kotlinlang #arrow

unsafe {
    runBlocking {
        fx {
            val posts = !mPostsApi.getPosts(authToken)        
            mView.showPosts(posts)
        }.handleError {
            Main.shift().map { mView.showGenericErrorAlert() }
        }
    }
}

jacob

05/11/2019, 3:37 PM

not quite ready to demo, but getting close

👏 1

Garth Gilmour

05/11/2019, 9:19 PM

Trying to install the KEEP87 plug-in on latest version of CE (2019.1). The install progress bar runs, but after restart the plug-in does not appear and hence the sample project will not compile. Any thoughts much appreciated…

Imran/Malic

05/12/2019, 4:55 PM

Does anyone know how to create a kotlin file programmatically?

jacob

05/13/2019, 1:00 PM

possible something is broken on the docs site

starke

05/15/2019, 5:31 PM

hello! I have a (hopefully) simple question regarding dealing with multiple

Option

values

val option1 = Option.just("foo")
    val option2 = Option.just(1)
    val option3 = Option.just(Bar())

    val result: Option<FooBar> = option1.flatMap { unwrappedValue1 ->
        option2.flatMap { unwrappedValue2 ->
            option3.map { unwrappedValue3 ->
                FooBar(unwrappedValue1, unwrappedValue2, unwrappedValue3)
            }
        }
    }
}

what is a better way to handle the use case where you want to return an

Option

that is, for example, a composite of multiple other types that are wrapped by

Option

I would like the resulting

Option

to be

None

if any of the other values are

None

, otherwise return

Some

that is a composite of the data from the other `Option`s…if that makes sense?

Robert Menke

05/17/2019, 5:38 PM

Hey, it seem like when using

Try

it’s common to want to write to a log if an exception occurs. I’ve currently been using a pattern like the following, but I’m curious if there’s a more succinct or abstract way I can go about this:

Try { db.update(model) }
   .toEither {
       logger.error("My message", it.localizedMessage)
       it
   }
   .fold(
       { /*handle failure*/ },
       { /*handle success*/ }
    )

Thoughts on how to better handle this?

raulraja

05/17/2019, 5:57 PM

Heads up,

Try

is impure.

simon.vergauwen

05/17/2019, 6:05 PM

So hopefully that will also help you passed the tricky bits. Feedback and questions are always welcomed here 🙂

Derek Berner

05/17/2019, 7:58 PM

So if a suspend function with side effects is pure, but its invocation is not, is that not also true of a non-suspending function with side effects?

ciriti

05/18/2019, 1:13 PM

Thanks a lot @jacob and @pakoito!!!

raulraja

05/20/2019, 6:55 PM

We need a Kotlin serialization interop module for arrow data types so we can provide all that boilerplate to users

raulraja

05/20/2019, 7:45 PM

For those of you in Madrid or nearby https://twitter.com/raulraja/status/1130472054816464896

:arrow: 1

Jorge Castillo

05/22/2019, 1:58 PM

Our aim is to support MPP by the end of the year, before releasing 1.0.0. There’s still roadmap to get there but we aready got people putting time into that so hopefully we can meet the deadline 🙂

Bob Glamm

05/22/2019, 4:05 PM

I’ve duct-taped OCaml to NodeJS

😳 3

simon.vergauwen

05/23/2019, 3:12 PM

Try.applicative().map(a(), b()) { a, b -> .. }

this should work 🙂

pakoito

05/23/2019, 10:44 PM

@Francesco megna https://github.com/JorgeCastilloPrz/ArrowAndroidSamples/tree/jc-upgrade-to-0.9.0-fx

Joan Colmenero

05/24/2019, 1:26 PM

And one with MVP?

raulraja

05/26/2019, 10:02 PM

@Derek Berner @Bob Glamm If

objectMapper.convertValue

can fail with an exception then it should be suspended and you should be using

IO

instead of

Try

. Try is going away or going to be modified because it can’t suspend effects or run suspend functions. If we make Try support suspend then we have IO. On a different token you can encode internally all your error handling efforts including validation without implying fail fast or error accumulation and then export a la carte in your public API the error strategies for each data type that you want to support. This link supports with the same code both fail fast and error accumulation https://arrow-kt.io/docs/patterns/error_handling/#example--alternative-validation-strategies-using-applicativeerror

Tom Hall

05/28/2019, 10:40 AM

I've been reading Scala with Cats (https://books.underscore.io/scala-with-cats/scala-with-cats.pdf) - would that be a good intro? (it seems to me that Arrow is very similar to Cats)

simon.vergauwen

05/28/2019, 11:09 AM

The fact that this is possible is because of the lazyness. We describe a program with Fx and translate (interpret) it to a runtime (Rx, Reactor, IO).

👍 1

🔝 3

raulraja

05/28/2019, 8:18 PM

IO_validation.kt

:arrow: 2

pakoito

05/29/2019, 1:27 PM

I believe Scala's Red Book has some too

noiano

05/29/2019, 2:27 PM

newbie question: I have a method which returns Unit because it just writes out a file … such method can obviously fail is something goes wrong… I’ve used Try<Unit> to capture this … it doesn’t feel quite right 😅

Title

noiano

05/29/2019, 2:27 PM

pakoito

05/29/2019, 4:30 PM

As Derek said, what you're doing is described as an effect. It's a side-effect in the sense that you're changing the world, in this case writing a file, but you'd like for that side effect to be pure, composable, error handling and all things nice. Wrapping a side-effect makes it what we call an effect. Try has most of those properties but fails one, it isn't deferred and gets executed immediately. Others like IO and Observable do have the same properties of Try, and laziness on top. Fun thing is, suspend functions also behave the same way. So at some point instead of composing Try inside suspend functions or IO or Observable, you just use those everywhere in a way that's compatible and consistent. And this reason is why we're going to phase out Try over time. We want the effect APIs to be used to prevent IO<Try<A>> or suspend fun bla(): Try<A> just for error handling

raulraja

05/29/2019, 7:34 PM

@noiano complementing all that Paco said here is an example of using IO for safe resource acquisition and release which is similar to your file use case. (Posting file outside this thread)

👏 1

noiano

05/30/2019, 1:55 PM

thank you very much for your answers ….

I actually did want to have a synchronous evaluation because I was invoking a spark action … I needed to be synchronous .. I’m now reading whether IO supports synchronous invocation

raulraja

05/30/2019, 2:11 PM

IO is heavily optimized for synchronous invocations

all effects are synchronous unless you move to new pools

noiano

05/30/2019, 2:57 PM

ook … I’ll check out the docs … basically try.invoke is similar to IO.invoke for what concerns synchronous invocation, right?

pakoito

05/30/2019, 2:58 PM

yes

noiano

05/30/2019, 2:59 PM

thank you 🙂

View count: 5