Hi All! I am looking for some ideas and perspectives on deserialisation, delegation, and type-safety, please. For the sake of argument, I am using

Protobuf

and Square's

Wire

to serialise data, and to deserialise into Kotlin. Just like the example at https://github.com/square/wire#protocol-buffers I have a common

Dinosaur

message which has got a

name

string  The first and obvious change in the direction that I want to go, would be to replace the

name

with a protobuf

enum

so that I get

Dinosaur.STEGOSAURUS

(etc) as a nameEnum, which means that the code can find out the soft-type of the message; but that doesn't give me what I want. The second approach would be to use the protobuf

oneof

type - and it would essentially do exactly the same soft-type thing - I could use

Dinosaur.STEGOSAURUS.foo.bar

- but that still does not give me what I want. What I want is for the deserialisation process of a

Dinosaur

message to yield a

Stegosaurus

type; or a

TRex

or

Apatosaurus

type, or whatever, so that resulting code can guarantee having an

Apatosaurus

passed into it. I am presuming that some sort of

DinosaurParser.peekType()

and

DinosaurParser.getApatosaurusOrThrow()

will be needed, along with a type-enforcement

is/as

check. But that leaves me with the question of delegation so that I can access the other fields of the

Dinosaur

message, like height, weight, etc. • what's the best way for me to convert a "Dinosaur-message-with-TRex-soft-type" in the serialisation domain • into a kotlin-domain strongly-typed-instance-of-

TRex

that the compiler can check • but also inherit (delegate?) all the accessors and methods from the underlying

Dinosaur

message, into the

TRex

instance? It would be nice to do:

class TRex(d: Dinosaur) : Dinosaur by d

...however the protobuf compilers are generating

Dinosaur

classes, not interfaces, so I don't think that option is open to me?

Our backend sends me ARGB values where the RGB is 0-255, but the alpha channel is 0-1. What’s the best way for me to scale the 0-1 to be the same as 0-255?

I have the following code

private const val TIME_OUT = 10000 // in milliseconds
        private val config = RequestConfig.custom()
            .setConnectTimeout(TIME_OUT)
            .setConnectionRequestTimeout(TIME_OUT)
            .setSocketTimeout(TIME_OUT)
            .build()

I only need

TIME_OUT

for this initialization. How can I avoid having it hanging around? One idea I had was

private val config = 10_000.let { timeOut ->
            RequestConfig.custom()
                .setConnectTimeout(timeOut)
                .setConnectionRequestTimeout(timeOut)
                .setSocketTimeout(timeOut)
                .build()
        }

but that is not pretty either. Are there better ways?

Does anyone have any idea why

kotlinCompilerClasspath

would downgrade my Kotlin version from 1.4.20 → 1.4.10, or could perhaps point me in the right direction on where to look? I see this when I run `./gradlew dependencies`:

kotlinCompilerClasspath
+--- org.jetbrains.kotlin:kotlin-compiler-embeddable:1.4.20 -> 1.4.10
|    +--- org.jetbrains.kotlin:kotlin-stdlib:1.4.10
|    |    +--- org.jetbrains.kotlin:kotlin-stdlib-common:1.4.10
|    |    \--- org.jetbrains:annotations:13.0
|    +--- org.jetbrains.kotlin:kotlin-script-runtime:1.4.10
|    +--- org.jetbrains.kotlin:kotlin-reflect:1.4.10
|    |    \--- org.jetbrains.kotlin:kotlin-stdlib:1.4.10 (*)
|    +--- org.jetbrains.kotlin:kotlin-daemon-embeddable:1.4.10
|    \--- org.jetbrains.intellij.deps:trove4j:1.0.20181211
\--- org.jetbrains.kotlin:kotlin-compiler-embeddable:1.4.10 (*)

There are no (direct) references to 1.4.10 in my dependency declarations, either. And that's the only place where I see the downgrade to 1.4.10. My Kotlin dependencies:

object Kotlin {
    const val version = "1.4.20"

    const val gradlePlugin = "org.jetbrains.kotlin:kotlin-gradle-plugin:$version"
    const val stdLib = "org.jetbrains.kotlin:kotlin-stdlib-jdk7:$version"
}

