TestClock

feels like it should have `plusAssign`/`minusAssign` of a

Duration

as a convenience.

clock += 2.minutes

Hi there, why would 2 different lists return the same hash?

println(listOf(3, 263).hashCode())
 println(listOf(5, 201).hashCode())
// -> 1317

println(Arrays.asList(3, 263).hashCode())
println(Arrays.asList(5, 201).hashCode())
// --> 40830

It's a more general problem as also the java example suffers from exactly the same problem. Interestingly, arrays with the same data give differnt hashes as documented at https://github.com/holgerbrandl/krangl/issues/75

How do you convert a data class to a map object?

sometimes you can back your fields with a map

var x: Int by map

Hello people 🖖 Does anybody know why

any

or

all

returns

true

with an empty list no matter what the predicate? Seems counterintuitive. I tried to google it quickly but no luck. Just wondering, maybe someone's aware of it and can explain 🙂

hi i just noticed the following: in java,

Stream.reduce(accumulator: BinaryOperator<T>)

returns an

Optional<T>

it makes sense because if the stream is empty, the result will be empty there’s another flavor providing an initial value and that returns a

T

Stream.reduce(identity: T, accumulator: BinaryOperator<T>): T

that makes sense too however, in kotlin both

Sequence.reduce()

(without initial value) and

Sequence.fold()

(with an initial value) return the same non-nullable type actually, the code throws an exception if the sequence is empty which seems not in line with what i would expect a nullable value why this behavior? am i missing something?

reduce

would be great extension for

NonEmptyIterable

https://youtrack.jetbrains.com/issue/KT-21217

oh, sorry Mike, I see what you mean, reduce would work on a NonEmptyIterable, I thought you meant as a return from

sequence.reduce

Is there a

List<E>.associateWithIndex(): Map<E, Int>

or something similar that I am missing. (edited)

Hi there, is there a reason not to have a type parameter for the delegate on a

KProperty

? if there was, one could write extension functions, depending on the delegate type. I.e.

fun KProperty<*, BehaviorRelayDelegate>.subscribe(O: Observer)

and use it like

val v : Int by BehaviorRelayDelegate(0)
::v.subscribe { Log("Value changed to $it" }
v = 2
val v1 = v

This is a repost (from #general) about extending the Boolean type: https://kotlinlang.slack.com/archives/C0922A726/p1569444125303200

Can we kotlinify Java’s

Objects.hash(foo, bar, baz, …)

to something like this?

hash { foo x bar x baz x … }

Just the boxing is a little tricky as well as getting it to work with just a single non-

Int

value without introducing boxing. Implementation:

inline fun hash(block: HashScope.() -> Int) =
	HashScope.block()

object HashScope {

	inline infix fun Any?.x(hashable: Any?) =
		(31 * hashCode()) + hashable.hashCode()

	inline infix fun Any?.x(hashable: Int) =
		(31 * hashCode()) + hashable

	inline infix fun Int.x(hashable: Any?) =
		(31 * this) + hashable.hashCode()

	inline infix fun Int.x(hashable: Int) =
		(31 * this) + hashable
}

class Example(val foo: String, val bar: List<Any>?, val baz: Int) {

	override fun hashCode() =
		hash { foo x bar x baz }
}

Another question: Why doesn’t

ClosedRange

have operators for destructuring?

Proposal.

@kotlin.internal.InlineOnly
public inline fun buildByteArray(builderAction: ByteArrayOutputStream.() -> Unit): ByteArray

Is there a better way to write this:

val CJK_EXTENSIONS_STR = fun (): String {
    var result = "\\p{InCJK_UNIFIED_IDEOGRAPHS_EXTENSION_A}\\p{InCJK_UNIFIED_IDEOGRAPHS_EXTENSION_B}\\p{InCJK_UNIFIED_IDEOGRAPHS_EXTENSION_C}\\p{InCJK_UNIFIED_IDEOGRAPHS_EXTENSION_D}"
    if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.N) {
        result += "\\p{InCJK_UNIFIED_IDEOGRAPHS_EXTENSION_E}"
        if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.P) {
            result += "\\p{InCJK_UNIFIED_IDEOGRAPHS_EXTENSION_F}"
        }
    }
    return result
}.invoke()

Is there a YouTrack issue for

Closeable

in stdlib? I remember the verdict being stdlib itself doesn't use it so it won't provide it, but there should a ticket somewhere.

Any particular reason

Iterable<T>.groupByTo

requires destination to be a list and not a collection ?

Suggest adding

fun Pair<A, B>.flip(): Pair<B, A> = Pair(second, first)

to the standard library

Hi! I couldn't find in stdlib common implementation fo formatting Double to String by pattern. Is there any crossplatform solution for that?

Encoding/Decoding Base64 without depending on Java methods

I see that kotlin is getting its own implementation of

Duration

in the stdlib, and I'm wondering if it could also have its own version of

Date

that cleans up the mess that is Java's

Date

,

DateTime

,

LocalDate

,

LocalDateTime

Hello! Is there any plans to support multiplatform BigDecimal?

When is it best to use

!!

over

checkNotNull

?

did something change with

Result

between 1.3.21 and 1.3.50? in 1.3.21 i can access it fine but can't find it in 1.3.50 (

-Xallow-result-return-type

doesn't help either)

What do you guys think?

when i run this.iterator() which calls invoke(Int) should i assume origin is the vtable for IDataframe::invoke(Int) ? do i need to manually delegate each method i wish to expose if I iwsh to use specialized methods I write in the delegator ?

class FrameGrouper(
    private val origin: IDataFrame,
    vararg val gby: Int
) :  IDataFrame  by origin  {
    override fun invoke(row: Int) =...}

inline fun <T, reified R> Array<T>.mapToArray(transform: (T) -> R) = 
    Array(size) { transform(get(it)) }

etc for

List

, primitive arrays...

Just throwing this out there, does anyone think it would be a good idea to include such a function in std-lib:

inline infix fun Any?.hash(other: Any?): Int = hashCode() * 31 + other.hashCode()

It would simplify writing

hashCode

functions:

class ABC<A, B, C>(val a: A, val b: B, val c: C) {
    override fun hashCode() = a hash b hash c
}

Why is the function

List(7) { it + 3 }

a top level function instead of an static invoke operator? Is the latter somehow unintended usage of the language, or against standard code style? static invoke operator is just so much more efficient than constructors, given you can error check prior to instantiation, among many other things. I can see how it can read very alien to those not kotlin-savvy, however.

Proposal.

fun CharSequence.splitToSequences(
    vararg delimiters: String,
    ignoreCase: Boolean = false,
    limit: Int = 0
): Sequence<CharSequence>