@user I am not a township in New Jersey

At anyone or anything

So, I'm trying to write a custom sync/backup tool for my laptop and desktop

So that I can just run a command and sync them up, and work on a project on whichever device I feel like

And Docker is trying to kill me

I want to use OverlayFS for incremental backups/syncing

So I have a known stable point, and whenever I sync one computer to the server, it creates a directory containing only the changes to the filesystem, so that once I sync the other computer the system can negotiate how to merge the changes

But if I have Docker files on my computer, since Docker uses OverlayFS too, it would break things since you can't put an OverlayFS filesystem inside an OverlayFS filesystem due to some weirdness with how they work

(If I ever become a billionaire and make that OS, a filesystem that natively supports (potentially nested) overlays/unions will be a part of it)

@RydwolfPrograms noice

why not git

I guess what I'm gonna have to do is just special case any directories containing block/character devices

@Seggan Git makes no promises about maintaining the integrity of your files in edge cases

Empty directories get removed, for example

and? just create them again

I don't want to imagine how it'd handle block devices, unusual permissions, or xattrs

@Seggan That rather defeats the point of a backup tool

true

And also, I don't want things silently breaking all the time

i still use git

It would probably work fine for backing up like, code and documents

which is what ure doing, no?

No, this is for the full disk

Well, the backup side of things

I'll only be syncing (mostly) code and documents, but if some weird stuff slips into my homedir I don't want it breaking everything

Like I've got some dumps of overlay filesystems which contain whiteouts in my laptop's Linux homedir

@RydwolfPrograms Are you sure such a thing doesn't already exist, with all the features you want?

There must be such tools for sysadmins and stuff

Maybe, but I haven't found anything good enough and it's pretty easy to get what I want working

presenting: the Light Emitting Resistor

lmao

@Ginger How lol

Looks like your resistor got promoted to a diode

@Ginger aw, I thought this was an original joke

Doing factory work

Items include

You can buy one

One consideration: the best option I have for the server to host these backups on is rto_wolf...which has pretty limited storage space and also will be hosting untrusted code (which shouldn't be able to break out, but if it does, maybe it shouldn't have access to my files :p)

Ooh unrelated but I just had a great idea

I should make a bootable thumb drive that, on a fresh Fedora install, will go through all of the steps necessary to install RTO

So if I ever need to start from a clean slate or someone else wants to spin up an instance, there'd be no mucking around with trying to get it running

I guess I could just run this on my droplet

Given that what I'll be backing up will mostly be code and screenshots and stuff, so it won't be more than a few gigabytes

@RydwolfPrograms there’s tons of software out there to sorta automate installation

Ansible, puppet, chef

Yeah but this isn't like, setting up some software on a computer

This is setting up the whole computer and OS specifically for RTO

It would also be taking care of things like the SSH config

@RydwolfPrograms Thought you could use Ansible for that

Maybe idk

I guess making your own tool would be more fun anyway

> No matter your role, or what your automation goals are, Ansible can help you demonstrate value, connect teams, and deliver efficiencies for your organization. Built on open source, Red Hat® Ansible® Automation Platform is a hardened, tested subscription product that offers full life cycle support for organizations. Explore how Ansible can help you automate today—and scale for the future.

Pure nonsense

I hate marketing speak like this

Everyone does it

No one takes it seriously anyway

I sure do love demonstrating value

> deliver efficiencies

Would you like one efficiency or two efficiencies with that image, sir?

Can I have an extra efficiency with my soup please

Also how does a software installation automation tool "connect teams" lol

Helps teams share a common config?

I guess lol

Or perhaps they have actual ansibles so teams can communicate instantly

@RydwolfPrograms i'm disappointed this is only 5 hits on the reveal bullshit bookmarklet

Ooh, what's that?