Hi there, quick question about split method for String if I have val v = "" then I do v.split(".") I expect the method to return an empty List, but it actually returns a list with one element, and that element is my original string "" I can confirm that by looking at the code:

private fun CharSequence.split(delimiter: String, ignoreCase: Boolean, limit: Int): List<String> {
    require(limit >= 0, { "Limit must be non-negative, but was $limit." })

    var currentOffset = 0
    var nextIndex = indexOf(delimiter, currentOffset, ignoreCase)
    if (nextIndex == -1 || limit == 1) {
        return listOf(this.toString())  -- HERE IT IS
    }

Is this a behavior i should expect?

Can someone help me with this please I keep getting failed to compiled values files every time I run the app. I have invalidated the cache and restarted the project but still getting the same error. I also added this dependency nothing worked implementation ‘androidx.core:core-ktx:1.3.2’

It seems kind of concerning that everything in the standard library that counts things, returns integers, accepts an index, etc, always seems to use

Int

. This seems like a pretty odd choice. Is there any way to instead have the standard library use a 64 bit integer?

Suppose I have:

annotation class MyAnnotation(val someValue: String, val optionalClass: KClass<*> = MyDefaultImplementation::class)

interface MyInterface {
    fun doSomething(input: Any?): Any?
}

class MyDefaultImplementation: MyInterface {
    override fun doSomething(input: Any?): Any? = input.toString()
}

How can I set a bound on

optionalClass

to make sure only a class which implements

MyInterface

is provided?

how are enum classes implemented on the JVM? do they produce code that’s similar to a java

enum

? asking because i’m trying to figure out how swagger annotations work and if i can rename them

Hello people, I have a question about the hotifix versioning. Suppose I released the app with minor fix which is from 

1.1.1

 to 

1.1.2

  and in this version I stopped the rollout because we found out that there is some bugs. So I created the new branch 

hotfix/1.1.2

 from the 

release/1.1.2

 branch and merge the changes into 

release/1.1.2

 branch. But now the question is 1. Is this correct to fix the issue on the 

release/1.1.2

  branch? 2. What version we should use for the hotfix release? should we use 

release/1.1.2

 or 

release/1.1.3

 ? 3. Should we merge the 

hotfix

 branch to 

release/1.1.2

 branch or first do we need to create the new branch 

release/1.1.3

 and merge hotfix changes into new branch?

The kotlin duration is killing me:

(200.days + 5.nanoseconds - 200.days).toLongNanoseconds()

This evaluates to

4

, not

5

.

When calling

String.replace(String, String)

for an unlikely replacement, does it ever makes sense to wrap in a

CharSequence.contains(CharSequence)

call, for efficiency?

Hi, I am trying to understand why this 'when' expression is not exhaustive:

sealed class Fruit {
    object Apple : Fruit()
    object Orange : Fruit()
}

fun <T : Fruit> eat(fruit: T): Unit =
    // 'when' expression must be exhaustive, add necessary 'else' branch
    when (fruit) {
        is Fruit.Apple -> {}
        is Fruit.Orange -> {}
    }

This is just a contrived example, I know that I could simply inline 'T' with 'Fruit'. I am more interested to understand why Kotlin fails to see that the 'when' expression is exhaustive.

Apparently it’s OK to pass null down a Flow, and Flow is intended to be compliant with the Reactive Streams spec. But the Reactive Streams spec disallows null, which is why RxJava 2 had that massive breaking change. I’m getting confused about something… why can a Flow handle null, but Rx2 can’t?

is there a built-in method that lets me filter a

Map<K,V?>

to

Map<K,V>

? I.e. filter all null values and cast the map?

