#development
1 messages ยท Page 239 of 1
yes
so```sql
ON CONFLICT (location) DO UPDATE SET location=location
RETURNING id
{
result: [],
success: false,
errors: [
{
code: 7500,
message: 'ON CONFLICT clause does not match any PRIMARY KEY or UNIQUE constraint'
}
],
messages: []
}```
o wait
config_id is part of the unique index
very nice
works
thanks guys
whats the number 10 do, and what is function t?
i recommend making it a bit less magic-number'y and easier to read ๐
what if you wanted to change that value, do you use it in other places too?
?
dont look at my vars then
beautiful
XXDDD
as long as its fast its acceptable
ima be honest, these functions were made with help from chatgpt
what do you use them for
astronomy right?
ye
the second pic is for converting ecliptic coordinates into cartesian
interesting
ie longitude+latitude+distance -> xyz
whats ur game again
yeah ik
smol db
D1 gives you 5gb for free right?
yes
well I use workers paid anyway
ah
yes
thats why you index everything to infinity
this only reads 1400 rows
with 120k builds
ngl thats pretty garbage from D1 then
still way cheaper than serverless pg/mysql
ye
this needs to be in the cloud though
way too many people rely on this api
and hosts
with all the hype around edge, from what ui've seen its pretty lackluster
if you were using pg instead of sqlite you could code almost everything within the db tbh
ye however all serverless pg i found is 20โฌ/month/location
then return everything processed to the api
like wtf
I mean yes thats what im currently doing
however I cant find good providers everywhere
(and providers that have good cf routing)
not even needed
pg has that built in
ye i am using built in master-slave replication
manager-paid_worker 
mmmm
ohno
I still need to optimize this mess https://github.com/mcjars/versions-worker/blob/main/src/api/v2/build.ts
ooo somebody's copying bun ๐คญ
interesting
better-sqlite3 is cooked
yeah they literally copied its syntax for the most part
dont blame them because its good
It's good yeah
node is coping /s 
what's this, a line drawing function?
calculating the position of a planet's Lagrange points
:O
i mean, are they actually basing it off of better-sqlite3?
if so, thats pretty pog
... well its not rocket science....
-# oh, wait...
it's planet science smh my head
planet science is crazy
I don't exactly know but someone pointed out this is where it was added: https://github.com/nodejs/node/pull/53752
The design seems very similar to that of better-sqlite3
i went to read most of those discussions to try to understand what they are doing and it seems they are not following any particular lib, but trying to take points raised by other libs into consideration
there were talks of having both a sync and an async api for sqlite
and the author of better-sqlite3 recommeded them to expose the raw sqlite C api
but there are still a lot of issues for them to solve, like sqlite's compilation flags and extensions
so it seems they are going to try doing their own thing
I guess, this comment https://github.com/nodejs/node/pull/53752#issuecomment-2228852505 also seems enticing
From a quick test, compiling SQLite in Node with the same settings used by better-sqlite3 yielded roughly the same performance. There may be other things to optimize as well, but it's good to know we aren't doing anything too awful ๐.
ye i guess the author of the initial implementation did follow better-sqlite's footsteps
but lets see where it goes from here
It'll probably get optimized a lot
Follow your father's footsteps, C++
just-js had a lot of potetial, i wonder which direction that will take
What is that?
they moved to a new project called lo
its an experimental v8 runtime for linux only
ew
nodejs-light when?
when the author launched just-js, he did make a barebones webserver and added it to those big ass web framework benchmark competitions
it was the only js thing in the top 10
You mean just runtime? Not sure how V8 is related here
v8 is the js engine
just is like a barebones wrapper around it
like node but without the entire node api
Considering V8 supports all platforms they could probably add support for all platforms
yeah, it heaviliy interacts with linux syscalls
They seem to have builds for all platforms
besides 99% of use-cases will be linux anyway
Or at least Linux distributions, macOS, and Windows
it's JavaScript Jim, but not as we know it. :space_invader: - just-js/lo
yeah, lo was a change of direction from the original goals
but its still in very early stages
after the bun hype, its time to hype a new thing
:^)
Coming soon to a browser near you in 2089
indeed
I wonder who even uses Bun now
bun is still so broken.
the http lib polyfill literally leaks memory
they still havent fixed it properly
Did they fix the runtime running into a segfault every 2 seconds?
no
Amazing
still segfaults constantly
People really be hyping broken new stuff up like crazy
Reminds me of the VSCode killer stuff
Zed?
