#💽Programming Chat v2
1 messages · Page 51 of 1
very nice and simple
oh please I've had this fixation for a while but it used to only get expressed when there was a tornado near me and the weather people were doing non-stop weather updates
exactly.
rare point in my life where i actually have too much food
hm
now that we actually have a wildcard cert with all the right chains n all
im gonna see if i can finally distribute videos yet
(of course i havent added Range so i expect something bad to hapen)
don’t you need to reimplement smtp for the third time
why do you keep redoing it 😭
that’s not an answer
evil answer
yknow i plan to supersede the DNS stack in our server too
and THANKS TO YOU i actually know how
yes
which is basically just better Java
i will concede on that point.
wow really
yes
surprised
ihopeyoucrumplelikeapieceofpaper
Need to make you use Rust again..
thats what concede means fats for brains
need to make you use ada again...

too bad there’s no CLR analog to Kotlin
whats a clr again
Common Language Runtime
oh you mean the jvm
i mean for Kotlin/JVM there is obv
for kotlin multiplatform its a bit more involved
ask jetbrains to make a compiler for that
real except they already have married Java
iirc their compiler is written in kotlin...
Married the JVM*
idk what to tell you
Honestly they could make a Kotlin/CLR target
ok anyways I sleep
also ngl I do wanna see how fast I could make TLS work in Ada
if I’d bang my head in the same spots as Rust
using a premade library or the protocol itself
"rn"
idk maybe later
no you'll never want to implement tls in ada or rust or anything
i'd say maybe kotlin but 99.99% of the work is done for you by JDK all the way back in like
1.4
dare i say it the JVM has the best TLS stack out of any platform
speaking from experience
yes
more specifically the encryption part though
if you have little to work on (e.g. in ada) you're essentially gonna have to learn the entire field of cryptography
makes sense
C#’s doesn’t look bad either but I’ve never used it
looks similar to the JDK's tls stack
JDK's is a tiny bit contrived in how you import certificates/keys though
since it mainly relies on the JKS format
absolutely none as its not standardized in the language
library wise you have either bindings to openssl or wolfssl
hopefully it becomes standardized in the 2030s though
itll last
privatized into the only entity that uses it: the military
$3000 license to use it outside of that
it's always been public
the first compiler made for it was a university after all
then it was moved to GNU
got mailboxes working with authentication without needing to write a login prompt
even works on phone
ughh
gotta email IANA back on how our Assignee email is working soon
also gotta make a mailbox for iana.contact
Is this Kotlin or Ada
Kotlin
I’m CURRENTLY hyperfixated on a bayachao character and you’ll never guess who
kohaku
Good.
no that’s Pain
駅員みたい
That’s correct grammar
I know it is, but it’s really basic
I’ve got でも… 私はも言った, but I don’t know how to continue it correctly
trying to say “but, I said that initially too”
Maybe 「でも、もう言ったよ」
What’s the よ doing?
In this kind of construct, 「はも」 is incorrect, も replaces は completely
When used like that, it’s kind of asserting what was said
Translated it’s like “but I already said that, yknow”
That makes sense
“As well” …
Also like if someone is like 「これは何ですか?」you can say 「警察官ケモノですよ」with the よ kinda asserting that the new information you’re presenting is right
It’s like “you’ve asked for new information and I’m presenting it”
It can be used a few ways like everything in Japanese
But it generally has some kind of tone of asserting something
Adaは良いよ。
悪いよ悪い
😭
最悪だよ
I mean the “already” part is handled by the もう
Hmmm…
hold on gotta learn what adverbs are
this is probably gonna look real awkward buttt
wow smh
what!!
I’m gonna be honest I only know of ですか from
like, that game where you ask questions until you get what the people are thinking
except it’s about guessing hot dogs
I know か alone also indicates a question but I dunno where either are super appropriate
(also know of の at the end being more emotionally charged… thank you bayachao fox game)
immersion 💪
の at the end is a form of another grammar structure んだ which is broadly the “explanatory form”
it’s uhhh
Idk it’s a little tough to explain because it carries a lot of nuance and I’m not sure I totally understand it either
I only know two meanings of it, really
The possessive particle and, again, in more emotional contexts
Jisho says it can be used as a question/confident end interestingly
ペコペコ…
yeah that makes sense
like idk the best way I could explain it is that it adds "it's because" somewhere in the sentence but it's also more than that
so like
あついです - "It's hot"
あついんです - "[It's because] it's hot"
there's a ton of other nuances
so like if you were to say "sorry I couldn't [do something]" I could say なにがあったの?("[Is it because] something happened?")
