« first day (1463 days earlier)      last day (1181 days later) » 
00:00 - 14:0015:00 - 23:00

12:14 AM
Wait why is the pfp for the dyalog sales email address a duck?
 
I remember those ducks in a previous competition
 
there's an entire bookshelf of ducks at the office.
There's a solar powered dancing duck by the door, and the office kitchen dish brush is a duck too.
 
Duckalog
 
 
 
4 hours later…
4:27 AM
 
wow
"Exciting life of off-by-one errors" 😝
 
5:27 AM
So f⍣g runs f till g is truthy
and f*num runs n times
how do I use ⍣ with both?
aka: run till g is truthy or otherwise, n times
I can think of a counter variable, but just wondering if there's a shorter way
 
You could iterate over counter,value
or (counter value)
then increment the counter in f and test the counter in g
 
hmm
that sounds more complicated
 
Could be actually shorter using an external counter variable
 
Let's see: {(f⍵),1+⊃⌽⍵}⍣{((⊃⍺)≡⊂⍵)∨⊃⌽⍵=n}
or {i←1⋄f⍣{(⍺≡⍵)∨(i+←1)⊣i=n}}
 
5:45 AM
@Razetime Slightly more correct: {i←0⋄f⍣{(⍺≡⍵)∨i=n⊣i+←1}⍵}
 
oh yep
 
Or {i←0⋄f⍣(≡∨{i=n⊣i+←1})⍵}
or {i←0⋄f⍣(≡∨{n=i⊢←i+1})⍵}
Not sure if I can remove from ⊢←
 
Only way to know is to try
 
If you want to take n as , {i←0⋄f⍣(≡∨⍺={i⊢←i+1})⍵} should do
 
interesting
idk if I can do that
maybe I can bind it as ⍺
but n is constant here so I won't need that approach
 
6:07 AM
@Razetime Does g need to compare two "generations" or is it a simple condition on the argument(s?)?
 
two generations
 
Does f have side effects?
 
side effects, meaning?
f doesn't modify anything outside of it's argument
 
OK.
Is g commutative?
 
lemme google that
 
6:12 AM
commutative means ⍺ g ⍵ is always equal to ⍵ g ⍺
and I think so, given that g is simply there
 
no, it's {~⊃⊃⌽⊖⍺}
 
Then no
 
OK.
 
what if g was commutative?
 
I'd save a
 
6:14 AM
interesting
 
@Razetime I think {f⍣(⍵g⍨f⍵)⊢⍵}⍣n or f{⍺⍺⍣(⍵g⍨⍺⍺⍵)⊢⍵}⍣n works.
 
this'd be a cool APLCart snippet
 
Yeah, but can you test it first?
 
sure
      a←2 3 ⍴ 1 1 1 1 0 0⋄f←⌽∘⍉⋄g←{~⊃⊃⌽⊖⍺}⋄n←4⋄({f⍣(⍵g⍨f⍵)⊢⍵}⍣n)a
1 1 1
1 0 0
      a←2 3 ⍴0 0 1 1 1 1⋄f←⌽∘⍉⋄g←{~⊃⊃⌽⊖⍺}⋄n←4⋄({f⍣(⍵g⍨f⍵)⊢⍵}⍣n)a
0 0 1
1 1 1
seems to return it unaltered?
 
Oh, my bad, there's a missing ~
Should be {f⍣(~⍵g⍨f⍵)⊢⍵}⍣n
 
6:25 AM
      a←2 3 ⍴0 0 1 1 1 1⋄f←⌽∘⍉⋄g←{~⊃⌽⊖⍺}⋄n←4⋄({f⍣(~⍵g⍨f⍵)⊢⍵}⍣n)a
1 0
1 0
1 1
it should return an iteration where 0 is at the bottom right corner
 
Ah, I see the problem. The left argument of g becomes the next generation, not the current one. The right argument is the current one.
     a←2 3 ⍴0 0 1 1 1 1⋄f←⌽∘⍉⋄g←{~⊃⌽⊖⍵}⋄n←4⋄({f⍣(~⍵g⍨f⍵)⊢⍵}⍣n)a
1 1 1
1 0 0
@Razetime I think it is too involved like this, and a simple "f until g or at most n times" would suffice.
 
