The Nineteenth Byte

XAT (eXtended Advanced Technology)

ATX++Pro Gold A**

ultimate edition

ATXXX

Isn't that the adult version of Advanced Technology?

3:21 PM

So interesting thing: SE (obviously) complies by GDPR, which include a "right to be forgotten" or a "right to erasure". Given that deleted content is (almost) always visible for some users, how does this jive with that right? I'd guess through dissociation of account and content?

@cairdcoinheringaahing In fact, stackexchange.com does not comply with GDPR.

@Adám How so?

@cairdcoinheringaahing GA tracker without consent.

@cairdcoinheringaahing and maybe also

> Given that deleted content is (almost) always visible for some users, how does this jive with that right?

Although IIRC the "right to be forgotten" is significantly less of a protection on the users that one might hope

technically if necessary we are able to completely demolish a post's content (i believe this not only applies to what's visible but literally destroys it from the servers) but it takes two mods to manually do it to a specific post

if a user wants their presence to be forgotten/erased then dissociation is about the only tool we have

either via deletion or specific posts can be dissociated by CMs on request

3:29 PM

Yeah, the right to be forgotten is definitely presented in a strong way by GDPR info, but arguably dissociation isn't strong enough - I can still grab close to all data I could on Mego before he removed his account now, it just takes slightly more effort

@Adám Have you brought that up on MSE/with SE?

@cairdcoinheringaahing especially given that SEDE lists OwnerDisplayName in its Posts table iff the OP has a deleted account, in which case, it shows the unique userXXX name, where XXX is their user id

userXXX may be unique, but it has no relation to the user's actual information though, does it? It's dissociated with the person, no? (even if the posts they created are associated with each other)

technically, yes, realistically, everyone can easily find which ID mego is and know it was mego

@pxeger I suppose it doesn't allow you to find the same info as from a user's account page (which deleted users don't have), but you can find plenty of info that could be considered PII just by looking at their posts

it would make sense for accounts to just show as "anon" (like they do on dissociation request) for deleted users and maybe not make it so easy in the SEDE to list a deleted user's posts

@cairdcoinheringaahing No; it is a hopeless battle. GDPR compliant sites are the rare exception these days… mostly government sites and the occasional GDPR info site. Most sites advertising services to help you become compliant are not compliant themselves. Case in point: duckduckgo.com/?q=become+gdpr+compliant where only the 16th result is compliant, and that's a PDF where the URL root is non-compliant. Other than that, you have to go to result number 24 for a compliant site!

3:35 PM

Mego's Posts

also the deleted user page has a link to all of the posts that user has ever created (♦)

@hyper-neutrino Interestingly, Mego's chats show up as "Anonymous", but it's fairly easy to figure out who that anonymous user who spent a lot of time in TNB was :/

oh that's interesting. mine show up as user###### for whatever my old chat ID was, IIRC

@hyper-neutrino Yep:

in Jelly Hypertraining, Apr 27 '17 at 12:13, by user165474

Yay :) Thanks for making this room

vs:

Jun 29 '17 at 7:24, by Anonymous

This room was placed in timeout for 5 minutes; Seriously, stop abusing stars.

ah

Aaron Miller

3:38 PM

It looks like the meta identicon thing is fixed now.

should we have blamed caching? [.com]

Not this time for once :P

Aaron Miller

The funny thing is that I think I might have preferred the glitched meta identicon

@AaronMiller I just noticed that you have CGCC in dark mode, which isn't officially supported, therefore clearly whatever is making it dark mode is causing the bug status-completed :P

@cairdcoinheringaahing except messages shown as replying to those, despite not containing your name in their actual contents, show up with your old username there

Aaron Miller

3:41 PM

@cairdcoinheringaahing Oh, of course. I should've known.

What happens if I quote them - will the mention be recalculated or something?

ಠ_ಠ

lmfao deletion ninjas

why can't I undelete it? (can mods?)

no

No

Deleted chats are gone forever

3:43 PM

except not really, because room owners can see the history

@cairdcoinheringaahing except in their history lol

Man, this room is full of ninjas :P

@cairdcoinheringaahing Haya

CMC Input position < 0 || position >= this->length output position >= this->length && position < 0 Using any form of regex / re

i am not sure what the task is meant to be

print("position >= this->length && position < 0") ??

@hyper-neutrino Updoted

3:48 PM

@N3buchadnezzar Can you rephrase that in English?

@Adám Using regular expressions / some form of search and replace take in some string on the form xxx || ... and return... || xxx. <- Is this clearer?