in that case the nuance is like "I have some interest in why you couldn't [do whatever] and would like you to tell me what happened"
you can google "Japanese explanatory form" and you should get a lot of good answers
what
taking that as a no
what did you send them
":3"?
you should totally email that to an iana employee
"bro I just use thunderbird if ur 'custom' mail server aint gonna receive mail then ur not getting ur port"
LOL 😭
but it should receive mail from any mail server if they're properly operating on 587/25
SMTP is honestly the simplest protocol to implement (on par with HTTP)
just wait till we make our own DNS server and not have any form of contact with them for a week 🧌
what if I make a rust dns server faster
you wont.
oh chatgpt
kotlin or ada
mmmmmmno
I thought you were on a good streak with all this Kotlin you've been writing
like sure it's less ideal but at least it's not Ada
IT'S TRUE
its false.
true
ada == outdated
rust is basically just better ada
with actual heap allocation w/ automatic memory mgmt w/o gc
"actual heap allocation" "automatic memory mgmt"
ADA HAS BOTH
ok u have auto mem mgmt on the stack
heap allocation u gotta Unchecked_Free or whatever
maybe dont allocate stuff
have you thuoght about that
?
how do you do shared memroy in rust
procedure B allocates a new array or smth and returns it to procedure A
stack reference is invalid
we uh
okay so im trying not to laugh cause
teh serious answer to that is ada has a secondary stack for returns like that 😭
in rust we'd probably just uh
either return a &Vec which iirc is heap allocated
or return a Box which is 100% heap allocated
: <
調べたらマッマのiPadにアプリが入ってたものめちゃやり込んで全クリしたやつでした!!おもしろいですよね!!!!毎回一番奥にめちゃひまわり植えてその前にくるみ(?)置きます
#ほめて箱 #褒めて #homete_BAYACHAO
https://t.co/x6d5JMAPRV
vec is heap allocated
in ada/gnat vectors are "special"
also @spare quartz what happens if you have a vector (i.e., a dynamically sized array -- iirc these are "containers" in Ada) and you wanna grow it but it's on the stack and has variables allocated above it
that depends on the type of vector
elaborate
a vector with discrete types can be stack allocated, theoretically
actually no
they are
the exact semantics depends on how GNAT implements vectors but iirc functions return them on the normal stack: i'm not certain on how they're grown
for vectors with indiscrete/variable size types they're heap allocated with automatic memory management (for the vector itself)
since in ada the only way as a developer you're supposed to handle heap allocated stuff is access types which are manually handled by the developer..
but how does it know when to free the heap memory if there's no borrow checker :3
IF the vector isnt using an access type value type it might be handled automatically
Controlled types
then its up to you to handle it, duh
same thing in rust
it doesnt make sense to auto-free up a heap allocated object thats "global"
mm in rust you have to start dealing with lifetimes (probably) but the compiler handles it all for you
because the programmer is stupid
and will make memory mistakes
a fatal flaw of ada
ada memory mistakes can't cause errors though
me when I overflow ur buffer cause u compiled in release mode w/ opts which removes bounds checks
the worst that could possibly happen is memory that's never reclaimed which isn't something the language or any compiler can solve
it doesn't, actually
the compiler will only remove checks if you have formally proved they can't happen mathematically
even if you compile with -gnatp which suppresses checks it will force checks in your code where it sees fit
this is why kernel crashes can happen in ada with a barebones environment cause the compiler forced a GNAT check, and you guessed it: there is no GNAT implementation
but what solution is there to the problem you have given
if you heap allocate an object that is GLOBALLY managed
and moreso than tells you, automatically inserts memory freeing
it is impossible to automatically determine where to free memory
rust has done it
i need an example
in ada the only way to leak memory (in a "safe" way) is an access type which you are expected to know what you're doing with
it is not very often you use access types
I mean the simplest example is with a reference
not applicable
mm yes it is
it isn't
in ada "references" are not a thing and 99.99% are stack allocated/automatically managed
ok well in rust, other stupid things in ada dont exist so
if its not applicable then it's not a valid point for you...