mid diff no mia ff 15
btw cloudflare d1 is suprisingly fast
I never expected serverless sqlite to be that fast
I mean
sqlite is slow
I didnt expect it to search a 800k row table thats 800mb in postgres within 100ms
UNINDEXED
sqlite is far from slow, it's just that it cant benefit from threads for write
it implicitly uses it
unless you declare a column as INTEGER AUTOINCREMENT PRIMARY KEY
in which case it overrides rowid
yknow whats weird
d1 just denies your request if it contains a ;
since it does not support multiple params
is it normal that this reads 2180 rows every time
DEB 18.07.2024, 20:09:38 Database Query
SELECT id, project_version_id, rehash
FROM builds
WHERE type = "SPONGE" AND version_id LIKE 1.11.2%
(rows read 2180, rows written 0)
INF 18.07.2024, 20:09:38 Found 90 builds for version 1.11.2 (SPONGE)
DEB 18.07.2024, 20:09:39 Database Query
SELECT id, project_version_id, rehash
FROM builds
WHERE type = "SPONGE" AND version_id LIKE 1.12%
(rows read 2180, rows written 0)```
its the total number of sponge builds
D1 sounds expensive
cheapest serverless db I know of
apparently LIKE will not use an index if the pattern starts with a digit
This constraint arises from the fact that numbers do not sort in lexicographical order. For example: 9<10 but '9'>'10'.
there is also the issue of case sensitivity, even if not applicable in this case
LIKE is case-insensitive by default. To have it use your index, you need to either make the index case-insensitive:
CREATE INDEX test_name ON test (name COLLATE NOCASE);
or make LIKE case-sensitive:
PRAGMA case_sensitive_like = 1;
also shuoldnt 1.11.2% be in quotes?
its a placeholder
ah
my logger replaces it for readability
you can also try this workaround
WHERE word >= 'search_string' AND word < 'search_strinh'
wtf
would WHERE version_id >= '1.11.2' AND version_id < '1.11.9' work too
that would be the same as LIKE 1.11.2%, LIKE 1.11.3%, LIKE 1.11.4%, LIKE 1.11.5%, ... LIKE 1.11.8%
dont think dbs accept semicolon separated queries anymore dure to injection reasons
or at least drivers dont
it works in lexicographic order, each character is converted to its char code, and each code is compared
oh wow it works perfectly
so "aaa" is smaller than "aab", and any more characters that are appended to that will always be bigger than "aaa" and smaller than "aab"
SELECT id, project_version_id, rehash
FROM builds
WHERE type = "SPONGE" AND version_id >= 1.11.2 AND version_id < 1.11.2.9
(rows read 91, rows written 0)
INF 18.07.2024, 20:34:14 Found 90 builds for version 1.11.2 (SPONGE)
DEB 18.07.2024, 20:34:15 Database Query
SELECT id, project_version_id, rehash
FROM builds
WHERE type = "SPONGE" AND version_id >= 1.12 AND version_id < 1.12.9
(rows read 499, rows written 0)
INF 18.07.2024, 20:34:15 Found 498 builds for version 1.12 (SPONGE)```
btw, if u want to match a pattern in LIKE but dont want to allow 2+ characters u can use _
like 1.__.3 would match any version that has major version 1 and build 3, and 2 digits for minor version
tho it requires the number of chars to be exact
do you want to specifically stop at .9? or you want to get all patch versions for 1.12.x?
all that start with 1.12
but its only normal semver in this case
then do 1.12 to 1.13
although there is a caveat
it will not work with 1.9 and 1.10 for example
because 1.9 is bigger than 1.10
the characters 1.9 are bigger than the characters 1.1 in lexicographiical order
let endTime = new Date().getTime() + duration * 60 * 60 * 1000;
let endString = `<t:${endTime}:F>`;```
why does this show me something utterly crazy
november 19th, 523523 or something tf
discorrd uses seconds
ya
js dates use ms
bomboclat
discord is dumb
tyy
discord doesnt care about ms!!!