I need a bookmarklet like that

javascript:(function(){var d=document,s=d.createElement('script');s.crossOrigin='anonymous';s.src='https://unpkg.com/@mourner/[email protected]/bullshit.js';d.body.appendChild(s);}())

thx

forget where i found it

I love this

it's beautiful

lol

omg the bullshit detector script

@UnrelatedString mourner.github.io/bullshit.js

i know that's where it's from; i forget how i found out about it

i found it on ycombinator.

i don't really use that but i might have found it in a thread there someone linked here

@RydwolfPrograms imagine :p

me when VSCode says my application running on port 8069 is available

Just wait until you hear about port 6969

Which apparently is used for "acmsoda" and was reserved by a Daniel Simms

what

That's what the port's officially registered for lol

I... didn't know there was an Official Port Registry

Every port has an official use case (aside from some reserved/unallocated ones)

Technically you're supposed to reserve ports for everything you use them for

But since nobody's going around running acmsoda on their server or whatever, people just use whatever port they feel like

It's a mess, we really need 32-bit port numbers and a PUA

@Seggan how should I do operators in ANTLR? Rabbit uses infix functions instead of defining plus, minus, etc. in the grammar file

I would hope that syntactically things parse identically regardless of what has and hasn't been defined as an infix op

Since if that wasn't the case A) you'd need to parse your code twice and B) a library could change how your code gets parsed

But that creates the issue of order of evaluation

Because we live in a society :(

@RydwolfPrograms yes, of course, but I'm not sure how to define that in my grammar file (or more accurately I'm worried that the way I'd do it is wrong, because I've never used this tool before)

also, how does order of evaluation work here?

Yeah that's the issue with custom infix ops

I wonder how Kotlin does it

I assume they'd all have identical selectiveness, or maybe two different levels you can choose between

evaluating them from left to right makes sense, but what about the mathematical operators, which everyone expects to have a set priority? Should in be an infix function?

Problem is that it would have to be operator-wide, the class can't choose or else two classes could define an operator with two different precedences

@Ginger What I'm suggesting is to give every custom op a precedence either tighter or looser than all the arithmetic stuff

but I want to define the arithmetic stuff as infix ops on basic types!

You'll need to special case them in the parser

so if you want to define a custom + you just override the function instead of fooling around with __add__ and such

@RydwolfPrograms >:(

I'd suggest also having a string pattern that can be used in your operator names which gives it a precedence comparable to boolean operators

@RydwolfPrograms wdym?

So that you can define things like a nor or xor operator

You'd need it to be determinable at tokenization-time whether it's a tight or loose custom op tho, so I'd suggest using capitalization or something to differentiate them

I'm still confused

E.g., tight custom ops are lowercase ("abcd" slice 2) and loose custom ops are capitalized (1 == 2 Nor x < y or 1 == x XOR 2 < y)

I'm not sure what "tight" and "loose" mean here

Tight is tighter precedence, loose is loose precedence. + is looser than *

Since * takes things closer to it than +

(It's way easier for me to remember than high/low precedence)

okay, I can basically do whatever I want at tokenization time so I can just add a keyword for precedence

No you can't

how dare you tell me what I can and cannot do

What if Foo declares nor as tight and Bar declares it as loose

how is that an issue? it'd depend on whose nor is being called

it'd be bad practice, but not an issue

And you don't know that until after you've parsed the code

And it's ambiguous anyway

x + y nor z

I do in fact know that: the left-hand class is the one whose function is being called

Is that x + (y nor z) or (x + y) nor z?

@RydwolfPrograms that'd be x.plus(y.nor(z)) with default precedence

If x and x + y are different types, and x's type declares nor as one precedence and x + y declares it as a second, it's ambiguous

@Ginger But I'm saying there should be a second precdence and you're suggesting that as a keyword

In that second precedence that would be x.plus(y).nor(z)

You need to know every operator's precedence class based purely on its name

🤮

I am not doing that

I suppose in theory you could pre-parse the code and ignore every non-class-defintion statement, and throw an error if there's conflicting operator precedences within the same scope, but that could cause all sorts of weirdness and slows your interpreter down

I would suggest just having a single precedence for custom ops

if you don't like the different-capitalization option

I'm still not sure why I can't handle this in tokenization, seeing as I can basically do whatever I want while turning the parse tree into a set of nodes