yeah, there's two uses of f as well
 
That makes it more complicated too, as it'd have to be an operator using an operator.
I f{⍺⍺{⍺⍺⍣(~⍵⍵ ⍵)⊢⍵}⍵⍵⍣⍺⊢⍵}g Y f applied repeatedly on Y until g Y or at most I times, whichever comes first
Or remove the ~ and replace "until" with "while".
This last case would be F⍟G⍟n in BQN, if I'm not mistaken.
     4 f{⍺⍺{⍺⍺⍣(⍵⍵ ⍵)⊢⍵}⍵⍵⍣⍺⊢⍵}{⊃⌽⊖⍵}a
1 1 1
1 0 0
 
6:58 AM
@Razetime It is actually missing one more application of f because you're used to the sometimes annoying aspect of that it always applies f at least once:
 
Nice work, though porting Jonah's solution should work much better (codegolf-wise)
 
@EliasMårtenson amazing
 
@Bubbler yep lol
I'll get on that after a little while
 
I guess half of the art of APL golfing is in finding an algorithm that avoids
 
7:03 AM
maybe it is
I need ot peek at J documentation
I need to see how Jonah handles padded matrices
 
@KamilaSzewczyk Glad to be of service :-)
 
@Razetime I've not tested it yet, but I have a suspicion (⊢≡∘,⊣/+∘⍳1+⊃∘⌽-⊃)⍸ works.
 
I'll put it through my testcase runner then
 
You need to represent "empty space" as 1, filled space as 0.
 
7:07 AM
Oh right, ⎕IO←0 ;-)
 
Try it online! Third testcase fails, and another one causes domain error
 
@Razetime I don't see it. The 3rd test case fails?
##
####
It gives 1 for that. Isn't that correct?
 
the two branches need to have the same thickness
(That is, if I read it correctly)
 
@Razetime You read it correctly
 
Ah, I missed that.
OK, but a square pivot simply means that the empty space is smaller than the full area by the same amount horizontally and vertically…
 
7:14 AM
given that the matrix is cropped to fit the filled area
 
It doesn't say, but I think we can assume that.
 
well, it is indirectly stated
 
> You should handle inputs containing or not containing zero margins
 
For example, the output must be True if the input is this:

0 0 1 1 1 1 0
0 0 1 1 1 1 0
0 0 1 1 0 0 0
 
@Bubbler Oh. Too early for me to read, clearly.
 
7:16 AM
Well, it happens
 
Also:
> the lengths of the two legs should also be at least 1.
But even with stripping padding and adding additional tests, it might be shorter.
@Razetime @Bubbler Will {1∊⍵:(=/⍵-⍥⍴e)∧i≡,e←(⊣/+∘⍳1+⊃∘⌽-⊃)i←⍸⍵ ⋄ 0}{⍉⌽⍵⌿⍨~∧\∧/⍵}⍣4 work?
 
:56877658
domain error on 7th one
but it looks correct
 
Ah, because get's a negative arg.
 
7:34 AM
@Adám I get domain error for g ' ' ' ## ' ' ## ' ' '
 
cool!
 
Now, can it be golfed?
 
yep, looks like it
Here's my(likely to be wrong) translation of jonah's answer Try it online!
 
7:41 AM
oh, deb works on integer arrays.
 
@Razetime Isn't Jonah's answer wrong?
 
oh, it is?
well then
I just wanted to try figuring out a J train
 
@Razetime Wait, only if given a left argument of what to strip, so I'll have to add 2 bytes.
 
hm, how's it working without the argument?
is there a default value for ⍺?
 
Maybe it's stripping the prototype? (too lazy to open the doc)
 
7:45 AM
@Razetime Yeah, space. I think it should have been the prototype.
@Razetime Want 40-byter (Extended)?
 
uh, sure?
 
 
Oh, I thought you were talking about the l shape
@Adám no used, heresy
 
:-D
 
is under, right?
 
7:51 AM
Yes.
iirc '-'⍪⍢(1∘⊖) doesn't actually save anything, but it is a cool trick.
'-'⍪⍢(1∘⊖)
⊖@(⍳2)'-'⍪
The approach is more general, though. (Makes sense; is a generalised @)
 
is there any place I can read up on ?
 