just substring till the 2nd period
then cast it to float
oh wait nvm, 9 is still bigger than 10
huh tf
let endTime = new Date().getTime() + duration * 60 * 60;
let endString = `<t:${endTime}:F>`;```
pretty sure it's seconds rn
sir
getTime returns ms
omg
it's not like mojang will ever release minecraft 2
what if minecraft 0
nobody plays on alphas
but well, u can WHERE substring(version_id, 0, 1) > 0
to deal with 1.9 - 1.10 you will need to do 1.9 - 1.: to match everything 1.9.x
also mojang doesn't honor semver
they use it but it doesn't work in the same way at all
nvm there are snapshots
or i stick with tims solutions because I think thats enough

this only runs every hour
no need to optimize row reads to infinity
inb4 query takes 1h and 1 minute to complete so they overlap and crash the server
the what
yes
it takes quite a bit starting a server on each mc version of each software of each build ever
or just abuse the lexicographic order:
WHERE (a >= '1.9' AND a < '1.:') OR (a >= '1.10' AND a < '1.12')
= match everything from 1.9 until 1.12
:^)
wait how do i create an actual timer in djs?
setTimeout()
Like one of the formatting did that but i forgot it's format
setTimeou = setTimeout
or do u want the discord "in X seconds"?
yeah
arigato ox7d8 kun
what the fuck
great, now im reading that name in a japanese accent anime girl voice in my head
oeksusevandeeito-kun
first 4 rewritten ๐ฅ
you're hereby nicknamed "lime"
raimu-desu!
sasuga timu-sama
๐ญ
lmao loved that part
just watched it yesterday xDD
waiting for the day pnpm adds an option to automatically add the @types/ package when the package ur installing does not include ts types and you use ts
when they use actual code instead of random matrix shit in hacking scenes

probably hired a programmer just for this scene
as its the only scene that shows code
getKurata new Rectangle

SELECT jar_url, jar_size, version_id, config_id, config_value_id
FROM builds
INNER JOIN buildConfigs ON builds.id = buildConfigs.build_id
WHERE type = "VANILLA"
LIMIT 1
what happens when buildConfigs has no match
is config_id null?
lmao not only java, but swing
tho it might be some other framework too, idk what Kurata is
its probably part of the game that the code was about
How do you know its swing tho
Region
he's dealing with either swing or java2d in general
tho getInstanceKurata might mean that's a class from his game
is Kurata a relevant thing in the anime? is it a character?
Its not relavent at all tbh
The only thing relevent is his programming background
He literally dies 3m later in the anime
๐
His programming background makes it easy to pick up the magic in the world he is transported to though
is that death march?
yeah no, that's definitely swing
region might not be, but the rest is without doubt
actually, not swing, AWT
Knights & Magic
ah the one with mechas
no
not when I have 1s db latency
1s is too much
ok somehow I got it working first try
nice
the queries seem very fast tbh
2-3 business days per query
yea
hey, maybe a dumb question? but if I want to add/change some code to my bot that's already on top.gg, and restart it -- will current/future users see these new changes?
on discord
yes, the bot is just listed on topgg, it doesn't "clone" your bot
all processing is done on your side
Looks like you fucked up
idk how
@sharp geyser
dockerfile
FROM xxx as xxx
or
FROM xxx AS xxx
yea capitalized
sorgy for ping
anything dockerfile the keynames will always be capitalized
it worked before
so im confused why this is happening
i didnt change anything and now it's ๐
okay... downgrade to 21 worked
node devs did an oopsie
do they even test their versions ๐
push now fix later
its not LTS yet
ya
it will be in a couple months or so
but still 
Once its LTS you know its stable enough
Anything not marked as LTS is usually still in the works
Meaning bugs are expected
You can't expect them to have EVERYTHING working
Nodejs is a vast ecosystem
Granted even LTS still has bugs, but they are mainly flukes that are happenstance
of course i can, node is supposed to be perfect
Right
As if a big ass runtime that adds new features every major update is going to be stable right out the box
๐
Most of those new features require other things being reworked
Which just so happens to cause bugs
ngl tim
make your own runtime
you have the knowledge
says who
its all superficial knowledge
Buddy
does your <@&1016059109130375168> role blacklist you from all channels but #development? xd
Your knowledge is immense
my only deep knowlege is how to sleep and eat
it does but i've hacked it
:^)
Wait
Did they actually lock you to development only

We been joking about it for years
nah xd
it finally happened
Im honestly tempted to work on my website a bit
but I dont have the patience
i figured out what im gonna do to make this exercise that cant be easily solved by chatgpt
I want to try and make my portfolio again
but do I have the patience to sit there and do it
when I can watch anime instead
gonna make it use mapbox api v6, which is new, and chatgpt doesnt know about it, it only knows about v5
i hate designing
I hate twitter rn
I cant go on there without seeing some shit about donald trump getting shot
bing ai might know about it
has access to internet
so it can pull documentation
but it struggles with complex enough questions
it starts to outright make up parameters
("hallucinations" is the buzzword for this)
nobody told me to check bing ai
so i wont
:^)
Nah, run your own ai.
