird coinheringaahing

7 bytes, assuming 1 can output as 1 rather than [1]

Unrelated String

nice

it's always where it's useless that I try to do weird stuff with replacing normal stuff with ƒ

Adám

@Bubbler Oh wow, you're right.

Unrelated String

23:45

I've just been trying extreme rearrangements and seeing if I've done the accounting wrong in my head and they turn out shorter

caird coinheringaahing

5 bytes porting xash's J answer :(

Unrelated String

ofc

caird coinheringaahing

I hate it when a port is shorter than an original, clever answer

(cc @hyper-neutrino - some golfs for you)

hyper-neutrino

oh :/ well i'll probably just edit both in, thanks

will do it in a bit

caird coinheringaahing

It annoys me to no end that Ṛ is reverse digits on integers, but U is a no-op, rather than reverse range ಠ_ಠ

Unrelated String

23:53

the classic

caird coinheringaahing

If it were, 4 bytes in Jelly :/

Adám

Uh, FU Jelly, for not being golfier.

hyper-neutrino

i wish jelly had grade down

user

@Adám Hey, at least this way other languages have a small chance of beating it :P

Adám

@hyper-neutrino Adding grade down would be such an upgrade, wouldn't it?

hyper-neutrino

23:56

reverse+grade up doesn't work on non unique arrays right?

@Adám lol, good one

caird coinheringaahing

Also, for some reason, when reducing n-wise with ƒ, the initial value is prepended to each list (in this example with ;): tio.run/##y0rNyan8/z/I2ujYpP@GBkGH2x81rTk66eHOGUA68j8A

Adám

@hyper-neutrino No. Try it online!

user

@hyper-neutrino Negate and grade up maybe?

hyper-neutrino

oh.

no still won't work

user

:( It's a common trick in Scala

@hyper-neutrino It seems to wok here

*work (not editing that because pings)

Transcript for