Hmm, APL wiki doesn't have an article yet
 
I couldn't find one on the bqn page either
 
Where I was looking too…
 
dzaima implemented it pretty well in dzaima/APL, and we had quite heavy discussion on it here
though I doubt how searchable the chat log is
 
8:02 AM
The bqn website is great for understanding modifiers like over and under
 
I really thought BQN had a thorough discussion of it.
 
i don't understand why and when i'd use atop and over instead of bind
i tried to read the apl orchard discussion, but it still didn't make much sense
 
tl;dr: It only makes a difference when the derived function is used dyadically.
@KamilaSzewczyk btw, "beside", not "bind".
 
I think the use cases were well covered in a dyalog.tv video
 
it says beside and bind
no clue
 
8:06 AM
Bind is when used to bind or curry an argument to a dyadic function, deriving a monadic function. E.g. Increment ← 1∘+ and Decrement ← -∘1
 
f∘g is beside (composition of two functions), x∘f and f∘y are bind (currying one side of a dyadic function)
 
the only description of under I can find is here: mlochbaum.github.io/BQN/doc/primitive.html
I'm gonna try the explai feature and see what it does
 
8:22 AM
@Razetime There's this but notice that he uses ¨ to mean a hybrid between and and .
 
I'll read it
is this the correct equivalent code?
 
@Razetime Yes, although it doesn't allow rotating a scalar. Try this!
Oh no, that's not it, it doesn't extend the scalar to join a scalar as a row to a matrix.
 
Still waiting for AST....
this explain feature takes a while
 
Yikes, yes.
 
mr Mortal Kombat is here
 
8:32 AM
@Bubbler We have an internal Dyalog wiki that discusses it.
@MortenKromberg Can I borrow material from there for APL Wiki?
 
⋄ 2 +⍤+ ¯4 ⋄ 2 +∘+ ¯4
 
@rak1507
¯2
¯2
 
I thought they did different things! :P
 
They do, the first is +2+¯4 and the second is 2++¯4 but with these precise arguments, the results are identical.
 
⋄ 2 ÷⍤+ ¯4 ⋄ 2 ÷∘+ ¯4
 
8:42 AM
@rak1507
¯0.5
¯0.5
 
these are the same too!
I'm not buying it
 
⋄ 2j3 +⍤+ ¯4j5 ⋄ 2j3 +∘+ ¯4j5
 
@Adám
¯2J¯8
¯2J¯2
 
@rak1507 +2÷¯4 and 2÷+¯4 lmao
 
yeah I'm just kidding
 
9:02 AM
⋄ 2+2 ⋄ 2×2 ⋄ 2*2
 
@RikedyP
4
4
4
 
I am confusion
 
lol
 
9:28 AM
@RikedyP Like the epitome?
 
9:56 AM
@ngn @Bubbler You folks are doing Project Euler? Friends key: 52715_d337e9574c3aea0ad76fd977063b16cc (I mainly did this stuff in high school and university though, it's been a few years)
 
166, nice
 
I guess I could try sorting by difficulty and find some low hanging fruit in the later hundreds...
 
I am currently in the process of ⎕IO←0-ifying all of my solutions
because some used ⎕IO←0, some used ⎕IO←1, no consistency
 
10:16 AM
@ngn yep (empty list element) - but i don't think k9 dmend covers n-dimensions yet
l2021.01.25 3GB (c)shakti 2.0
 .[((1 2;3 4);(5 6;7 8));(;1);+/]
(1 2;7)
(5 6;15)
 
10:31 AM
@user87247 Hi Richard. So nice of you to stop by!
 
@Razetime this
 
cool
 
I like the biographies, do people write them themselves?
'John sits in the corner in the Dyalog office and can frequently be heard shouting at his computer.' lol
 
@rak1507 Yes.
 
10:43 AM
@Adám I want to make a site similar to tio which executes code using tryAPL and has some convenient things like postifying with SBCS link, CMCifying, converting plaintext to arrays and so on
 
@Razetime OK. So you also need the API? I think @RikedyP was going to publish it.
 
@Razetime If you want it ASAP I can push it up the list
 
I have used the API before, just asking if it's ok to do that
 
@Adám The API rocks. I'm still baffled at how many cool things the statelessness unblocks
 
probably host it on gh pages
 
10:45 AM
@Razetime It certainly is. From our perspective, it is really no different from using TryAPL itself.
 