Custom RAG, so you can just make it check the top 10 results on Google instead of the first 2 or 3 ^-^
But yeah, bing does great with it.
what src/xxx folder should I use for classes
src/classes 
/tmp/src/classes
/dev/null
root@google.com:/var
could have downgraded to 22.4
easier to change a 22 to a 21 in docker
:^)
How would I de-bounce re-rendering a component?
I have a form with a field with an image URL which I'm displaying in a child component.
My main issue is that the image URL is fetched EVERY time a character is typed, so someone typing a URL would send a request every time a character is typed.
"use client";
import { ChangeEvent, useState } from "react";
import DisplayFormData from "./someRenderedComp";
export default function Home() {
const [formData, setFormData] = useState({
description: "",
image: "",
});
const handleDescriptionChange = (e: ChangeEvent<HTMLInputElement>) =>
setFormData((prevState) => ({ ...prevState, description: e.target.value }));
const handleImageChange = (e: ChangeEvent<HTMLInputElement>) =>
setFormData((prevState) => ({ ...prevState, image: e.target.value }));
return (
<main className="flex min-h-screen flex-col items-center p-24">
<form className="flex flex-col">
<label htmlFor="description">Description</label>
<input
name="description"
value={formData.description}
onChange={handleDescriptionChange}
placeholder="some fancy description"
className="text-black"
/>
<label htmlFor="image">Image</label>
<input
name="image"
value={formData.image}
onChange={handleImageChange}
placeholder="imageURL"
className="text-black"
/>
</form>
{/* Dont re-render unless the user hasnt typed in the 'image' field for 500ms */}
<DisplayFormData {...formData} />
</main>
);
}
"use client"; // in my real example this has to be a client component, here it doesnt.
import React from "react";
function DisplayFormData({ description, image }: { description: string; image: string }) {
return (
<div>
<div>{description}</div>
<img src={image} alt={"Cant load"} />
</div>
);
}
export default DisplayFormData;
React - nextjs ๐
(ignore the bad practices, I let AI write most of this becuase its just an example)
useDeferredValue seems okay, but it has limited control.
I wouldn't mind using it, but it doesn't appear to have any way to control it.
setTimeout
only update inside the timeout
whenever they type, cancel current timeout and restart it
setTimeout would be a throttle.... never mind, yeah
Can I do that from within the child component?
I'm not exactly sure how re-rendering works.
Honestly, I'm just going to try just to learn.
Thanks
setState is what causes the update
simply make a timeout that calls it, instead of calling it directly
in your case setFormData
hmm, I'm doing something wrong.
let timer: NodeJS.Timeout | undefined;
const handleImageChange = (e: ChangeEvent<HTMLInputElement>) => {
console.log("Image change: ", e.target.value);
clearTimeout(timer);
timer = setTimeout(() => {
setFormData((prevState) => ({ ...prevState, image: e.target.value }));
}, 500);
};
It almost feels like I need a seperate hook to do this right
I just typed a, waited, then bc
Or a buffer for the input.
But then it would look awful, the input wouldnt show typed characters for 500ms
Something like this
But this wont work either,
const useDebounceInput = (ms: number, value: string): string => {
let timer: NodeJS.Timeout | undefined;
const [debouncedValue, setDebouncedValue] = useState(value);
useEffect(() => {
clearTimeout(timer);
timer = setTimeout(() => {
setDebouncedValue(value);
}, ms);
}, [value]);
return debouncedValue;
}
But timer will be lost each re-render
why not use https://www.npmjs.com/package/debounce
import debounce from 'debounce';
const [value, setValue] = useState()
const debouncedSetValue = debounce(setValue, 500)```
I try to avoid libraries unless I understand how they work
Especially for relatively small things like this
this is actually quite complex in react
since there is no proper proper way to do this
React sometimes 
I don't think this will work, will it?
When setValue is called it will re-render after debounce is called, destroying the timers it uses too.