reason i stated 99.99% of the time is cause the 0.01% is System.Address or using an access/alias
but access/aliases are automatically managed as they're tied to an object in a block, and System.Address is just directly modifying memory
ok fine
struct Stupid<'a> {
// the lifetime 'a is associated with
// instances of struct Stupid
// the lifetime annotation here means
// that field dumb lives for as long
// as the outer struct
dumb: &'a str,
}
this is used by the borrow checker
to know when to free things
where can you use this without having to wrap it in a Box or a reference?
you could also have a shorter lived lifetime given to dumb (idk rlly how but it's possible)
anywhere
are you sure?
what does it need to be accounted with to be used
either stack or heap
i mean code wise
because im not super familiar with how you code in rust
and you clearly can't use that as an object like in kotlin
and if it's "freed when the block ends" we have that in ada, return it in the secondary stack and initialize a variable with it
struct Stupid<'a> {
dumb: &'a str,
}
fn main() {
let bruh = Stupid { dumb: "lo" };
println!("{}", bruh.dumb);
}
it'll get freed by the end
yeah ada has this just fine
declare
Not_Stupid : String := "meow";
begin
Ada.Text_IO.Put_Line (Not_Stupid);
end;
well a &str was an easy example
but in general lifetimes are tags of information that help the borrow checker determine how long something lives
ergo when the memory can be freed
and probably where memory might need to be allocated
eg if it needs to be heap
formally ada does not define where things need to go, compilers can decide wherever (but we've spoken about this)
but RAII is very powerful in ada when determining when things should no longer exist
raii is goated
if im being honest
the main reason i havent been coding in ada recently is just
lack of tooling IDE wise 😭
well if you come to rust land...
vs code works but its inferior to intellij
and abandon everything that makes ada good?
noooo thank you
no u still get everything that makes ada good
nope
i get it in a weird package full of nonsensical syntax
AGHHJ
WHYY ARE JAPANESE PEOPLE SO WEIIIRD
a good package full of good syntax
yo you got any good syntax?
no, we got
YOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
struct Container<'a, T> {
value: &'a T,
}
impl<'a, T> Container<'a, T> {
fn new(value: &'a T) -> Self {
Container { value }
}
fn get_value(&self) -> &T {
self.value
}
}
struct Wrapper<T> {
inner: Container<'static, T>,
}
impl<T> Wrapper<T> {
fn new(value: T) -> Self {
let value_ref: &'static T = Box::leak(Box::new(value));
Wrapper {
inner: Container::new(value_ref),
}
}
fn display(&self) {
println!("{:?}", self.inner.get_value());
}
}
fn main() {
let wrapped = Wrapper::new(42);
wrapped.display();
let data = String::from("Rust");
let container = Container::new(&data);
println!("{}", container.get_value());
}
beautiful syntax
UGGGLY
it looks like perfectly good kotlin code after 500 years of bit rot
hold on
looks like perfectly bad kotlin code after 500 years of perfection
THINK RUST
THINK
WHAT WILL YOU HAVE AFTER 500 WINDOWS
class Container<T>(val value: T)
class Wrapper<T>(value: T) {
private val inner: Container<T> = Container(value)
fun display() {
println(inner.value)
}
}
fun main() {
val wrapped = Wrapper(42)
wrapped.display()
val data = "Kotlin"
val container = Container(data)
println(container.value)
}
gross kotlin
5,000,000 times more simpler and probaly just as performant
AHHHH!!! KEYWORD USED!!
with Ada.Text_IO;
procedure Main is
type Container (T : Integer) is record
Value : Integer;
end record;
procedure Display(C : in Container) is
begin
Ada.Text_IO.Put_Line("Value: " & Integer'Image(C.Value));
end Display;
type Wrapper (T : Integer) is record
Inner : Container (T => 0);
end record;
procedure Create_Wrapper (W : out Wrapper; Value : Integer) is
begin
W.Inner.Value := Value;
end Create_Wrapper;
My_Wrapper : Wrapper (T => 0);
My_Container : Container (T => 0);
begin
-- Using the Wrapper and Container
Create_Wrapper(My_Wrapper, 42);
Display(My_Wrapper.Inner);
My_Container.Value := 100;
Ada.Text_IO.Put_Line("Container Value: " & Integer'Image(My_Container.Value));
end Main;
even UGLIER ada
oh my god its so pretty 🥺
🤢
imagine being so old that you need to write out your keywords completely otherwise you get confused
nah
imagine being so old that you basically READ AND WRITE ENGLISH when you code
rust syntax is composed entirely of ascii :3
nope
yup
nope
proof?