...oh, right

yeah I see what you mean :/

it looks like Kotlin has predefined precedence for symbolic operators

If this was my language I'd do a trichotomy based on capitalization

I'm going to do the Kotlin thing, at least for now, because it's the simplest option

nor for tighter-than-arithmetic, Nor for tighter-than-boolean, and NOR for looser-than-boolean

the downside of that is that you can't define custom symbolic operators, but ah well, who was going to define a <:-:> operator anyway q:

(besides you Radvylf)

That's probably a requirement

Otherwise you'd run into ambiguous cases almost instantly

Like if someone defines a ++ operator, now a++b (a plus unary positive b) is ambiguous

But since the parser always has to be able to handle ++ existing since it doesn't know if it does yet, a++b is always ambiguous no matter what

ok, so there will be a fixed and overridable set of custom symbolic operators, and you can define custom named operators which will all have the same (left-to-right) precedence

And same goes for literally any symbol-infix-symbol-prefix-unary pair

Well I am donating blood

Currently

not sure how you're using TNB while doing so :p

Only need one arm for bleeding and one arm for typing :p

I don't think I was designed to lose a pint of blood in 20m lol

Also wtf apple:

I know I'm an MCR fan but...

I'm not all gottik and deprezzd

> I don't think I was dead

@Ginger left associative, but thatd break math

kotlin how it does it

the other way to do it is since antlr rules can be self-left-recursive

check rols grammar

@Seggan now Kotlin is unhappy about me trying to access a property in an ANTLR context

it says ctx.prefixOp is an unresolved reference

Well I have survived :p

I don't think I was dead

good job

I'm just baaarely above the 110 pounds minimum limit and I was apparently pretty dehydrated (despite drinking a whole water bottle of water before donating), so I felt preeetty bad afterward

I'm talking like, crazy lightheaded and arm-fell-asleep across my whole body (which I'm pretty sure I was supposed to tell them about but I'm good now so meh :p)

Probably the first time I've ever been truly thankful to be given a grape juice box lol

weirdly, it also says ctx.location is an unresolved reference

something sus is going on

@Ginger have you run generateGrammarSource?

@Ginger well...

wait no lol

@Seggan yes I have

@Ginger thats because location is an extension property i defined

me when

lemme just yoink that

ure totally not gonna break the license, are you

uhhhhhhhhhhhhhhhhhh

dw, i give you full permission to use it (as long as u attribute it)

there was a license?

ok lol

you'll probably have to remind me about that once I release Rabbit

Well if there wasn't a license it'd be even worse lol

Since it'd be copyright infringement no matter what

yeah without license is all rights reserved

Licenses are whitelists not blacklists

@Ginger github.com/Seggan/Rol/blob/master/LICENSE

ok I added the code and regenerated the sources

error did not go away

is it a vscode error or a gradle compile error

> Unresolved reference: location kotlin(UNRESOLVED_REFERENCE)

I'd hope as a modern language Kotlin provides more information than that in its error messages lol

it does

@RydwolfPrograms that's all it gave me

thats a very weird error message

nvm

you havent imported location

@RydwolfPrograms its just a weird format. can be rewritten as Kotlin (UNRESOLVED_REFERENCE) Unresolved reference: location

added import community.rto.ginger.rabbit.tree.common.Location to the visitor file, error is still there

Yeah but like...if I do the same thing in Rust, it gives me way more information and stuff

error[E0412]: cannot find type `RiprState` in this scope
   --> src/cipr.rs:140:12
    |
8   | struct CiprState {
    | ---------------- similarly named struct `CiprState` defined here
...
140 | type Xyz = RiprState;
    |            ^^^^^^^^^ help: a struct with a similar name exists: `CiprState`
    |

hey guys

I have an issue

Hi!

also it's still yelling about prefixOp

@AiraThunberg don't we all

I want to query that if I start a bounty

will it impact my leaderboard stat?