There are react specific libraries for it
I'll try it ^-^
so: No inputs are received at all
export default function Home() {
const [formData, setFormData] = useState({
description: "",
image: "",
});
const debouncedSetValue = debounce(setFormData, 500);
const handleDescriptionChange = (e: ChangeEvent<HTMLInputElement>) =>
debouncedSetValue((prevState) => ({ ...prevState, description: e.target.value }));
const handleImageChange = (e: ChangeEvent<HTMLInputElement>) => {
debouncedSetValue((prevState) => ({ ...prevState, image: e.target.value }));
};
return (
<main className="flex min-h-screen flex-col items-center p-24">
<form className="flex flex-col">
<label htmlFor="description">Description</label>
<input
name="description"
value={formData.description}
onChange={handleDescriptionChange}
placeholder="some fancy description"
className="text-black"
/>
<label htmlFor="image">Image</label>
<input
name="image"
value={formData.image}
onChange={handleImageChange}
placeholder="imageURL"
className="text-black"
/>
</form>
<DisplayFormData {...formData} />
</main>
);
}
I goofed around with it for awhile
I think you are understanding wrong
Always ๐
you need to use a variable for the text
it's cuz when the timeout runs, it'll grab the stale value
make smth like let val = "", then update it normally
inside the debouncer, you get that variable's value
Outside of handleImageChange right?
yes
Yeah, okay I think I understand it.
It doesn't work great, but it works.
in what way?
No matter what, this system inputs wont be shown in the field until 500ms after the user stops typing.
If I just use:
let timer: NodeJS.Timeout | undefined;
let value: string = "";
const handleImageChange = (e: ChangeEvent<HTMLInputElement>) => {
value = e.target.value;
console.log(value);
clearTimeout(timer);
timer = setTimeout(() => {
setFormData((prevState) => ({ ...prevState, image: value }));
}, 500);
};
It will only ever take the last character typed
If I wanted to fix that I'd probably need a second variable to track what was there before the debounce, and add only things after that to value.
Its just fiddling with the logic.
I think I understand enough to get to a working example, or at least struggle with it until I do ^-^
it's annoying to get right, but aint too hard
usehooks-ts might have something for this
not sure if it's your use case
In the end, I'll probably use https://www.npmjs.com/package/use-debounce
A hook just seems like the best way to do this.
But I'll look into usehooks-ts too
What are you trying to do?
i think ive done this sort of thing but in svelte and its very similar
do you want to wait before sending a request while the user types?
If I was smart I would just make it a modal for images
I'm not sending a request, just updating the value inside an image element. Which auto sends a request when it updates.
same thing basically
Use debounce works fine, I still want to implement it myself in the end
But for now, I at least understand it enough to do so
youre doing it wrong
export default function Home() {
const [formData, setFormData] = useState({
description: "",
image: "",
});
const [image, setImage] = useState('')
const debouncedSetImage = debounce(setImage, 500);
const handleDescriptionChange = (e: ChangeEvent<HTMLInputElement>) =>
setFormData((prevState) => ({ ...prevState, description: e.target.value }));
const handleImageChange = (e: ChangeEvent<HTMLInputElement>) => {
setFormData((prevState) => ({ ...prevState, image: e.target.value }));
debouncedSetImage(e.target.value)
};
return (
<main className="flex min-h-screen flex-col items-center p-24">
<form className="flex flex-col">
<label htmlFor="description">Description</label>
<input
name="description"
value={formData.description}
onChange={handleDescriptionChange}
placeholder="some fancy description"
className="text-black"
/>
<label htmlFor="image">Image</label>
<input
name="image"
value={formData.image}
onChange={handleImageChange}
placeholder="imageURL"
className="text-black"
/>
</form>
{image /* use this for the actual image */}
<DisplayFormData {...formData} />
</main>
);
}
youve tied the input to the formData state
That's a controlled input
so until that updates the text that you type wont reflect until that timeout runs which then sets the data using setFormData
you need a seperate state for the debounced value
this for example
Isn't two states just causing multiple re-renders?
But a hook wouldn't? Idk do hooks cause re-renders?
first one will always keep up to date with the formData, the second will set the typed in text after the timeout, which is tied to the image src
yes, one while typing, one after 500ms
tbf you could also do this with just one state
the formData state isnt strictly needed since you can detach it from that and just get the value from the onChange event anyways
faster as well
but if you want to do it the "react" way, sure
gotta save that 0.05ms!
๐ my render is way slower than that right now.
Quite a lot is going on behind the scenes
Which is part of the reason I am debouncing
I'm also using react-hooks-form for their isolated rendering
^-^
Trying to implement all of the custom elements Discord supports, and user facing variables.
As usual, this whole thing is too much work.
Auto complete almost killed me too.
๐
react isnt known for its remarkable performance either
tbf if youre using any framework youll get some performance loss anyways
if you dont want any you have to use vanilla js
but that doesnt matter too much you just need to make sure long or blocking code is async where possible
so it doesnt delay any painting or ui
use react if you don't care that much
use solid if you do
i don't completely get why all the blame is being placed on crowdstrike
like okay they published a bad driver
but shouldn't there have been checks/automatic undo by Microsoft on clients?