@N3buchadnezzar Yes, but can you confirm that xxx and ... are arbitrary strings that do not contain ||?

@N3buchadnezzar Also, did you want us to change || to &&?

@Adám Yes to both (I was too late to change the last message)

@N3buchadnezzar QuadR, 25 bytes: (.*) \|\| (.*) \2 \&\& \1 Try it online!

lambda x: re.sub(r"(.*) \|\| (.*)", r"\2 && \1", x) should work

ah yeah what adam did

3:59 PM

CMC: Takewhile; Given a list of integers and a black-box function f : Z -> B (integer in, boolean out), take elements from the start of the list until one returns false under f

e.g. [1,2,3,0,4,5,6], f = x != 0, outputs [1,2,3]

@cairdcoinheringaahing Are we allowed to apply f as much as we want?

Yep

Normal code-golf ruules and whatever

taking f as a link, Jelly, 7 bytes: ¹ƤÇƇƑƇṪ

@cairdcoinheringaahing APL, 10: l/⍨∧\⎕¨l←⎕ Try it online!

last element filtering by invariance under filtering by the black-box for each prefix

i feel like there is some way to put something into the body of the prefix quick that isn't identity, but then you can't filter like that

4:03 PM

@hyper-neutrino invariance?

@Adám given f, is f(x) = x

@hyper-neutrino Also 7 bytes

@hyper-neutrino What happens if the first element gives false?

it's equivalent to Ç€Ạ$ meaning "apply last link to each, then check if all are true"

@Adám oh it returns 0 (not [0]) because the tail receives empty list

totallyhuman

@cairdcoinheringaahing Haskell, 9 bytes: takeWhile

4:04 PM

you can append ọ$ to fix that if returning a true empty list is necessary though jelly will not output anything in that case

@totallyhuman wow, how does that work? :P

@hyper-neutrino Also 7 bytes

@totallyhuman Love it when the implementation is shorter than the name.

these are more correct technically because it returns true empty list

i keep forgetting that tailing empty lists isn't really correct

@hyper-neutrino It should return [] though, no?

also 7: Ç€i0’ḣ@ (actually works)

@Adám yes

2 mins ago, by hyper-neutrino

you can append ọ$ to fix that if returning a true empty list is necessary though jelly will not output anything in that case

(taking that trick from jonathan allan's comment to my answer to the zero sum sublist problem

4:09 PM

This nicely illustrates why a "filter-by-black-box-function" function is inferior to "filter-by-array": For the former, you can only customise the black-box-function, but not make a decision based on the results for (in our case) all previous elements.

why does vyxal have a one-byte element for 26 lol

because it's Vyxal :P

it is actually kinda helpful for any alphabet based ascii challenge which there are a considerable amount of

~~god only knows~~ actually lyxal will know lol

@Lyxal i refuse to turn to the dark side of (ab)using flags :P — hyper-neutrino ♦ May 2 at 3:03

@hyper-neutrino You became the very thing you swore to destroy!

4:12 PM

you are not the first to point that out :p

FTFY

totallyhuman

@totallyhuman alternatively and perhaps more interestingly, 11 bytes: (fst.).span

Razetime

@totallyhuman husk, 1: ↑

lol

totallyhuman

oh yeah

i should've learned husk at some point

4:33 PM

Why the hell does Julia use * as its string concatenation operator

Razetime

if / is split then * can make sense as concat

* makes more sense as join to me (that's what Ruby uses)

Razetime

whatever it chose it's gonna stick with it

5:08 PM

Well, I've officially formalised the spec for Mixology, a Chef-style language but for cocktails. Time for the worst part of making a language: actually making it :P

:P

flagz

what the hell

lol

why tho

5:23 PM

y not?

have we had a question about expressing numbers as the sum of 3 cube numbers?

I think I've got something in the sandbox about it

Or, I've been meaning to post something in the sandbox about it, I can't remember

recent, or old?

Nothing here

because if you're "meaning to post" something, I'm bagsying it!

5:30 PM

@pxeger I'll take a look, feel free to have it if not :P

Doesn't look like I've posted it yet :P

It would be a subset of this

That'll teach me to be quicker with drafts

lol

But I think it's ok because using only integers is a fair bit simpler?

> Your submission should be able to compute a solution for every input given enough time and memory, that means having for instance two 32-bit int representing a fraction is not sufficient.

yeah I'm not happy about that either

well, time to go write it up!

