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

    jeremy

    03/17/2019, 1:11 AM
    Yeah
  • b

    Bob Glamm

    03/18/2019, 2:37 PM
    Copy code
    fx {
            val (rs: ListK<NamedResource>) = service.loadNamedResources()
            rs.forEach { r ->
                val name: String = r.name
                val (ins: InputStream) = !effect { r.resource }
                val (contents: ByteArray) = !effect { service.loadContents(ins) }
                val (jsonNode: JsonNode) = !effect { service.parseContents(contents) }
                val id: String = service.getId(jsonNode).getOrElse { "" }
                !effect { <http://LOG.info|LOG.info>("{} -> {}", name, id) }
            }
        }.unsafeRunSync()
    fails with StackOverflowError after processing 98 
    rs
    (there are a total of 443 for it to go through.) Is there some sort of tail recursion method I'm missing?
  • b

    Bob Glamm

    03/18/2019, 2:41 PM
    I see a lot of the following repeated in the stack trace:
    Copy code
    at kotlin.coroutines.ContinuationKt.startCoroutine(Continuation.kt:114)
	at arrow.effects.typeclasses.suspended.FxSyntax$effect$1.invoke(FxSyntax.kt:41)
	at arrow.effects.typeclasses.suspended.FxSyntax$effect$1.invoke(FxSyntax.kt:19)
	at arrow.effects.extensions.IOConcurrent$async$1.invoke(io.kt:135)
	at arrow.effects.extensions.IOConcurrent$async$1.invoke(io.kt:120)
	at <http://arrow.effects.IO|arrow.effects.IO>$Companion$async$1.invoke(IO.kt:40)
	at <http://arrow.effects.IO|arrow.effects.IO>$Companion$async$1.invoke(IO.kt:26)
	at arrow.effects.IORunLoop.loop(IORunLoop.kt:220)
	at arrow.effects.IORunLoop.access$loop(IORunLoop.kt:18)
	at arrow.effects.IORunLoop$RestartCallback.invoke(IORunLoop.kt:378)
	at arrow.effects.IORunLoop$RestartCallback.invoke(IORunLoop.kt:356)
	at arrow.effects.internal.Platform$onceOnly$1.invoke(Utils.kt:43)
	at arrow.effects.internal.Platform$onceOnly$1.invoke(Utils.kt:18)
	at arrow.effects.internal.UtilsKt$asyncContinuation$1.resume(Utils.kt:132)
	at arrow.core.Continuation$DefaultImpls.resumeWith(Continuation.kt:12)
	at arrow.effects.internal.UtilsKt$asyncContinuation$1.resumeWith(Utils.kt:128)
	at kotlin.coroutines.jvm.internal.BaseContinuationImpl.resumeWith(ContinuationImpl.kt:45)
  • t

    thanerian

    03/18/2019, 2:43 PM
    I see you are deconstruct some effects 
    val (ins: InputStream) = ...
    and also extract the value from the effect 
    ... !effect{}
  • t

    thanerian

    03/18/2019, 2:43 PM
    have you tried using only one of them?
  • b

    Bob Glamm

    03/18/2019, 2:47 PM
    hmm, that definitely helped, I'm now up to 390 out of 443
  • b

    Bob Glamm

    03/18/2019, 2:51 PM
    So what is idiomatic for the final line in 
    rs.forEach()
    ? Using 
    val (x: Unit) = IO { <http://LOG.info|LOG.info>("...") }
    works, but using 
    !effect { IO { <http://LOG.info|LOG.info>("...") } }
    yields 
    StackOverflowError
  • n

    noiano

    03/18/2019, 2:52 PM
    good morning everybody
  • n

    noiano

    03/18/2019, 2:52 PM
    I have the weirdest exception when running a unit test involving an either
  • n

    noiano

    03/18/2019, 2:52 PM
    Copy code
    java.lang.AssertionError: expected: arrow.core.Either$Right<Right(b=GeosetStatus(status=UNKNOWN_GEOSET_STATUS, additionalInfo=Some(GeosetInquiryHttpResponse(errorType=Some(END_DATE_BEFORE_START_DATE), message=Some(wrong dates)))))> but was: arrow.core.Either$Right<Right(b=GeosetStatus(status=UNKNOWN_GEOSET_STATUS, additionalInfo=Some(GeosetInquiryHttpResponse(errorType=Some(END_DATE_BEFORE_START_DATE), message=Some(wrong dates)))))>
  • n

    noiano

    03/18/2019, 2:52 PM
    Looks like a bad joke to me since the type signature is exactly identical
  • n

    noiano

    03/18/2019, 2:52 PM
    What am I missing?
  • b

    Bob Glamm

    03/18/2019, 2:53 PM
    Something doing reference equality comparison instead of value equality comparison?
  • n

    noiano

    03/18/2019, 2:53 PM
    🤔
  • n

    noiano

    03/18/2019, 2:53 PM
    everything is data class
  • n

    noiano

    03/18/2019, 2:54 PM
    and I’m using kotlin’s assert.AssertEquals
  • p

    pakoito

    03/18/2019, 3:17 PM
    @raulraja
  • p

    pakoito

    03/18/2019, 3:18 PM
    assert.AssertEquals
    use == and see what happens, maybe the dates are not comparable
  • t

    thanerian

    03/18/2019, 3:20 PM
  • t

    thanerian

    03/18/2019, 3:20 PM
    and parTraverse
  • t

    thanerian

    03/18/2019, 3:21 PM
    List like Option and other data types are not stack safe
  • t

    thanerian

    03/18/2019, 6:17 PM
    They does not garantice stack safety
  • r

    raulraja

    03/18/2019, 6:29 PM
    @Bob Glamm is this something you can reproduce with a reduced example that compiles? Initially if that part of the stack trace is repeating it seems to be an issue with the IO loop. Also how are you releasing resources in those input streams? Normally you'd use bracket for interacting with something like that.
  • p

    pakoito

    03/18/2019, 6:39 PM
    Untitled.java
    Untitled.java
  • p

    pakoito

    03/18/2019, 6:40 PM
    that test completes correctly
  • p

    pakoito

    03/18/2019, 6:40 PM
    so it’s not on the suspension
  • p

    pakoito

    03/18/2019, 6:40 PM
    or the IO run loop
  • p

    pakoito

    03/18/2019, 6:42 PM
    wheck what you’re doing with the service, maybe it’s loading things on a loop?
  • p

    pakoito

    03/18/2019, 6:42 PM
    and in any case 
    .unsafeRunSync()
    should be called once per program/activity
  • b

    Bob Glamm

    03/18/2019, 6:59 PM
    Easiest question first: I'm actually calling 
    .unsafeRunSync()
    at the end of a 
    @Test
    so that is essentially once per program
1...246247248...416Latest