Shreyas Mididoddi

......uhhhh h..hi

all installed on my system btw

-> yay -Q | rg -i electron electron32 32.3.3-1 electron33 33.4.11-2 electron34 34.5.8-1 electron35 35.7.2-1 electron36 36.7.3-1

based

I use double emojis all the time 🙃😅

you use apt to install docker to run make and ship the image to AWS

so obviously they all do cloud things wdym

one day I'm sure

I actually got told by Vue core team people and it's actually not really more magic than any other framework

react use
svelte async
solid use resource

all work (***almost***) exactly in the same way

correct me if I'm wrong but in the pic that daniel shared the await just is regular await that suspends the current render until it is resolved but then it just returns a reactive value that is controlled by an effect so everything after the first await is just regular client side vue right?

okay this actually makes sooo much more sense now thank you

oooooo that is cool

ohhhh

so use fetch is a computed/derived value with some syntax sugar? is that right or does use fetch literally just returns a promise?

My question is like can I do arbitrary awaits that depend on reactive values and those statements automatically re run?

or maybe you do that differently in nuxt? idk enough about vue tbh don't mind me asking dumb questions😅

if I pass a reactive thing to use fetch will it refetch on changes to that thing?

Packages · UnJS A gateway to our galaxy of JavaScript utilities, libraries, and tools created to empower developers.

unjs.io/packages/

unjs packages are sooo fking amazing! Actively trying to make no dependency cross runtime js server tooling sooo easy.

Doing all the right things!! unjs is based

yep

srvx, nitro, h3 all of the unjs stuff is amazing. Love what you people built

nix fixes this

even tho it's not really competition, competition (even in ideas space) is such a beautiful thing

but all of them are converging and moving in the right direction which is awesome!!

only on paper and it's kinda painful to do any sort of cross runtime standard web server and we both know that lol

my parser does file parallelism tho when I hit an import statement I do resolve and lex/parse that in parallel (in a thread pool)

single file and single core

hmmm what does that mean 🤔 is there like a thing I can read line code or article or something?

I was just wondering if there's a special magic algorithm some dude discovered in the 1980s or something

Also it's getting harder to see what's slow cuz the time per call to nextToken is super super tiny

Lol absolutely not 😅 i just want to push this thing as far as possible.

Well yeah but I've already done that and the slowness is almost evenly spread throughout from what I can see. There are no hotspots

i tried to do simd but idk i couldn't get it to be faster than regular character by character approach

I need help with my skill issues

What is a fast algorithm to build a fast tokenizer/lexer? I am hand rolling a state machine with special case optimisations for identifiers and stuff

But I can't get it to go faster than 500 ish megs a sec what do

+1

oh wait missed the version bump part nvm actually

....are you a javascript developer by any chance?