The Nineteenth Byte

lyxal

00:47

@Neil 3 bytes now

OldSandboxPosts

01:00

Sandbox posts last active a week ago: Langton's Ant

ATaco

01:13

I've had a theory for a new Parser for a PracLang

Instead of parsing token types by requesting specific tokens, I just claim whatever tokens I can, and then check if they have certain properties.

Which should make adding Syntax aware Macros much easier

Will probably be a bit slower than the other method though, as it has to check more times per token...

Rydwolf Programs

01:34

@cairdcoinheringaahing No, the bots are hosted in the cloud

Nothing is hosted on site except for Ginger's new VM and, in the future, RTO's back-end

Also I haven't gotten around to re-adding STATUS yet sorry

Bbrk24

03:14

@lyxal That gives the odd numbers ascending, not descending

lyxal

03:48

@Bbrk24 then Neil's answer wasn't valid either

hey wait that's not u$e that's just $e

the whole point of the outgolf was that -1 ** n was added as a built-in last night

Unrelated String

lol nice

Simd

06:08

Has anyone tried mojo?

AviFS

07:16

Wordle 688 4/6

🟨🟨🟨⬛⬛
⬛⬛🟨⬛⬛
🟨🟨🟨🟨⬛
🟩🟩🟩🟩🟩

I know we're over Wordle, but cool-ish pattern, maybe?

@Simd No, but it came up in several places in my feeds.

Doesn't seem too inherently fascinating, like syntax-wise and feature-wise in terms of PL design, from what I could tell briefly. More interesting to see how it's executed and if it delivers on performance. That's what I took away from it, at least. I'm usually really interested in new PL.

Neil

07:36

@lyxal µǍ* works, although maybe someone already suggested it in scrollback

and µN$e did work, I don't know why you thought it didn't

lyxal

@Neil because I thought it was u$e which is very different to N$e

New Posts

09:17

0

Q: Freedom Fence Freedom Fence

Freedom Fence is the top fence installation company for vinyl fencing, wood fencing, aluminum fencing, and privacy fencing in Sarasota and Bradenton Florida. Offering some same day complimentary estimates if the schedule allows! Fence company Sarasota ,Sarasota Fence Company Address: Sarasota, FL...

browser

Unrelated String

for a split second when i just saw the title i thought that might be an actual challenge somehow

maybe because usually spam titles aren't usually the name of the company repeated twice

mousetail

Freedom fence sounds like it could be a mathemetical concept or a puzzle

lyxal

Gee that was a quick spam nuke

The Thonnu

@UnrelatedString I think titles need to be at least 15 characters

Unrelated String

ahhhh

mousetail

11:52

Simd

12:36

@AviFS if only Julia could usefully produce stand-alone executables I think it would be my favoured choice

Sandbox Posts

13:20

0

A: Sandbox for Proposed Challenges

Complex remainder Complex remainder can be defined in a similar way to real remainder as \$ y - x \lfloor \frac y x \rfloor \$. However this depends on a suitable complex floor function. There are a number of ways of defining a complex floor function (for at least one of which there has been a pr...

mousetail

14:17

> Since this is a new module, no major backward compatible issues are expected. The module name “subprocess” might collide with other, previous modules [3] with the same name, but the name “subprocess” seems to be the best suggested name so far. The first name of this module was “popen5”, but this name was considered too unintuitive. For a while, the module was called “process”, but this name is already used by Trent Mick’s module

Imgine if we had popen5

urllib3 is confusing enough

Ginger

~~come to the Rabbit side, we don't have confusing names~~

mousetail

Easy to say before you have needed to make backwards incompatible changes

Ginger

I what now

mousetail

Make changes to the language that are not backwards compatible

Ginger

why'd that be an issue?

mousetail

14:24

It's every language designers nightmare

And leads to situations like urllib3

Oops, I made a decision that made sense at the time but I now realize is terrible but I can't change it because a ton of code depends on the old behavior

Ginger

I don't get what the issue'd be

if I make a non-backwards-compatible change to the Rabbit interpreter I just release a new version and wait for everyone to update their code to support it

mousetail

So you are going to maintain every old version indefinatly?

Ginger

it's not like they'll suddenly cease working

mousetail

There could be security critical issues raised

Ginger

sucks for them

mousetail

14:37

That's usually the reason why you don't want to use old versions

Ginger

install updates kids

mousetail

But updating code to work with new versions is extremly difficult

Ginger

¯\_(ಠ_ಠ)_/¯

mousetail

it took a team of 8 of us 2 months to upgrade PHP to 8.0

Ginger

that's because you're using PHP, and you have my condolences

mousetail

14:38

It look like 10 years before python 3 was widely used

Still many projects haven't updated

That's why there is urllib3 instead of python 4

Ginger

there's only going to be one canonical Rabbit version: the latest one

when a new major version releases the old one will no longer get updates, because that's not how semver works

mousetail

That would make it impossible to use for any genuine purpose

Ginger

?????

mousetail

I need guarentees that every version is going to be supported for X time

Ginger

well ofc, I'm not releasing a new major version every week

mousetail

14:40

But when you do I have 0 warning and my application is instantly using a outdated version?