5:32 PM

That makes me think that an integers-only one should be acceptable, albeit a very close challenge

Also, I don't think this formula could be used on integer-only inputs

yeah

A: Sandbox for Proposed Challenges

7:35 PM

Mobile view is being deprecated in favour of the responsive design

8

New Posts

8:09 PM

0

Q: A Suitable Mission

Summer Klerance, a senior in college, is what her teachers referred to as GBL*. Students in her probability class have been assigned individual problems to work on and turn in as part of their final grade. Summer, as usual, procrastinated much too long, and, having finally looked at her problem, ...

Sandbox Posts

8:38 PM

0

Sum of three cubes code-golf math integer number-theory polynomial Given a positive integer \$ n \$, find three integers \$ a, b, c \in \mathbb Z \$ (they may be positive, negative, or zero) such that \$ a^3 + b^3 + c^3 = n \$. If it is currently known that no solution exists (i.e., for \$ n \in...

How would people feel about an infinite-output tag?

@cairdcoinheringaahing deprecated?

@pxeger I can good words :P

want an 11? :p

Q: Proposal for tag: [infinite-output]

Feel free :P

1

I recently created the infinite-output tag for my newest challenge. After a discussion in chat, there's a bit of disagreement on whether it's needed. Here are some questions I think the tag would apply to: Shortest code to produce infinite output (For obvious reasons) Implement a truth machine (...

discussion tags

@NewPosts Isn't this unobservable?

8:45 PM

@Adám I don't know if you saw, but I fixed the tio-api bug you noticed!

@AviFS I did.

Curious what were you using it for!

As TIO gets increasingly out of date, we might want to move APLcart's links to a different service, and for that I might need to "recompile" all the hardcoded TIO URLs in the database.

Ah, that makes sense!

Let me know how using tio-api goes, and if you have any requests!

Can someone explain the "Suitable Mission" challenge to me?

user

8:49 PM

The OP can probably do it best :P

Also, I'm planning on adding support for any/all online interpreters as soon as there's a demand. Are there any in particular you're thinking about switching to?

user

Might want to ask them for clarification instead

@user I just want to check if it is just unclear for me.

@AviFS afaik, there aren't any that support Dyalog APL yet.

Eg. ATO or TryAPL? Both are planned, but I haven't started as no one has shown interest yet

Well, TryAPL's URL API is super simple.

8:50 PM

Still, for convenience!

I can just have a json with functions to parse a bunch

And you could then switch between permalinks for different sites within the api

Well, it is literally https://tryapl.org/?clear&q=<code>&run where you can leave out clear if you don't want to start off a clean slate.

@Adám What about just storing the hardcoded urls as eg. [input, code, flags]?

@Adám True, but you'd have to find the relevant docs or notice the pattern, first!

@AviFS They'd take way more space.

People don't have to use tio-api functions to have them be helpful

Eg. if they look for tryapl and they see function makeLink(code) { return `https://tryapl.org/?clear&q=${code}&run;` }

Try double-backticking.

8:55 PM

if you want backticks in a code block do \`

Or simply use double outer backticks.

@Adám The things you know!

Thanks hyper!

``function makeLink(code) { return `tryapl.org/?clear&q=${code}&run`; }``

For anyone who's still confused :)

Haha, thanks!!

@Adám In that case, storing them as is, as TIO urls is prob the best way to go

Just remove the leading https://tio.run/#

@AviFS Yeah, especially as the entire database is sent to the frontend on pageload.

8:57 PM

I think TIO urls are as compressed as possible

You can take whatever functions you need to parse the url every time

Except the superfluous / after run

And then you can just generate a link to whatever you're using at the time

You can also send them over to APLgolf and let Razetime figure out who will host :P

Given that he also has it hooked up to TryAPL, I think that should be safe

@AviFS But they are just links. If I intercept a click, how can I reliably let the user continue?

@Adám Oh, I see...

I've no idea how/if you can do that in js

I'd think you could though

But if not, then I agree you should just recompile the links in the database!

I think the browser would block it to prevent hijacking.

9:01 PM

But also, what about TryAPL and what about APLgolf?

APLgolf has no execution engine, and is mainly intended for golf, which APLcart isn't.

TryAPL is whitelisted APL, not sandboxed.

@Adám True, but it finds an execution engine. And since all the entries are short snippets I think it works APLgolf is a nice interface

I don't think it's very different from TIO's

@AviFS Isn't the execution engine built into its URL?

@Adám Yeah, that!

But it has a default, if you don't specify

Oh.