smh
僕は人間の善性をそこまで信頼してないというか、ネガティブな意味ではなく"そういうもの"だと思ってるから(人間みんな大分不器用なので)仮にその手の良くないエピソードが事実としてあったとしても「まぁそういう部分もあったりするのかもしれんな〜」ぐらいに思ってる。
i think mr hoshikake is talking about you here
just admit ur lang is obsolete and inferior
what's one thing you use regularly in Ada that isn't in Rust
my internet
inferior
superior*
who needs em
mmm there's a raw_socket crate
CRATE.
at least one crate
if you delegate everything to your ecosystem
im sorry to say it but the JVM floors you and your entire community
and a similar trait for unix
drink once vomit anywhere
you're just mad you dont have your own interpreter.
why would I be mad about not being interpreted lmao
compiled to native code >> interpreted

kys ❤️
what does that even mean ??
Mesoscale Discussion 0431
NWS Storm Prediction Center Norman OK
1014 PM CDT Sun Apr 06 2025
Areas affected...Southeast AL...FL Panhandle...southwest GA
Concerning...Severe potential...Watch unlikely
Valid 070314Z - 070545Z
Probability of Watch Issuance...20 percent
SUMMARY...Locally damaging wind and perhaps a tornado remain
possible into the early morning hours.
DISCUSSION...Widespread convection is ongoing late this evening
across parts of south AL into the western FL Panhandle, in response
to an approaching mid/upper-level trough, and a persistent
southwesterly low-level jet.
Several small cells with occasional weak rotation have been noted
across southeast AL. These cells may continue to develop within the
low-level warm advection regime. Modest MLCAPE and weak midlevel
lapse rates will tend to limit updraft strength, but rich
boundary-layer moisture and favorable wind profiles (as noted on
regional VWPs) will support potential for at least transient
supercells, which may be accompanied by the threat of a tornado
and/or isolated strong gusts.
Farther west, multiple loosely organized clusters have occasionally
emerged from regenerative convection near/east of Mobile. Some
threat for locally damaging wind and a tornado could develop if any
of these clusters can become more organized and sustained as they
spread northeastward with time.
Watch issuance continues to be considered unlikely, due to the
expectation that the severe threat will remain rather isolated, but
trends will continue to be monitored for an uptick in coverage of
organizing storms.
..Dean/Hart.. 04/07/2025
...Please see www.spc.noaa.gov for graphic product...
why did you just send me an nws mesasge. .
:33
you are NOT "silly"
you are a college student
do your homework or whatever it is you do.
why don't you go watch your bayachao videos or whatever it is you do
we could collab on an lsp
or make a custom editor idk
poor man’s GNAT Pro
git
but like
you're gonan be coding in rust
im gonna be coding in kotlin
how do they join..
unless you'll be writing JNI
oh I figured you’d just use rust /j
you wouldn’t want to implement it in Ada?
you'd have an even harder time joining the two
since interop in rust isn't the greatest from what i know even if i export convention as C
you wouldn't write kotlin.
I have before…
laaammee
i mean if you're actually serious i do have some windowing work in place already
only for windwos thouggh
windowing for what
i dunno how lsps work 😭
giant spec
plus static analysis
This document describes the 3.17.x version of the language server protocol. An implementation for node of the 3.17.x version of the protocol can be found here.
i assume by giant spec you mean the protocol for microsofts lsp protocol rigth
yes
just usee this tiny incomplete thing im ade .. so easier..
except the
Peer A opens socket on port 30 (TCP)
Peer B listens to socket on port 30 (TCP)
part is deprecated since of the existence of BEG-DSCNP
virus
no .
json rpc is mandated by the lsp spec
horrible..
no email yet..
imagine they have emailed you but it bounced
it'd resend
the thing is we're not allowed to SEND email
at least not until we get our second server
theres only one issue with relying on WWW-Authenticate
browsers won't get the memo that two mailbox paths use different authorization headers
also emails are so fat
wait the iana contact email needs auth to send email to it?
no, just read it's mailbox
like you can't read my mailbox without the right password
@timid quartz
you're confusing it,...
I realize
oh cmon
i havent added any DKIM/SPF yet so
try not to abuse that too much yet..
should probably add a favicon soon but idk what to pick
check ur inbox
i think you're right..