I need months if not years to schedule time to upgrade my program

Ginger

hmmm, fair point

but Python's only had 3 major version increments in all the time it's existed

and minor versions are backwards compatible

regardless, that's an issue for later

mousetail

Python is maybe a bit too convervative in breaking backwards compatibility. PHP is ironically actually very good at how to smoothly introduce changes. One version makes it a notice, then a warning, and only then the next version is something removed

Ginger

agreed (with Python's conservativity)

mousetail

Making sure there is overlap between the old and new functions so you can move them over a bit at a time and not need to fix everything at once

Ginger

ideally Rabbit's core modules would always be backwards compatible, but that's of course not possible

(by "core" modules I mean ones that are part of the interpreter and updated when it is, like os)

mousetail

14:46

Well then you would need a python situation where you just add more ways of doing things but never remove old ones

Ginger

but that's also bad because it causes feature rot

unfortunately the core modules are fixed to the interpreter version, and therefore can only have breaking changes when the interpreter jumps a major version

mousetail

That seems reasonable

Ginger

since there are only going to be a few dozen core modules I think that's a fair price to pay

the rest of the standard library can have its own update schedule

mousetail

I'd suggest using the same version for the whole standard library and the interpreter.

Ginger

nah, that's not how Rabbit's going to do things

mousetail

14:49

It's not really a standard library at that point

Ginger

the core modules implement essential functionality and are part of the interpreter; the rest of the stdlib (everything in the rabbit namespace) doesn't come with the interpreter but are officially supported and maintained by the Rabbit organization

mousetail

If you have to install the correct version yourself

Ok so it's not a standard library then

Ginger

yeah, not in the traditional sense

mousetail

Just some official modules

Ginger

they're more like Java's javax stuff

mousetail

14:50

Yea then it makes sense to version them seperatly

Ginger

so there will be official modules for things like high-level networking, but low-level sockets will be a core module

requests vs socket

mousetail

And I imagine you could just fork requests and import your own version if you don't like the official one?

Ginger

sure, they'd all be open-source

I could make ginger.requests if I don't like rabbit.requests

mousetail

I mean they wouldn't use magic to interface with the interpreter in a way that wouldn't be accesible to third party modules

Ginger

correct, they'd just be easier-to-use interfaces to the core modules

the same way requests is a layer over socket or smth

mousetail

14:53

requests uses urllib3 I think

Ginger

F

well I will be having none of that

mousetail

I would suggest a library with basic data types for url parsing and query parameters escaping etc. separate from the one that can actually make HTTP requests

Ginger

(I mean if you really wanted you could skip socket entirely and bundle a JAR to use Kotlin's native networking, but I have no clue why you'd do that because that's what socket does for you)

mousetail

(for speed probably)

Ginger

@mousetail the core modules will only include socket; all that stuff would be a standard module

I'm keeping the core modules as slim as possible to cut down on the interpreter size

since all that stuff will be part of the interpreter

mousetail

14:56

Yea in standard. Similar to PHP's PSR, to have a standardized interface that all libraries can share

Ginger

yup

the only downside of this idea is that if the Rabbit site goes down you won't be able to use any standard modules you don't already have copies of

~~hopefully the servers don't get hit by lightning~~

mousetail

You could pre-install them to your module cache

So you have at least one version available always

Ginger

I could probably have the Rabbit installer automatically put all available ones in the root store during install

but that'd be of limited utility if you're using software that depends on older versions

mousetail

Yea might not be worth it

Ginger

SLaaS: Standard Library as a Service

:b

mousetail

15:00

You probably want to use some kind of CDN for the modules so if any individual server goes down you can still download the packages

Ginger

that's a later thing

mousetail

15:39

Cloudflare is free

mousetail

18:48

@cairdcoinheringaahing Can you please post a answer codegolf.meta.stackexchange.com/questions/25758/… to either give official approval or suggest some alternative approach? Your edit to the editing guidlines suggest I need meta approval but I don't think anyone is going to answer so not sure how to get the approval

caird coinheringaahin g

Sure, I can write a quick answer :)

caird coinheringaahin g

19:12

@mousetail Posted, I threw together a super quick SEDE query to help find the links that aren't associated with a post id

Effort to update anonymous imgur images

4

mousetail

Thank you

Ginger

👍

Transcript for