@Razetime Yep sounds like A Good Idea™
 
@RikedyP feel free to tell me once done, then I'll start on a rough implementation
 
@RikedyP I don't even remember, why didn't we go with GH pages for the frontend?
 
@Adám No it's a single page hosted with Jarvis
since Jarvis is handling the API anyway
 
> why
Early versions of TryAPL 3 did use GH pages, but TIO as execution engine.
@RikedyP Would it make a difference who serves the frontend?
 
10:48 AM
@Adám No but if we're running a webserver to do Jarvis anyway there's not really a reason for it not to be part of it - at this point it would be more work to create the github repo
 
Right.
 
https://janiczek.github.io/tryapl-elm/#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
 
@Adám Also it's not particularly clean JS so
@MartinJaniczek the drop shadow on completions makes me instinctively feel that I should be able to click away / dismiss it
 
@MartinJaniczek the ⍺ and ⍵ refer to the previous and current value of the iteration, not the left argument, so you could do something like {⍺=10}
 
@RikedyP Hm, good input. Earlier I was experimenting with it being on right vs below the title, but I didn't think about the shadow too much.
 
10:51 AM
⋄ {⍵+1}⍣{⍺=10} 0
 
@rak1507 10
 
@Razetime The API is useable, if you email me rpark at then maybe later today or certainly later this week I'll send updated API description materials (which might become part of TryAPL help tab) if you'd be willing to give feedback / just check it makes sense?
 
⋄ {⍵+1}⍣{⎕←⍺ ⍵ ⋄ ⍺=5} 0
 
@rak1507 Illegal code
 
@rak1507 Why alpha though? It does one more iteration than I would expect I guess
 
10:52 AM
oh well that's boring
lol caps
 
@RikedyP sure, I'll send one now
 
@MartinJaniczek ⍺ is the current value, ⍵ is the previous one I think
 
@MartinJaniczek Yeah, that's a common problem with . In hindsight, it wasn't really defined so sensibly.
 
so its r parrk at die log dot com right?
 
@DyalogAPL though do note: ⋄ {⍵+1}⍣{⍺=10} 10
 
10:54 AM
@dzaima Execution timed out
 
@Razetime aye - just two r's is fine
 
basically, is often kind of very awful for anything that's not ⍣≡
 
⍣≡ is the cool one anyway
 
Maybe I should add {⍺⍺⍣(⍵⍵⊣)⍣(~⍵⍵ ⍵)⊢⍵} to APLcart.
 
I'm guessing I'll end up with ⋄ {⍵=10: ⍵ ⋄ ∇⍵+1} 0
 
10:56 AM
This is apply f until g ⍵ is true, including 0 times.
 
@MartinJaniczek yeah, recursion is often simpler than
 
I actually speculate that the annoying definition of was chosen because we don't have infinity in APL…
 
@Adám isn't that just {⍵⍵ ⍵:⍵⋄∇⍺⍺ ⍵}?
 
@dzaima Sure it is, and maybe with tail recursion it is as performant.
 
@dzaima I had the ngn/apl thing running and this crashed my browser
 
11:01 AM
@Razetime yay
 
boo
 
:D
@Adám huh
 
11:22 AM
Would this be acceptable "sum of digits"? +/⍎¨⍕1234
Can't think of anything more straightforward
 
@MartinJaniczek Yes, for code golf, though it isn't a function or full program. Just watch out for precision problems.
 
there's always 10(⊥⍣¯1)
 
so 10(⊥⍣¯1)1234 is kinda like (infinite list of 10s)⊤1234?
Ah I knew I read about it somewhere sacrideo.us/decoding-inverses
 
11:39 AM
@MartinJaniczek More like, "enough 10s", but you can also think of it as "10⊥1 2 3 4 computes the base-10 evaluation of those digits, so 10(⊥⍣¯1)1234 asks "what value of Y satisfies 1234 ≡ 10⊥Y?" and since you only need a single 10 to do base evaluation, you also only need a single 10 for ⊥⍣¯1.
 