9:04 PM

It still worries about the execution engine for you

@Adám You mean eg ⍎ isn't allowed?

@AviFS It is, but with the same restrictions as TryAPL has.

Sandboxed means it runs in a closed environment so everything is allowed, and whitelisted means it only runs stuff deemed safe?

Exactly. And many things in APLcart use files etc.

@Adám That's what I mean, I'm asking about TryAPL

@Adám Asking about this comment

@AviFS Yes, that's what I was answering. TryAPL's ⍎ has the same limitations as TryAPL.

9:06 PM

Understood! Maybe ATO will get there, or maybe Dennis will come back, or maybe something else

I agree that it's too early to tell

If only there was a simple sandboxing system with timeouts, we could host our own.

@Adám Ah, it's just slightly circular then. Because I don't know what it's limitations are, that's what I'm wondering about!

@Adám What about allowing the user to toggle between modes in TryAPL?

You can choose either: whitelisted with no time-out || sandboxed with time-out

@AviFS Circular, or just a recursive definition!

@AviFS TryAPL also has a timeout. 10s, vs TIO's 60.

@Adám (asking here rather than the Orchard cause it's a small-ish question), how does tacit APL have while loops? Is it the ⍣ operator?

Oh...

@Adám But then what was this saying?

9:09 PM

@AviFS If we had a simple sandboxing system with timeouts, we could link to that instead of to TIO.

@cairdcoinheringaahing You mean ⍣?

@cairdcoinheringaahing Yes.

Oh yeah! Here: aplcart.info/?q=while%20loop#

@AviFS Is that not what I typed?

@AviFS '⍣'≠'⍣' ‽

9:11 PM

Eg. the famous golden ratio calculator: 1∘+∘÷⍣= 1

@AviFS Clicking (#) or accesskey c should copy the URL to your clipboard with a little less noise.

Or better: +∘÷⍣=⍨1

@Adám I did that though, I thought

@AviFS Would look like aplcart.info?q=while%20loop then.

@cairdcoinheringaahing Even better: +∘÷⍣=⍨1

@AviFS aplcart.info?q=phi%20limit

9:13 PM

@Adám Isn't that what I have

No, you have a slash and a hash.

Ah, don't know how that got there!

9:30 PM

@SandboxPosts (cc @pxeger) I'd suggest using \bmod instead of \mod

Reduces the spacing, and makes it look better

9:59 PM

Took me way too long, but I just PRed adding TIO links to both golden ratio entries

Searching golden ratio right now gives 2 results with no links

@cairdcoinheringaahing Wahooo!! You're really doing it! Nice :)

Starred it is

And adding to Bartender chat room description :P

@AviFS Haha, thanks :P

I've resorted to using regex to match the lines, feel free to burn me in hell :P

Don't blame you even a little!

Does anyone know if you can add newlines to SE room descriptions?

Or even just how to pad with spaces?

Or any other markup-esque things you can do?

@AviFS I don't think you can

Descriptions are plain text IIRC (aside from rendering hyperlinks clickable)

Fun fact: I tried the example provided in spec tonight, it does not taste nice (or really make sense):

An "Example Program". Serve up with a straw, in a Martini glass with a lemon twist garnish

- 2 oz of vodka
- 1 oz of pineapple juice
- 1/2 oz of Campari
- 1/2 oz of Kahlua
- 2 dashes of chocolate bitters
- 1 mint leaf, muddled

Add the pineapple juice to the shaker. Muddle the mint leaf. Add the vodka, Campari, Kahlua and chocolate bitters to the tin. Dry shake over crushed ice. Double strain into glass.

Campari, mint and Kahlua is a ...unique mix :P

@cairdcoinheringaahing Well I tried!

Now I just have a wall of auto-gen text about changing room description

See Bartender room for new description

@cairdcoinheringaahing Wow wow! That's def fancy

I tried c/p unicode nonbreaking spaces, but that didn't work

Still collapsed

So, I mean, you can pad with - or . etc

But wondering if there's a nonvisible char to do that with

@AviFS I don't know how much you know about cocktails, mixology and all that, but Campari is a bitter, grapefruit style liqueur, Kahlúa is coffee and "Dry shake" means "shake without ice", so it both doesn't taste nice, nor make sense to make :P

10:14 PM

SE IS SO LAME

@AviFS The chat system for sure :P

Everything about chat rooms

@cairdcoinheringaahing Oh, I didn't know Campari, but def the rest!

It's just a tropical mocha :P

Hmm, the bitterness is very strong :P