If we're writing some kind of not-completely-trivial transformation that can work on both sequences and iterables (as most can), it's pretty common to have to write the function twice. Is there a convention for which one should call which, in terms of the best performance? Or is there some preferred way to share code in a third function/object? Or should the code just be copy pasted? I guess usually the version that runs on iterable is the one that gets

inline

, so perhaps you'd want the sequence version to call the iterable version?

That is not true: “It is not required to match directories and packages: source files can be placed arbitrarily in the file system.” — Language Guide https://kotlinlang.org/docs/reference/basic-syntax.html

This is also wrong: “The return type is not specified, which means that the function returns nothing.” — Examples https://play.kotlinlang.org/byExample/01_introduction/01_Hello%20world cause it returns Unit — https://pl.kotl.in/-E3uqRzER

i am not seeing KMM project template.

Why is there no pow function for integers?

I have a

List<String>

that i'd like to split into a

List<List<String>>

by indicating a delimiter

String

, similar to how i can split a

String

into a

List<String>

by indicating a delimiter

Char

using

.split()

. unfortunately lists dont have a

.split()

function, so what would be the best way to implement it ? looking for a functional approach if possible

I have this -

interface A {
    val na: Int
}
class B {
    val <T: A> T.nb: Int by lazy {
        <http://this.na|this.na> // !!! I want to reference T/A here, not B.
    }
}

I want to reference

T : A

but the compiler won't let me.

this

-s type is

B

. Is there another way?

📢 Revamped :kotlin: Documentation – Give It a Try Hi all! We’re revamping our Kotlin documentation to bring you more helpful features and improve your user experience. To name just a few advantages, the new Kotlin documentation: • Is mobile friendly. • Has a completely new look and improved structure. • Provides easy navigation on each page. • Lets you share feedback on every page. • Lets you copy code with a single click. More features such as dark theme support are coming soon. Before going to production and deprecating the current version of the documentation, we want to ensure that we haven’t missed any critical issues. We’d be extremely grateful if you would look through the new revamped documentation and share your feedback with us. 👀 View new documentation When you view the Kotlin documentation you’ll still see it in the old format. To view the documentation in the new revamped format, click this link. This will add a cookie to your browser which will enable you to see the documentation in the new format. If you want to check anything in the old version of the documentation, open the usual link in another browser or use an incognito window in your current browser. 🗣️ Share your feedback 🤗 Your feedback is very important for us. You can: • Share feedback in the #docs-revamped channel. • Report an issue to our issue tracker. • Email us at doc-feedback@kotlinlang.org Try out the new documentation now!

Does anyone know how to make something like this 🙂 ? I want to have type hints for reaction based on T class extended DomainEvent.

val chainReactions: ChainReactions = mutableMapOf()
typealias ChainReactions = MutableMap<DomainEvent, DomainEvent>

I am using

Thread.currentThread()

to obtain the line number, the problem is when I use a

inline fun

this number is not valid and I can’t see a pattern to fix it. Is there a workaround to this problem?

I noticed that if I create a class whose primary constructor contains only parameters with default values - an empty constructor is available for Java code to call, without the use of

@JvmOverloads

e.g.

class Foo(s: String? = null, i: Int = 3) // `new Foo();` from Java compiles

However, that’s not the case when using a secondary constructor e.g.

class Foo {
   constructor(s: String? = null, i: Int = 3) // `new Foo();` from Java does not compiles
}

Is this behavior documented somewhere? I couldn’t find anything..

How could I design the following code in a more functional, immutable way? There's a single string where I want to throw multiple regexes on it. The size of the list of regex is flexible, so I can't operate over a set range. I'm actually refactoring some guys code and it grinds my gears that I cannot come up with a simpler solution for that:

var label = when {
    ..someConditions.. -> "someString"
    else -> ""
}

regexes.forEach {
    label = label.replace(it.toRegex(), "").trim()
}

return label

I should maybe just step away from it for some time.. 😄 any hint is appreciated, thanks!

Hi everyone, I would like to know where the kotlin project people usually to save? For example, the android project usually saved in

~/AndroidStudioProjects/

. How about non-android project?