https://kotlinlang.org logo
Channels
Powered by
#arrow
  • j

    jacob

    05/12/2019, 11:20 AM
    course I guess by the time MPP re: native is more stable a lot of this stuff is going to have to be sorted out by jb one way or the other
  • r

    raulraja

    05/12/2019, 11:28 AM
    I didn't gather much interest in kapt regarding MPP support which is what seems to be recommended to most users given the state of compiler plugins
  • r

    raulraja

    05/12/2019, 11:28 AM
    And others are using PSI like sqldelight
  • r

    raulraja

    05/12/2019, 11:28 AM
    metaprogramming in Kotlin feels in no mans land.
  • s

    simon.vergauwen

    05/12/2019, 11:29 AM
    Everything out there is currently by Google or JB AFAIK. So all inside track.
  • j

    jacob

    05/12/2019, 11:29 AM
    what does PSI stand for?
  • s

    simon.vergauwen

    05/12/2019, 11:30 AM
  • j

    jacob

    05/12/2019, 11:30 AM
    cheers
  • j

    jacob

    05/12/2019, 11:30 AM
    AST I knew, never heard of PSI
  • s

    simon.vergauwen

    05/12/2019, 11:32 AM
    It's not the same AFAIK. Psi is an interface to query info
  • r

    raulraja

    05/12/2019, 11:32 AM
    PSI is basically the IDEA Apis, you can use them programmatically
  • s

    simon.vergauwen

    05/12/2019, 4:58 PM
    In Kotlin-jvm, using the Java 
    File
    api
  • j

    jacob

    05/13/2019, 10:02 AM
    can't wait for 87 to drop so I can lose the first 3 params
    Copy code
    /* calculate :: MonadThrow A => HasRepositories F => HasWarehouseService F => 
                ShoppingCart -> Kleisli A (F a) ShoppingCart */
fun <F, A> calculate(
    AMT: MonadThrow<A>,
    FHR: HasRepositories<F>,
    FHW: HasWarehouseService<F>,
    cart: ShoppingCart
): Kleisli<A, Kind<F, A>, ShoppingCart> {
    return calculateDiscount<F, A>(FHR, cart).flatMap(AMT) { cart -> calculateErrors<F, A>(FHW, cart) }
}
    ➕ 1
  • s

    simon.vergauwen

    05/13/2019, 10:07 AM
    Copy code
    fun <F, A> calculate(cart: ShoppingCart): Kleisli<A, Kind<F, A>, ShoppingCart> =
 calculateDiscount<F, A>(cart).flatMap(::calculateErrors)
  • s

    simon.vergauwen

    05/13/2019, 10:07 AM
    Not just the first 3 params
  • j

    jacob

    05/13/2019, 10:08 AM
    oh ahaha, I didn't even think about the body
  • j

    jacob

    05/13/2019, 10:08 AM
    yeah, that tidies it up so nicely!
  • s

    simon.vergauwen

    05/13/2019, 10:08 AM
    Not sure if Kotlin can defer 
    <F, A>
    in that lambda. I feel like that currently wouldn’t be support to maybe 
    flatMap { cart -> calculateErrors<F, A>(cart) }
  • j

    jacob

    05/13/2019, 10:08 AM
    it's having trouble inferring <F, A> in those functions, so I guess not
  • j

    jacob

    05/13/2019, 10:13 AM
    that function was part of a big block of functions that was specific to ForMonoK. I refactored the whole thing to be generic in A rather than specific, and arrow made that really really easy
    👍 1
  • s

    simon.vergauwen

    05/13/2019, 10:18 AM
    Happy to hear.
  • j

    jacob

    05/13/2019, 11:20 AM
    I think I might have asked this before, but with the talk of kapt and psi, do those transforms only run on code in my project, or do they run on the arrow lib when I compile as well?
  • j

    jacob

    05/13/2019, 11:20 AM
    i.e. if I didn't run kapt, but wrote out interfaces and companions by hand, would it still all work?
  • s

    simon.vergauwen

    05/13/2019, 11:21 AM
    Ye that would still work AFAIK
  • j

    jacob

    05/13/2019, 11:21 AM
    awesome! I wonder if that's an acceptable way around trying to use arrow in js. will give that a shot in a few days
  • s

    simon.vergauwen

    05/13/2019, 11:23 AM
    Uhm, how? I think I might have miss-understood you
  • s

    simon.vergauwen

    05/13/2019, 11:23 AM
    Arrow is not MMP so you’ll not be able to use it in a MPP project
  • j

    jacob

    05/13/2019, 11:24 AM
    but most of it could be right?
  • s

    simon.vergauwen

    05/13/2019, 11:24 AM
    Ye almost all
  • j

    jacob

    05/13/2019, 11:24 AM
    the io stuff would need a huuuge rework, but otherwise there's nothing really jvm about it
1...278279280...416Latest