/j
Provided to YouTube by Universal Music Group
Down's Syndrome Girl (From "Family Guy") · Cast - Family Guy
Down's Syndrome Girl
℗ 2010 Twentieth Century Fox Film Corporation
Released on: 2016-12-02
Composer Lyricist: Seth MacFarlane
Composer Lyricist: Walter Anthony Murphy
Auto-generated by YouTube.
LOL
you copy everything so what's new
this video download speed sucks ass :((((
Range hasn't been implemented yet so browsers are gonna get buggy responses
yeah cuz all ur stuff is buggy 
<he says, as he has no idea how the web works>
if i wasn't bound by the rules of this server
you're already maxing out the server storage space by being here..
belelelelele
talking bout making our own DNS with codev rn
as well as automation/succession of certbot
backup server being shipped tmrw
we'll be able to serve the entire eastern seaboard with this one
have you been sleeping for 14 hours
Yes..
@pastel tinsel
true
Astral Industries??
Bro thinks they’re BBI
I clicked on index.html
Give me back your Astro site
That isn’t mine.. but don’t you enjoy the new content
:3
cookie
Lmao
write kotlin and use Kotlin/JS
except not really because it's gonna output a bunch of weird-looking JS that will def make your teachers sus
is this not for a class
Kotlin/JS
what is this
intellij context menu
actually i forgot my computer education classes from 2015
its the menu thats apart of the title bar
oh gg
docker ❤️
useless
mmm nope
mm yes
very useful in fact
tool whos sole goal is to let developers stop developing their server applications correctly
horrible
false
tell me what else it does that isnt just letting you commit sins on file systems
ugh i feel like playing with the power brick suspended next to me but if i fidget with the led itll shut off my entire setup :<
easier deployment of server apps
because everything is set up correctly and bundled nicely in a container
so it's as easy as giving it environment variables/a config path and running it
just write your server to be a DS??
not that hard..
distributed system but SURE... yknow why
also easier to manage and restart containers if a service degrades/goes down
※うごメモからアクセスできますが非公式のスドメモです
▼お仕事のご連絡
[email protected]
▼ばやちゃおWebサイト
https://bayachao.com
▼Twitter
https://twitter.com/BAYACHAO
like sure you can do this with like idk a systemd service but the container is a nicer package
ALSO because they're containers they're sandboxed
me when I hijack your http server and get control over your entire system cause it wasn't sandboxed
really i just prefer more "traditional" distributed services with messaging
okay, which http servers?
thats what we have lines for
you can make "traditional" distributed services with messaging if you package them all in containers
you can't hijack GREEN/AMBER servers because those have no write access under your control
and BLUE servers are designed to exercise file system control through specified stores
see point about easier managing, scaling, restarting, etc. of services as needed and the setup being packaged and not needing to be redone each time
(which are inherently locked)
thats what we have libraries for, on the latter
on the left, just. design your servers right. its not that hard?
besides your point about bloat is irrelevant on linux
no
there's no virtualization done
you said wrapping a distributed service over a container
you're not really adding anything because it all still interfaces with the underlying kernel
you're wrapping a message twice essentially
mmm not really
[ container 1 ] --- vlan --- [ container 2 ]
is how it would look
the vlan is no different than a normal linux vlan (if running on linux)
okay, but inside those brackets, are the needless bloat
i was saying bloat towards a DS inside a container
inside those brackets are your services
not any regular server
refer to #1264496483957346346 message
and another question
what does docker really do if say
you only have one server and are expected to remain with just that server
you can run multiple containers on one server?
not my point
the point of deployment doesn't exist, and im back to square one
unless you mean one server for your distributed app
virtualization/sandboxing is effectively just letting you design your server badly
hard disagree
better to have a sandbox and not need it then need it and not have it
but when would you need a sandbox?
off the top of my head, barring any security vulns, at least to keep your app (config files, temporary stuff, etc.) organized in one spot
organization is a fair point
but ill admit im biased against it (i prefer no configuration files)
also the health checks and --restart unless-stopped flag in question
health checks as in?
