Devansh Jethmalani

[Devansh Jethmalani @devanshj__]

wrote something short https://t.co/LBD7nlE4rS

[Devansh Jethmalani @devanshj__]

an alternative to lifting state up in react https://t.co/vDfiYhY94t

[Devansh Jethmalani @devanshj__]

People say TypeScript is unsound.
No.
TypeScript just assumes you know what you're doing.

[Dmitrii Kovanikov @ChShersh]

People say C++ is unsafe.
No.
C++ just assumes you know what you’re doing.

[Devansh Jethmalani @devanshj__]

.@infinterenders a react paglu like you might like these posts of mine...

https://t.co/bMTXZld4nI

https://t.co/ePzK3j3XTa

[Devansh Jethmalani @devanshj__]

@techsavvytravvy @typescript It'll work if you explicitly type `utils`: https://t.co/MuIpFcIqqb... which is silly hence this feature request: https://t.co/RcUAJURxFx

[Devansh Jethmalani @devanshj__]

New experimental PR!! 🤭 https://t.co/iiKbGb8DfV

[Devansh Jethmalani @devanshj__]

ps: this is a hint to solve this: https://t.co/mAuJC73M88

pps: `distribute (x) { y }` transpiles to `{ y }` so it's an erasable syntax.

[Devansh Jethmalani @devanshj__]

TypeScript Challenge!!

1. is this code safe to compile? (TypeScript says no)

2. if yes, what is the best way to change the code so that it compiles? (without changing the logic obviously)

https://t.co/5aaE7Pfv9R

[Devansh Jethmalani @devanshj__]

ok ryan but I like my types precise 🤭

[Ryan Cavanaugh 👉 searyanc.dev on bsky @SeaRyanC]

Is the return type here:
{ a: "x" | "y", b: "x" | "y" }
or
{ a: "x", b: "x" } | { a: "y", b: "y" }
?

Both are correct, but the latter is more precise. But the only way to get the latter is through combinatoric explosion

[Devansh Jethmalani @devanshj__]

3. bonus: if yes, what changes should be introduced in the compiler/language to make it more "complete" and not error

[Devansh Jethmalani @devanshj__]

TypeScript Challenge!!

1. is this code safe to compile? (TypeScript says no)

2. if yes, what is the best way to change the code so that it compiles? (without changing the logic obviously)

https://t.co/5aaE7Pfv9R

[Devansh Jethmalani @devanshj__]

Never miss a chance to be a TypeScript reply guy 😌

[Devansh Jethmalani @devanshj__]

@fernandorojo I think you can simplify the props type to `Pick<ViewProps, "children" | "className" | Extract<keyof ViewProps, `aria-${string}`>>` and then you don't need that `AriaKeys` type

[Devansh Jethmalani @devanshj__]

@fernandorojo I think you can simplify the props type to `Pick<ViewProps, "children" | "className" | Extract<keyof ViewProps, `aria-${string}`>>` and then you don't need that `AriaKeys` type

[Devansh Jethmalani @devanshj__]

jcalz will be happy https://t.co/Z3gjA1DGTj

[Devansh Jethmalani @devanshj__]

onto my next pr... any guesses what this is / which issue this is going to solve? 🤪

[Devansh Jethmalani @devanshj__]

now I realise it's pretty ironic that I say "not the intended usage of existential types" in the comment 😭 but I had to or else I would have been dragged how it's a bad idea xD

[Devansh Jethmalani @devanshj__]

It's so cool how TypeScript has enabled non-type-theory programmers (me included) to come up with features that are well understood in type-theory... For example this person wants "anonymous type parameters" and that's exactly what existential types are! https://t.co/KjxXvsMgjO

[Devansh Jethmalani @devanshj__]

good thing is now I can verify my typescript PR behavior against purescript...

(i used to think what I have is universally quantified types and that `<T> ((t: T) => T` should be same as `<T>(t: T) => T` but that's not the case)

[Devansh Jethmalani @devanshj__]

okay I figured out how to write existentially quantified types in purescript... and look typescript is much better haha

also that means my types are existentially quantified which i was unsure about

[Devansh Jethmalani @devanshj__]

there much better

[Devansh Jethmalani @devanshj__]

did it bwhahaha