@Ginger try doing prefixOp()

@AiraThunberg Like in the rep leaderboards?

cz I am really going high in the weekly, yes

@RydwolfPrograms kotlin is a lot more dependent on ides for error highlighting and stuff

@Seggan didn't fix it

I believe so, but nobody pays much attention to the rep leaderboards. I started a bounty this week, lemme check if it changed my spot

@Ginger did you make sure the visitor is taking a non nullable parameter

can i see the code

I don't think bounties change your place

@RydwolfPrograms well, I do, I am showing this of, a show-of girl at school

you sure mate, 0_0

override fun visitExpression(ctx: RabbitParser.ExpressionContext): UntypedExpression {
        return when {
            ctx.atom() != null -> this.visitAtom(ctx.atom())
            ctx.prefixOp() != null ->
                    UntypedPrefixExpression(
                            this.visitExpression(ctx.expression(0)),
                            PrefixExpressionType.fromSymbol(ctx.prefixOp().text),
                            ctx.location
                    )
            else -> throw AssertionError("?!")

@Seggan ^

Oh it does

sikwww

the errors are on both occurrences of ctx.prefixOp, as well as ctx.location

sikww

bruu

._.

@Ginger you havent imported the location prop

as i said before

What were you going to bounty? I can put the bounty up for you if you want

@Ginger whats the grammar for the expression

@Seggan fixed that, I typed Location instead of location :|

good

expression
	: atom
	| expression NEWLINE* DOT NEWLINE* expression
	| expression NEWLINE* LPAREN NEWLINE* (
		args += expression (
			NEWLINE* COMMA NEWLINE* args += expression
		)*
	)? NEWLINE* RPAREN
	| prefixOp=NOT expression
	| expression NEWLINE* op=(ASTERISK | SLASH | MODULO) NEWLINE* expression
	| expression NEWLINE* op=(PLUS | MINUS) NEWLINE* expression
	| expression NEWLINE* op=(LESSTHAN | GREATERTHAN | LESSEQUAL | GREATEREQUAL) NEWLINE* expression
	| expression NEWLINE* op=AND NEWLINE* expression

@Seggan ^

no thanks so much @RydwolfPrograms

I heartedly appreciate your words though

@Ginger can you find the generated parser by any chance

the files ANTLR made from that? sure, easy

yeah check if theres a prefixOp field in it, and make sure the package is what you expect

the ExpressionContext definition starts with:

public static class ExpressionContext extends ParserRuleContext {
    public Token prefixOp;
    public Token op;
    public Token customOp;

mhm

that checks out

anyone good at math here?

package is correct

@AiraThunberg Depends, how good :p

@Ginger last thing: try deleting the build folder and regenerating the grammar sources

XD, I started a bounty on a question in Math SE :P

it is like 100 points, all my life earning there

you might check it out

though it was difficult before, Neil made an algorithm to solve it, the answer here must be mathematical representation of his algorithm

@Seggan well, the original 2 errors are gone, but now there are 10 new errors because Kotlin can't find the generated parser files :b

did you regenerate them :P

yes

hrm i probably skipped smth by accident in the build config

One thing that might help is finding a formula for the diagonals. If you can figure out a formula for going up the diagonals and down the antidiagonals, you just need to switch between those periodically

nope, nothing i skipped

I am really dumb :]

It seems that each diagonal is the previous diagonal plus the powers of two

@Ginger try closing and reopening the files in vscode

wait, lemme open the question

every diagonal has a different formula?

im assuming all the errors are from the vscode linter and not the gradle build?

@Seggan yeah, gradle seems fine

@Seggan restarted VSCode, errors are still there

Like if you look at 1,3,8,20,48 then 2,5,12,28, which are adjacent diagonals, the second one is the first one plus the powers of two

hrm

then vscode is bad with generated code :P

use intellij

bruu, I am real dumb, I literally copy my friends homework and cheat in exams

@AiraThunberg lmao welcome to the club

ty!

the language server output is full of cryptic, unhelpful errors :/

can you run tree for me again

@RydwolfPrograms thanks, I feel i got what you said, but still confusing of how it will help in those bumps

wait one sec

also have you tried importing it as a gradle project if vscode has that option

@AiraThunberg Well, if during the part where you're going upward you remember which numbers you're passing, you can add specific multiples of powers of two to find the numbers on the downhill part

I just did a full build and restarted the language server

let's see if it works now

@RydwolfPrograms which feels like it might be part of a solution, if not a solution itself

THAT FIXED IT

:D

yay

apparently all I had to do was delete and remake the build dir

classic jvm

Another approach might be trying to find a closed form formula for the triangle, and then a closed form formula for the bumping

I literally am so dumb, all questions I have even posted here are the inventions of my boy friend :/

I fail in math grr

like, I have never posted a question here on my own even, I take foreign help

huh???

