Hi Folks, I'm struggling to de-serialise a

data class

object from Google's Firestore. I can save my data class (lets call it Parent) to Firestore, but on retrieval it is a HashMap which seems nontrivial to convert back because some values on Parent are themselves data classes:

data class Parent(val id: String, val child1: Child1, ...)

data class Child1(val name: String)

I can't use GCP's DocumentSnapshot.toObject(Parent::java.class) because for that to work, every field on Parent needs to be nullable, and I'd have to recursively create nullable data class intermediaries (i.e., NullableChild1, etc) ... any suggestions on a better approach? I'm new to Kotlin so figure I'm missing something obvious here 🤷 Example code:

// What i'd like to do

fun getFromFirestore(id): Parent {
   ...
   val data = doc.data()
   data.toObject(Parent::java.class) // this throws because it needs a 0 argument constructor to unpack the hash map into
}

// Overly complex solution

fun getFromFirestore(id): Parent {
   ...
   val data = doc.data()
   data.toObject(NullableParent::java.class) // this throws because it needs a 0 argument constructor to unpack the hash map into
}

data class NullableParent(val id: String?, val child1: Child1?, ...) {
   constructor(): this(null, null) // I could also unpack the hash map here but then this constructor gets verbose!
 
   fun toParent(): Parent {
      if(id == null || child1 == null...) throw Error("blah")
  
      return Parent(id, child1, ...) // this doesn't work because Child1 doesn't have a 0 argument constructor so I'd have to create a NullableChild now 😢
 }
}

How do I implement this correctly?

Let’s say I have, as part of a DSL, the following class

class ListScope {
    internal val items = mutableListOf<BulletedListItem>()

    operator fun String.unaryMinus() = items.add(BulletedListItem(this))
}

allowing me to use

list {
    - "first item"
    - "second item"
}

Am I right to say that it’s impossible to do the same with an extension function of ListScope? I.e. something like

class ListScope {
    internal val items = mutableListOf<BulletedListItem>()
}

operator fun ListScope.String.unaryMinus() = items.add(BulletedListItem(this))

I was pointed out that it's not a good pattern to use

copy()

constructor of a data class as a kind of builder, essentially calling it multiple times to re-create an object adding some data inside in a pattern like:

data class TrackingEvent(
        val eventType: String = "",
        val eventFeature: String = "",
        val eventVariant: String = ""
)

fun TrackingEvent.click() = copy(eventType = "click")

fun TrackingEvent.onArticleFromStage(stage: String): TrackingEvent {
    return copy(
            eventFeature = "article_stage",
            eventVariant = "click on article from stage $stage"
    )
}

TrackingEvent().click().onArticleFromStage("stage")

On one hand I can understand why creating new object would result in GC cleaning up the old one. On the other hand, we have an immutable class and I think I read somewhere about compiler optimizations on this, not creating a new pointer a memory for properties and rather reusing old pointers? Am I wrong? Should I bring back builder back from Java days instead if I want similar functionality?

@Throws doesn't work in my code, "Unresolved reference: Throws". Do I need a special import to make it work?

Throwing this out there in case anyone else has run into it - I’ve got a multiproject with about 500 kotlin files spread over about 15 projects. Everything runs great!… until I’m in a higher memory pressure environment - the default CircleCI container in this case. I’ve been running into sporadic errors during my Gradle build that I’ve deduced are coming from OOM, usually during compile or DCE tasks. If anybody has any clues on how to make this perform more nicely under these conditions, that could save me a few forehead wrinkles 🙂

Is there a shorthand for this?

fun Date.toCalendar(): Calendar {
    val calendar = Calendar.getInstance()
    calendar.time = this
    return calendar
}

I updated to kotlin 1.3.61 and seeing errors like: Cannot access built-in declaration 'kotlin.Unit'. Ensure that you have a dependency on the Kotlin standard library . Cannot access built-in declaration 'kotlin.collections.MutableSet'. Ensure that you have a dependency on the Kotlin standard library

Maybe bit off-topic, but can somebody explain to me why the font text in Intellij IDEA looks better than in Android Studio?

can I safely increase an array in this way, without an intermediary temporary variable?

ptrId = Array(newBufLength) { i -> ptrId.getOrElse(i) { i } }

why no checked exception 😠

I think I found a compiler frontend bug. Can you guys get this piece to compile?

class A

inline fun <reified T> A.foo(): T = TODO("Not necessary to implement")

open class B {
    val a = A()
    inline fun <reified T> foo() = bar { foo() }
    fun <T> bar(baz: A.() -> T) = a.baz()
    companion object : B()
}

is using reified good or bad?