Solid isnt as relevant anymore
It had hype when it first launched
but svelte and react are still the most used along with vue
vue is just horrible to read
vue is similar to svelte
vue is just horrible to read
<script setup>
import { ref } from 'vue'
const count = ref(0)
function increment() {
count.value++
}
</script>
<template>
<button @click="increment">Count is: {{ count }}</button>
</template>
Doesnt look too bad
Similar to svelte in a way
svelte >
that looks very similar to svelte 5
apparently the svelte team isnt completely happy with the svelte 5 syntax either
they agree its very react-like
but they had many many other potential syntaxes proposed but none of them really worked out
i still think theyre taking away what made svelte svelte, with svelte 5 i might as well use vue.js honestly
@quartz kindle worker deployment to production has been without issues so far, generally faster too
ref incremental?
where the states at
thats the state
value++ will cause anything using count to rerender
or well any mutation to it
thats the point of ref yes
ref is similar to reacts useState
except it doesn't have a getter and a setter separately
its combined
what were they even thinking
did they think const [something, setSomething] = useState(); was a good ieda
uhm
[link](url)
not [link]('url')
@eternal osprey
why are refs signals
refs are supposed to not rerender components on change
Ask vue team
arigatoo
Also, in react that is the opposite
Refs do cause re-renders

Which is why when using state management libraries you pass a ref
refs dont rerenderrr ๐ญ๐ญ
So you can pass it down to all the children components and re-render as needed
Yes?
They literally do
else it wont re-render any of the changed values
Then dk what redux does
As you pass in a reference of the state when using it
weird stuff is what it does โค๏ธ
redux makes heavy use of refs
zig is so cool. i can build the binary on my windows machine for arm linux so easily. now my "builds" take less than 30s
.PHONY: build-DiscordSignatureVerifyZig clean
build-DiscordSignatureVerifyZig:
@if [ ! -f bootstrap ]; then \
echo "Bootstrap binary not found, building..."; \
zig build -Dtarget=aarch64-linux-musl -Dcpu=baseline -Doptimize=ReleaseFast; \
cp zig-out/bin/bootstrap $(ARTIFACTS_DIR); \
else \
echo "Bootstrap binary found, skipping build"; \
cp bootstrap $(ARTIFACTS_DIR); \
fi
clean:
rm -rf zig-cache zig-out
Okay
is the only good thing about zig is that it builds fast
rewrite chromium in zig
zig offers very little other benefits from what i can tell
ig if you want to write something in C use zig
because its the "better c"
other than that idk what else it is useful for
zig is really good for cold starts compared to Rust. rust inits a lot of stuff for the borrow checker
they both compile to binary
so they both are good for lambda
C, C++, Java, And basically any other compiled language then
I guess they all are alike
Java can compile to binary
C++ has too much junk
right
makes it a bit slower for cold starts
the binaries aren't
The only thing they share is some similarities with syntax
zig makes faster binaries for the lambda environment
yeah, rust and c++ both are slow for lambda
trying to compare zig and rust is the dumbest thing anyone can ever do
are you talking about AWS lambda?
yeah

cold starts and all
So you are comparing two languages that are unrelated based on one service hardly anyone uses
Well if zig is better for using aws lambda then use it
i am
but comparing zig and rust based on that alone is idiotic at best
rust has it's use cases too
but for aws lambda zig is better
That's why every month AWS Lambda has over a million monthly active customers who generate over 10 trillion invocations.
lol
A million is nothing
Also, AWS lambda sounds useless
Literally sounds like cloudflare workers
it has use cases i promise
it compiles to bytecode which can then be executed by the java virtual machine
there are tools that will make binary from java
Well yea normal java
but there are other jdks you can use to compile to binary
by that logic i can say python is a compiled language too
howd you know
the binaries that you get from java and python would be even slower than rust
do they use the blazingly fast โข๏ธ LLVM
they use cranelift

blazingly slow ig
likely uses cranelift
they wrote their own toolchain in zig
lol
yep, and it(stage 1 compiler) started off using llvm
just an update, i had one of my staff make bingo balls for it ๐ฌ they look good!