env vars are also super easy
basically it executes a command at regular intervals to make sure your service is still running optimally
what the command does is up to the person making the app
what does it do if it reaches a failure condition
you can configure that
that just sounds like a cron job to me in all honesty
it is but it's organized within docker/podman/whatever you use
even more powerful when you use say docker compose which lets you define one config file to deploy an entire set of services
fail2ban moment
containerization for deployment was a response to the shortcomings of the lack thereof
it has its upsides and personally I prefer it to "normal" deployments
mainly because of the ability to manage, restart, scale, etc. as needed, the ability of things like docker compose to declaratively create a set of services to be used together, and because organizing everything into containers keeps the underlying base os clean
the organization also includes dependencies
somewhat related to this i was looking into transactional operating systems a while ago
very interesting design that i feel like should be expanded upon
current implementations are strange to use
should take a look at nixos and fedora silverblue
well ig those are "immutable"
but the updates are "transactional"
also enjoy how macos does their method of application storage
it's not as contained as what you've described but
it's a lot better than what's present on many linux distros/windows
it's similar
because the core app stuff gets kept inside the "app"
but it does sprawl out
whereas a container would keep everything inside the container
*unless otherwise allowed
lemme just find my notes for something
# conduwuit
services:
homeserver:
### If you already built the conduwuit image with 'docker build' or want to use a registry image,
### then you are ready to go.
image: registry.gitlab.com/conduwuit/conduwuit:latest
restart: unless-stopped
network_mode: "host"
volumes:
- ./data/:/var/lib/conduwuit
- ./conduwuit.toml:/etc/conduwuit.toml
environment:
# CONDUWUIT_SERVER_NAME: aerasto.xyz
# CONDUWUIT_DATABASE_PATH: /var/lib/conduwuit
# CONDUWUIT_DATABASE_BACKEND: rocksdb
# CONDUWUIT_PORT: 6167
# CONDUWUIT_MAX_REQUEST_SIZE: 20_000_000 # in bytes, ~20 MB
# CONDUWUIT_ALLOW_REGISTRATION: 'false'
# CONDUWUIT_ALLOW_FEDERATION: 'true'
# CONDUWUIT_ALLOW_CHECK_FOR_UPDATES: 'true'
# CONDUWUIT_TRUSTED_SERVERS: '["matrix.org"]'
# CONDUWUIT_LOG: warn,state_res=warn
# CONDUWUIT_ADDRESS: 0.0.0.0
CONDUWUIT_CONFIG: '/etc/conduwuit.toml' # Uncomment if you mapped config toml above
#
### Uncomment if you want to use your own Element-Web App.
### Note: You need to provide a config.json for Element and you also need a second
### Domain or Subdomain for the communication between Element and conduwuit
### Config-Docs: https://github.com/vector-im/element-web/blob/develop/docs/config.md
# element-web:
# image: vectorim/element-web:latest
# restart: unless-stopped
# ports:
# - 8009:80
# volumes:
# - ./element_config.json:/app/config.json
# depends_on:
# - homeserver
like
this is a docker-compose.yml file for conduwuit (matrix server)
oh right conduwuit is something i need to succeed later
ugh!!! where did i put my ufedmi notes!!!
good luck, matrix protocol is massive
a protocol is the same as a file format
just different transmission
can't find them
will need 2 rewrite. just an os idea i had for nanokernel which defers most of its power to applications and servers
(delegate would be a better word)
and shouldnt say os either, its a kernel..
I'd like to say
the largest section numbers I can see
for client-server
is 10.41.x
thats not that much
ok buddy
ive read rfcs yes
oh the ip rfc is way smaller than the client-server api spec for matrix
thats because IP is a single protocol out of the entire IP stack
it by itself isnt very special as a lot of the logic that goes behind it is router-specific
the tcp and ip rfcs combined are probably smaller
sounds like a self-inflicted hell to me
The client-server API allows clients to send messages, control rooms and synchronise conversation history. It is designed to support both lightweight clients which store no state and lazy-load data from the server as required - as well as heavyweight clients which maintain a full local persistent copy of server state.
API Standards The mandatory...
idk scroll it for yourself
old people from 1990 did
im on it rn
it LOOKS big
but still managable
im sure you wouldn't need all of this
im sure you would
you implement the big stuff first and then weed out edge cases
i will after i write this SSH client
do you want this to be on our gituhb or my account
uh idk beg is your thing so your choice
do you want it to be published on a maven
why does my opinion matter in this lo
becuse YOU'RE the commissioner.
ada or bust
a matrix server and not client just to be sure
yes a server not a client
preferentially kotlin
wow you've really fallen out of love with ada this is crazy
matrix looks full of encryption stuff which kotlin/java is much more developed towards
it is
TOOLING!!!
im sorry our lang doesnt have a billion femboys being mad on stack overflow and github for everything you have
RUST!!!!!!!!!!!!!!