Idk, I think the fact that you're coming up with sequences and trying to reason about them shows that you have the sort of mathematical curiosity that matters way more than whether you're naturally "good" at it :)

@RydwolfPrograms if nothing else works, just prove it with induction

@mathcat [classic] *vscode

the sequence is not even mine, my bf named it AG Series in his book here amazon.com/dp/… and I insisted to rename it to bumping series

glad he allowed it

btw, wierd, its not even in OEIS :(

The OEIS tends to contain series' people come across while doing more practical stuff

so, it is not suitable for OEIS?

You can probably find both the triangle's contents and the bumping itself as separate sequences on there

Speaking of maths, believe it or not, I blew my chances of qualifying for a national-wide competition.

I am sure, the bumping is not there

trust me, my friend might be the best mathematician in his country if he did not crank his research

like, he likes naming his findings on people

@mathcat I messed up the second question (out of three) and two other nerds got it right :/

so yeah

"AG Series" what he called it was nickname of his teacher "AG", I really begged him hours to name it Bumping Series

he even named a basic Biology formula after someone here

github.com/AitzazImtiaz/Imtiaz-Wajid-Genotype-Function

otherwise, he was gonna be a famous mathematician :/ in short, I wish he stops cranking

Ooh interesting, the indices of the bumping in the triangle, ignoring the actual value of the triangle's contents, isn't in the OEIS either

Here's some code to generate it

var rc_to_i = (r, c) => r * (r - 1) / 2 + c;

function* bumping_indices() {
    var r = 1;

    for (var k, i = 2;; i++) {
        for (k = 1; k < i; k++)
            yield rc_to_i(r++, k);
        for (k = i; k > 1; k--)
            yield rc_to_i(r++, k);
    }
}

wait, why you don't post it to my question?

This is just part of a solution

It generates the coordinates within the triangle of the bumping motion

here

it asks to write code that make the series

Oh neat. I can try that challenge later

I gotta do some school stuff tho

sure!

don't tell me that we all are nine graders :\

11th here (US system tho, not sure if that's different from the German one)

I am in Belgium :D

Oh right sorry lol

We have same system almost, we are offered a Fleimish Secondary Certificate in Class 12

@AiraThunberg Hey, I live next door

and it's wierd she is on a social platform

I have existed since the beginning of time

Does your aunt's cat do math? :p

maybe? if the above claim is genuine :P

If she does, I might not have detentions in Class 5 for exposing exam to school

leak*

@mathcat i once did the natgeo geography competition, and i nearly qualified for state had i not mixed up east and west

@AiraThunberg whoa there, that kinda language is only appropriate in a Serious Screenplay

@Seggan oof, sadge

@Ginger verdorie, mijn engels :/

huh?

granted, it's pretty easy to qualify for the state round in the maths comp I'm talking about

my english sucks XD

@mathcat you need to practice a lot tho

oh, don't worry about it

it was a reference to a certain book series

@mathcat dutch speaker?

@Ginger you talking about that book?

I'm not even a german native speaker lol, I just live in germany.

@Seggan how tf does Rol handle function calls?

like I am real sad, he is wasting his talent

@AiraThunberg Your English is pretty good IMO :p

@mathcat ooh, i thought you dutch

@RydwolfPrograms I barely learnt it in class 5 :P

imagine learning a,b,c in Class 5

as far as I can tell if the visitor sees an RPAREN while processing an expression rule it just interprets the contents as an expression, but your grammar file only has RPAREN in the expression rule in the context of a function call

I'm specifically referring to github.com/Seggan/Rol/blob/master/src/main/kotlin/io/github/… and github.com/Seggan/Rol/blob/master/src/main/antlr/…

ig nice job @Seggan and @Ginger ?

I kinda like the repo

oh that was made by Seggan, I'm just using part of it for my own project

slay seggan XD

copy cat above me

oh wait, I'm stupid

@Seggan nevermind lol