@real rose
a bit of a contrast issue, make the black smaller or the white bigger
iโm kinda scared to ask her ๐ i see what youโre seeing tho
you should come play sometime man! youโd be proud of me
/clickwar
damn it
Hey tim
do you know how to read barcodes programatically
I dont necessarily want to use a lib
and all I get online is apps for it

just ask her if she can improve the contrast
nop, never tried that
he does know it
I knew it
itโs somewhere in his brain
he lied
he just has to find it behind the rest of his knowledge
cant believe Tim lied to me
but whatever you asking, itโs somewhere in his brain
here's some source code for different OSs
https://github.com/wildabeast/BarcodeScanner/tree/master/src
Official repository now at phonegap/phonegap-plugin-barcodescanner. - wildabeast/BarcodeScanner
it's super old though
ew java
Not that badly
lol

im managing to partially decode these damn proto buffers
still a god damn awful format
imagine
protobuf encoded barcodes
seriously, why do we not have encrypted barcodes/qrcodes
sounds like something that absolutely should exist
why not
You know how painful that would be for businesses
Not to mention fact checking
You'd have to have the key to decode it to know what it is
exactly
meaning businesses can scam the fuck outta you
Oh what you bought is a bar of soap????
nah its specifically so you can print barcodes and be sure only your target recipient can scan them
Nope it turns out its actually a rolex
I mean sure
But still a bad idea
i mean
There are standards to barcodes so you can track the product down to who made it
qrcode
barcode barcode ye, but i mean more the general printed code tech
like all the qr code variants
brb encoding all my passwords into an encrypted qr code and tattooing it on my arm
ye
changing subjects
i find this amusing
CrowdStrike Engineering has identified a content deployment related to this issue and reverted those changes.
Workaround Steps:
- Boot Windows into Safe Mode or the Windows
Recovery Environment- Navigate to the
C: Windows/System32|drivers/CrowdStrike
directory- Locate the file matching "C-00000291*sys", and delete it.
- Boot the host normally.
also, why brazil was not affected
wtf is crowdstrike
you didnt see the whole outage thing?
they are like a security company responsible for updating shit for other companies
they pushed a bad update and broke millions of pcs, including airports, banks, and shit
yup
but nah
I didnt hear about the outage
All I know is https://clownstrike.com exists
wasn't the binary a bunch of 0s?
Crowdstrike is known for producing antivirus software, intended to prevent hackers from causing this very type of disruption.
lmfao
lmfao
also, i found this comment on reddit
Fun fact: I'm a former Firefox dev. The leading cause of headaches was anti-viruses that just linked themselves to Firefox and started doing arbitrary things in memory, instead of using the APIs dedicated to let anti-viruses do their job properly. In my experience, all the crashes were attributed to Firefox by users who (of course) had no way of knowing better.
So this fiasco feels extremely familiar.
Perhaps now people will start being cautious about security software and realize that some of them are actually more dangerous than the harm they're supposed to avoid
very true
the problem was actually the code that referenced this code didn't check for null before dereferencing
null causing all the problems again
that sounds like every developer ever
I never check for null
I just let shit happen
I go with the flow yknow
if im developer and i submitted a new bot is it going to be quick than before or the same ?
Depends on the queue size
If you are asking if you get special treatment because you already have a bot on the site the answer is no.
You wait just like everyone else
should be the latter
as in performance issue?
YOOO that's so cool. Thanks for the update 
me when I deref 0 in my kernel driver: page fault
thing is if it wasn't for windows following good osdev practices this bug probably would've went unnoticed
the driver is still bound to page tables so when it tried to deref 0 that area of memory was unmapped which caused the blue screen
otherwise it would've just allowed it
since well the driver runs in kernel mode (probably) so it can do anything
including deref 0
@green kestrel why are so many people running D++ on windows @_@
why not
possibly but not necessarly, dropped frames can happen for a variety of reasons
wanna see the funny
yes
xD
why wouldnt they
its always linux this linux that
maybe some people want their trusted corporate overlord microsoft to run their app
ah yes, and also hire crowdstrike to manage it
hot
because windows is the most popular OS
most bots are run directly on peoples desktops not on a VPS
(unfortunately)
can tell that from what percentage of bots are offline when reviewed here on top.gg
is that a bad thing because i use windows
if it works for you
im planning on getting rid of windows 10 on this laptop and moving to kde
im fed up of upgrade to 11 by buying new hardware nags
you can also go w11 LTSC
iirc they removed those checks there
understandable have a nice day
i do actually use bitlocker and stuff, it cripples them
i mean, is it really crippled when all the stuff it removes is useless?
bitlocker isnt useless
it is for me :^)
so youre ok with id someone steals your device, they have all your data?
sure
do you use online banking? buy anything online?