Hi 👋 I'd like to declare a data class property as a function:

matcher: (x: Any) -> Boolean

. Later when i pass in a lambda for matcher, I have to cast from Any to the concrete type as per blew. Is there a better way to do this?

data class Query(
 matcher: (x: Any) -> Boolean
)
...
val query = Query(
 matcher: { x: Any? ->
   val value = x as Int?
   if(value == null || value == 0) false else true
 }
)

The Kotlin documentation says that if all of the primary constructor parameters for a class have default values, on the JVM, Kotlin will create an additional parameterless constructor. So I was surprised that this doesn't work:

class BadTest(val name: String? = null)
val badTestCreator: () -> BadTest = ::BadTest

But that this does work:

class GoodTest(val name: String? = null) {
    constructor(): this(name = null)
}
val goodTestCreator: () -> GoodTest = ::GoodTest

Shouldn't there be a no-arg constructor for

BadTest

that would satisfy the

() -> BadTest

type?

tbh hate forloop of kotlin is there anyway can mate with index via forloop instead of choosing while.

This code compiles

val xyz: Map<String, (String) -> Unit> = mapOf("abc" to { it -> println(it.length) })

with a suggestion that there is a 'redundant lambda arrow' (the

it ->

part in the lambda in the map key). However, if I remove that portion, I get a 'type inference failed' compile error. The compiler interprets the lambda as

() -> ???

rather than

(String) -> Unit

. Shouldn't the compiler interpret

{ println(it.length) }

as satisfying

(String) -> Unit

?

in kotlin native, are generics specialized? will

val list = arrayListOf<Int>()

result in boxed values?

Might there be a better way to multiply an integer with a float and getting a result of Integer? (paddingTop is Int, factor is Float)

paddingTop = (paddingTop * factor).toInt()

Maybe an extension function but this seems odd:

private fun Int.fMul(factor: Float) = (this * factor).toInt()

Im launching a simple python script and trying to read its output like this

val pp = Path.of(System.getProperty("user.dir"),"test.py")
        val p = ProcessBuilder("python3", pp.toString())
        p.redirectErrorStream(true)
        val pr = p.start()
        pr.outputStream.close()
        while (true) {

            if (pr.inputStream.available() < 1)
                Thread.sleep(10)
            else {
                val content = pr.inputStream.bufferedReader().use(BufferedReader::readText)
                println(content)
            }
        }

Python script:

import time
print("ABC")

i = 0
while True:
    print("ABC")
    i += 1
    time.sleep(1)

But nothing gets printed to screen. (Unless I remove the

while

part, then it actually prints "ABC" so python path, script path is correct) What is the correct way to do this?

is there any default logger in kotlin

Is it okay to use

@kotlin.internal.InlineOnly

? (esp. in a Kotlin-only library) 🙂

Why are delegates (as in

val a: Int by lazy { 5 }

not an interface? Instead they are weird classes that have to have certain methods without specifying an interface... I would like to write a

and

infix function that creates a delegate that combines two others, but I'm afraid that won't be possible because of this

is there any reason why Kotlin doesn't provide functions for stack in the stdlib? like

stackOf("value 1", "value 2")

Hey Kotlin fam! I'm sure there must be a more efficient way to do this, can't quite think of what though... Determine whether a list of objects contains an object with a specified property value? Example (bad):

data class Objects(
    val objects: List<MyObject>
) {

    fun containsFirstProperty(value: String): Boolean {
        var itDoes = false
        objects.forEach {
            if (it.firstProperty == value)
                itDoes = true
        }
        return itDoes
    }
}

data class MyObject(
    val firstProperty: String,
    val secondProperty: String
)

Thanks!

Hi Kotliners! We have been running kotlin in some greenfield projects for about 2-3 years and we've been super happy with it. It has started to gain a lot of traction in our company but lately we've heard about people who experience kotlin to be super slow build-time and development time. This is especially when you have a bit bigger projects who are running a mix of Java and Kotlin. Does anyone recognize this issue and knows if there's any intellij-trick you could do to get a smoother experience?

wait is 1.3.70 released? Just saw a notif to update Kotlin's plugin on IntelliJ and there have been some releases on the GitHub page

Kotlin 1.3.70 released Today we’re happy to present to you a new release of Kotlin – 1.3.70. This is an incremental release, so it doesn’t bring major new features. However, we did our best to improve the existing functionality, fix issues, and even … Continue reading →

Say I have the following four classes:

* DatabaseKennelModel
    + dogs: List<DatabaseDogModel>

* DatabaseDogModel
    + name: String

* Kennel
    + dogs: List<Dog>

* Dog
    + name: String

Is there a better way to transform a

List<DatabaseKennelModel>

into a

List<Kennel>

, than the following:

databaseKennelModelList.map { databaseKennelModel ->
    Kennel(
        dogs = databaseKennelModel.dogs.map { databaseDogModel ->
            Dog(
                name = databaseDogModel.name
            )
        }
    )
}

☝️ it works, but I'm wondering is there anything I can do to avoid that nested

map

?

Does anyone else find themselves wanting a shorthand operator to discard a result? Sometimes a function returns a result but you want it to parse/behave as though it's

Unit

. Any tricks to achieving this neatly?