EWW
https://github.com/matrix-org/vodozemac THE OFFICIAL MATRIX ENCRYPTION IMPL IS RUST
EWWWWWW
LETS GO
your langs could never imagine these kinds of Ws
says the furry
not a furry
are a furry
nope
covered in the black tar that is the jvm
alive more than ever
remember who you have to rely on for your compilation (LLVM)
id trust oracle hotspot with my life...
remember who you have to rely on for your compilation (gnu)
okay well
unlike you, adas compilation was actually matured with the fsf/gnu
this is why our code is superior
actually
i wonder what the first ada compiler was written in
EW
man
when our compiler manuals look like this you know its peak
holy crap
20 second compilation time for hello world
computers/compilers really have improved
MULTITHREADING SINCE THE START BABBYYY
Imagine having to stop your entire app dead in its tracks to clean up after the mess it made
actually not true anymore
and hasn't for a good decade
GCs encourage bad memory use patterns…
we've had many parallel GC designs over the years
most modern one is ZGC which is nearly infinitely scalable
GCs ARE UNNECESSARY BLOAT!!! 🗣️ 🔥
You are bloat(ed)
We’re finna trim you
im not an ssd.
So much text when I’ve already dropped you and forgotten you
"your"
This is also false just watch modded Minecraft load up you can literally watch the GC pauses in real time
SETL (SET Language) is a very high-level programming language based on the mathematical theory of sets. It was originally developed at the New York University (NYU) Courant Institute of Mathematical Sciences in the late 1960s, by a group containing (Jack) Jacob T. Schwartz, R.B.K. Dewar, and E. Schonberg. Schwartz is credited with designing the ...
they wrote it in SETL
gross
dudes using minecraft as an example to JVM GC technology 😭
it's the most widespread and well-known use of the jvm
it's also the worst example
well youre only as strong as your weakest link...
50% of the game is just native, and the parts that aren't are most likely for older JVMs like 8 whose only option is G1GC (or G1C? i forget the name..)
also the fact that most modders are just bad at coding
while your jvm program is sitting down doing its 50th gc pause in the last half-second, my rust program will already have finished
oh no it looks like your rust program panicked...
too bad my jvm program finished with flying colors
because all of its execution was safely threaded.
tell that to the RuntimeException
rust has safe threading
smh
nah
yep
your multiprocessing is weird as hell..
ever heard of "fearless concurrency"
☣️ concurrency
says the one who uses the language that automatically starts threads whenever you instantiate a package
wrong, actually
thats dependent on what the package does
(and really isnt a point against it! )
mm yeah it is
nope
EVERYONE
concept made up by the government to keep us in control
ok
considering that implementing olm and megolm in ada yourself may be slightly painful
you will be permitted to use kotlin
trol
i cant write microcode uipdates for my brain
you can put it under beg and publish it on your maven
cause ill probably still use conduwuit (rust 💪)
in whag
✦Story✦This is the magical world of Majirisia.
You are a summoner and you have summoned a demon.
You tell it your name in order to make a contract with it, but the demon won't tell you its name.
…
2025
yknow im 99% sure i can't send this here because of some of the dialog but...
its all in japanese nobody can understand it
NOOOOOOOOOOO
okay but
arren't you gonna get blamed for 9,000,000 server problems
like you were and ben last yr..
i blame rust
duality of man..
were you planning to use a java package
or were you planning to implement it yourself
im not authorized to use external libraries unless someone says i can..
you can idc
kades server for instance runs purely by itself
it would relieve the burden of implementing olm/megolm
but if you wanted or still want to implement it yourself, then use ada
i bet Sun already did it
actually yknow what
im not really even commissioning this
you said yourself that you wanted to "succeed conduwuit eventually"
so...you're under no obligation
okay butlike
eventually means whenever i want...
you're forcing me at gunpoint ..
yeah rela.y..
I was just saying
if you're so confident in making a matrix server
then do it in 2 weeks
but you don't have to
gunpoint..
ok listen
idc
even if you made it I probably wouldn't use it unless it was just that good
yes i knwo you dont care 😭
do u want to be forced to do it in 1
yes they do
i mean that depends
im sure i could get something FUNCTIONAL by that time
like being able to send messages
was thinking of getting a split ortho keyboard like that tbh