not on the pc no
but the accounts you pay for are on your pc?
some i guess, like facebook and shit
but my pc doesnt even have a password
it has literally zero security on it
youre a hackers dream
because i dont care :)
its because of people with machines like yours that stuff like crowdstrike exists.
its my personal machine, not a business machine tho
you not caring affects everyone else
i dont need security on my personal stuff
you can easily become part of some botnet, or otherwise end up participating in things that are detrimental to the stability of the internet as a whole
yeah im not that dumb
ok! then leave your front door unlocked each time you go out ๐
i dont run random software downloaded from strange websites
lmao
thats not really a good analogy but ok
irl houses dont have firewalls and network systems in front of it
sure it is
the stuff on your pc has value, the resources of your pc have value
just as your tv, your phone, your clothes etc have value
no your house has a lock, and a door, and walls
the only value the pc has is resale value for the thief
no
nobody cares about data
thats an ignorant assumption
i live in brazil
your pc still can run code
thats what a haker wants it for
they dont give a shit about what you do on facebook
why would they steal it and then run code on it
they want to make it part of a botnet to do criminal activities, or to mine coins, or send spam
they only steal to sell
no, they dont need to STEAL it
beause you dont encrypt it, you dont password it, you have left it wide open
no, but not passwording and encrypting it does weaken it massively, it only takes one zero day remote exploit, and there are no protections on the accounts
i bet you even run as administrator
yes i do
smh
its my machine, i own it
i dont need to have any software running on it with higher permissions than mine
yeah good luck hacking it then
do you install ALL windows updates the day they are released?
no
then how can you say youre secure and wont get hacked
i dont have hacker paranoia

i've literally never had any problem
i've had wayyyy more problems with windows update and with microsoft shit
then dont complain the day someone takes down all your bots and steals your token, from your un-updated dev machine via a key logger or trojan
yeah never gonna happen
Can I make it happen sir
making a note of this conversation for later
lmao
this kind of paranoia is what makes machines swlo af
antiviruses cause more problems than they solve
all security does
i dont run an antivirus
the whole crowdstrike problem is proof
or security software
but i do have a firewall (hardware), i do have a password, i do have my drives hardware encrypted, and i dont run all my stuff at highest privilege
explain to me how does hdd encryption protect from hacking
because an intruder cant just read raw sectors off your device
bypassing the OS security
if they hack into the machine, the machine is already running
thtas one reason
so the hdd is already decryteed
yes, but they dont have the key, just because its running
the key is stored in the TPM
they have access to the running machine
why need the key when the hacker can run remote commands on the already running machine?
because youre not running as administrator right? so they can only run commands as YOUR user
which means they cant read or change those OS files that arent accessible to your user...
thats just user privilege, not directly related to hdd encryption
this is the default in linux, NOBODY runs everything as root
if they get admin access, they can use the running machine anyway, regardless of hdd encryption
still
encryption costs you nothing in terms of cpu
it doesnt slow down your pc
you should use it if you have it
i costs a lot in terms of being recoverable
a lot of noobs especially in IT run all their applications under root because its easier
but its the worst thing you could do
a damaged encrypted drive is hell for data recovery
whats your facebook password, an empty string?
i've had to deal with that way mroe times than with hacking
if its damaged, youre fucked anyway
i think youre confusing hardware encryption that is part of the SSD firmware, with shitty software encryption
i would never use hardware encryption either
the only thing it protects again is from unauthorized physical access
it probably has 50 different backdoors
and if anything bad happens, yeah good luck getting your data back
my bitlocker doesnt use the tpm, only a strong password i type in on boot
windows 11?
yeah but its because i disabled it explicitly
btw if any of you have a phone youre basically forced to use all this security stuff you all profess to hate
well, google did try to encrypt my phone, i pressed no tho
thats why it kills me when i see people using their phones for OPSEC of any type
unless its some off brand with a custom OS you cant trust it for anything
how do you unlock your phone
my phone didnt even have a lock screen for a long time, i was forced to add one because of an app that doesnt work without it
so if someone steals your phone theyd have had everything?
all your accounts, your photos, possibly a 2fa app?
not right now because it has a lock screen
before they could yes
thats awful
lucky you put a lock screen on it now
ever had your phone stolen? i have
i have yes
and you know what they do?
they instantly shut it down and remove trhe battery to prevent tracking
then they facgtory reset it
and resell
i know what they did with mine, because i got it back ๐
the tim we see here now might be some random Brazilian thief from having his phone stolen