Hello! I've been thinking about expressing an `IsPrime` function with simple operations for a while.
I managed to cut it down to
`⋄{2=+/⍵(÷=(⌊÷))⍳⍵}¨⍳30`
 
⋄ {2=+/⍵(÷=(⌊÷))⍳⍵}¨⍳30
 
@Martin Janiczek
┌→──────────────────────────────────────────────────────────┐
│0 1 1 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 0 0 0 0 1 0│
└~──────────────────────────────────────────────────────────┘
 
⋄ {2=+/0=⍵|⍨⍳⍵}¨⍳30
 
 @Martin Janiczek
┌→──────────────────────────────────────────────────────────┐
│0 1 1 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 1 0 0 0 1 0 0 0 0 0 1 0│
└~──────────────────────────────────────────────────────────┘
 
11:51 AM
"floor(div) = div" replaced with modulo approach
⋄ {⍵|⍨⍳⍵}7
 
@Martin Janiczek
┌→────────────┐
│0 1 1 3 2 1 0│
└~────────────┘
 
⋄ {⍵|⍨⍳⍵}6
 
@Martin Janiczek
┌→──────────┐
│0 0 0 2 1 0│
└~──────────┘
 
⋄ ↑{7,'%',⍵,'=',⍵|7}¨⍳7
 
@Martin Janiczek
┌→────────┐
↓7 % 1 = 0│
│7 % 2 = 1│
│7 % 3 = 1│
│7 % 4 = 3│
│7 % 5 = 2│
│7 % 6 = 1│
│7 % 7 = 0│
└+────────┘
 
11:57 AM
should I turn on boxing in the chat box userscript?
 
@Adám Yes, it is OK to use definitions from the Dyalog Wiki pages on the proposals for dual
 
@Razetime Not sure what you mean. It uses TryAPL, so boxing is always on (albeit at -style=min, not max like the bot does).
 
oh ok
 
Thanks @MartinJaniczek, I'll try out the modulo operator now!
 
If needed it's also a small step away from getting the list of divisors: ⋄ {i/⍨0=⍵|⍨i←⍳⍵}7 ⋄ {i/⍨0=⍵|⍨i←⍳⍵}6
 
12:14 PM
@diogotito To avoid confusion in the future, note that APLers tend to call things like | and + functions, reserving the word operator for things like / in +/.
 
@MartinJaniczek Although that could be optimized a bit to only try for numbers up to sqrt(n), since there can be no divisors above the sqrt. Something like ⍳⌊0.5*⍨7
 
@Adám Thanks for the clarification. Do APLers also call user-defined functions functions?
⋄⍳⌊0.5*⍨7
 
@diogotito
┌→──┐
│1 2│
└~──┘
 
@diogotito Yes, and there's no syntactic difference between built-in functions and user-defined functions.
 
Oh wait that sqrt trick is for prime factors, not for divisors. For divisors you need to go up to n/2 inclusive. Sorry.
 
12:21 PM
⋄⍳⌊0.5×⍨7
 
@diogotito
┌→────┐
│1 2 3│
└~────┘
 
Ok, thanks for the correction
BTW I tought up a tiny problem that I'm wondering how to do in APL:
I have a tablet with
oops, sorry, the apl Xorg key layout hijacked my Shift+Return XD
 
@dzaima ^
 
No problem, I have a rather unusual setup and I'm still fiddling with stuff xD
Oh, didn't notice that troubleshooting guide, thanks :D
 
@Adám i don't think the default layout should mess with shift+return
 
12:28 PM
@dzaima should or does?
 
It doesn't, sorry for the confusion
 
Oh. No worries.
 
@Adám i don't have an actual aplkeys.sh file to test with :p
 
lol
 
I used Super+Return to spawn a terminal but that sent Enter to the chat
 
12:30 PM
@diogotito so whats the question?
 
@diogotito ah, it will mess with super though
 
@dzaima which I happen to use as a prefix for my window manager, but it's ok for now
@Adám So, as I as saying,
I have a tablet with dimensions `2160 1620`
I want to write a nice APL expression to have
2160 1620 ÷ 1
2160 1620 ÷ 2
2160 1620 ÷ 3
and so forth
 
@diogotito it is on most others of course too. you are somewhere between the 10th and 12th person to have a problem with it in this chatroom :P
 
@diogotito What structure result do you want? A matrix or a list of lists?
 
Oh, APL distinguishes between the two?
 
12:33 PM
 
Right now I have
{ 2160 1620÷⍵ }¨ ⍳5
 
You could use the rank operator, try entering 1 2 3 4 5+⍤1 0 ⊢ 1 2 3, this means take rank 1 cells from the left (the vector) and rank 0 cells from the right (each of the scalars)
 
@diogotito 2160 1620∘÷¨ ⍳5 is neater.
 
      2160 1620∘÷¨⍳5
┌─────────┬────────┬───────┬───────┬───────┐
│2160 1620│1080 810│720 540│540 405│432 324│
└─────────┴────────┴───────┴───────┴───────┘
 
⋄ 2160 1620÷⍤1 0⊢⍳5
 
12:34 PM
@rak1507
┌→────────┐
↓2160 1620│
│1080  810│
│ 720  540│
│ 540  405│
│ 432  324│
└~────────┘
 
50 rep bounty to codegolf.stackexchange.com/q/218059 for the shortest dyalog (extended or otherwise) solution that is at least half as short as the existing 48-byter, i.e. 24 bytes (can't start a bounty for such a new challenge, so here is my informal statement)
 
@user41805 And you managed to write a <24 byte solution yourself?
 
I think that'll be pretty tough
 
no :)
 
wait for jelly and port
 
12:36 PM
Ooh, Neil's method looks promising.
 
@diogotito (what i do is just clear /opt/mdyalog/*/64/unicode/aplkeys.sh and change the keyboard manually with something like setxkbmap -layout previousLayout,apl -variant ,dyalog -option 'grp:switch')
 
@dzaima That works like a charm! I'll incorporate that in my configuration
 
@dzaima Didnt't you add that method to the wiki?
 
@Adám I did add the note to clear the file, but the replacement isn't under troubleshooting of course
 
@Adám It's the first thing in the wiki page. I hadn't read it until now :'D
 
12:44 PM
Ah right.
 
do note that for whatever reason the us,apl (and au,apl) option makes alt+shift+j give but shift+alt+j doesn't. Both work on eu,apl and lv,apl
 
@dzaima How about uk?
 
@Adám uk is invalid ¯\_(ツ)_/¯
 
gb?
 
@Adám doesn't allow shift+alt+j
 
12:48 PM
Is there any reason the English layouts shouldn't allow it?
 
oh, shift+alt appears to start a compose sequence..?
 
Yeah, that lets you type all kinds of strange things.
 
@Adám it's what i use to type all of my ಠ_ಠs and ¯\_(ツ)_/¯s, but i have it set to start with right super, not shift+alt
 
setxkbmap -layout gb,apl -variant ,dyalog -option 'grp:switch' allows my to type ⍤ with Shift+J when the "apl" layout is enabled.
 
@diogotito no markdown in multiline messages because SE
 
12:55 PM
@dzaima thanks for pointing me that
 
@diogotito "when the "apl" layout is enabled" the purpose of that is to make it get enabled on altgr. Do both altgr+shift+j and shift+altgr+j work for you?
 
@dzaima You're right!
In my case, with PT layout enabled I can use AltGr+Shift+J and Shift+AltGR+J to type ⍤
But then I'm unable to type @ € { [ ] } the usual way
 
@diogotito right, on non-ASCII keyboards things like that happen. if you have some unused modifier keys, you could choose the APL key the be a different key (what I do is just switch between APL and LV layouts when needed, but all i'm losing is accented keys, you're also losing braces and brackets which are still pretty important)
@dzaima (see options with grep "grp:" /usr/share/X11/xkb/rules/base.lst)
 
@dzaima @dzaima Awesome! I tried setxkbmap -layout pt,apl -variant ,dyalog -option 'grp:caps_switch' and for now I'm happy with it!
 
:D (still had pt layout typing this lol)
 
1:10 PM
@Razetime The spec for Under is here but it's some difficult reading.
 
yep I can see that
 
Hahahaha
I didn't mean to ask for tech support for this particular issue because it looks like it was already well documented in the wiki and such, but thanks thanks a lot for your time and kindness :)
 
@dzaima (might be worth putting that in the wiki?)
 
@dzaima yes
 
@Adám oh, it already is written, just /usr/share/X11/xkb/rules/evdev.lst not base.lst. the files appear to be equivalent, so ¯\_(ツ)_/¯
 
00:00 - 14:0015:00 - 23:00

« first day (1463 days earlier)      last day (1181 days later) »