hello, this might be a dumb question, but when compiling project with IR, do all of the libs have to be compiled with IR aswell?

I’m seeing an error relating to

kotlin.Result

when I enable IR. No other changes. [This line of code](https://github.com/robfletcher/strikt/blob/main/strikt-core/src/main/kotlin/strikt/assertions/Result.kt#L17) starts throwing

java.lang.ClassCastException: class kotlin.Result$Failure cannot be cast to class kotlin.Result (kotlin.Result$Failure and kotlin.Result are in unnamed module of loader 'app')

The same issue as in the old BE, but with the different stacktrace: https://youtrack.jetbrains.com/issue/KT-28287#focus=Comments-27-4712502.0-0 [“Show Kotlin bytecode”: “Internal error: null” for inline property delegation from a different file]

I am getting the below error very regularly when I build in IntelliJ with Kotlin 1.4.31 and Compose for desktop 0.4.0-build171. The only way I have found to fix it is to do a gradle clean and build. Is there a way I can get it to go away permanently? It seems to happen only when I edit my composable functions.

Hello, everyone! My app got an error and the app is an Android app, but I think this is the right channel for this topic. As long as I remember, the last thing I've done that might be contributing to this error is enabling the IR compiler on the Android Studio by the following line

kotlinOptions.useIR = true

The view model has a LiveData that hold Kotlin's Result class

fun doSomething(code: String, price: Long)  {
        repository.doX(code, price)
                .doOnComplete { _priceSubmission.postValue(Result.success(true)) }
                .doOnError { _priceSubmission.postValue(Result.failure(it.cause ?: it)) }
                .fromIo2Ui()
                .subscribe()
                .disposeOnCleared()
}

and the activity observing it by

viewModel.priceSubmission.observe(this) { result ->
            if (result.isSuccess) {
                setResult(RESULT_OK)
                finish()
            } else {
                val error = result.exceptionOrNull()
                if (error is HttpException && error.code() == 400) {
                    showDialog()
                } else {
                    ToastUtils.showNetworkFailureToast()
                }
            }
        }

I got an error on runtime

java.lang.VerifyError: Bad type on operand stack
Exception Details:
  Location:
    com/app/package/ui/MyActivity$onCreate$1.onChanged(Lkotlin/Result;)V @31: invokevirtual
  Reason:
    Type 'java/lang/Object' (current frame, stack[0]) is not assignable to 'kotlin/Result'
  Current Frame:
    bci: @31
    flags: { }
    locals: { 'com/app/package/ui/MyActivity$onCreate$1', 'kotlin/Result' }
    stack: { 'java/lang/Object' }
  Bytecode:
    0x0000000: 2b59 c600 09b6 001c a700 0557 0112 1eb8
    0x0000010: 0024 2b59 c600 09b6 001c a700 0557 01b6
    0x0000020: 001c b800 2899 0015 2ab4 000d 02b6 002e
    0x0000030: 2ab4 000d b600 31a7 004d 2b59 c600 09b6
    0x0000040: 001c a700 0557 0112 1eb8 0024 2b59 c600
    0x0000050: 09b6 001c a700 0557 01b6 001c b800 354d
    0x0000060: 2cc1 0037 9900 1a2c c000 37b6 003b 1101
    0x0000070: 90a0 000d 2ab4 000d b800 3ea7 0009 b200
    0x0000080: 44b6 0047 b1                           
  Stackmap Table:
    same_locals_1_stack_item_frame(@11,Object[#24])
    same_locals_1_stack_item_frame(@13,Object[#5])
    same_locals_1_stack_item_frame(@29,Object[#24])
    same_locals_1_stack_item_frame(@31,Object[#5])
    same_frame(@58)
    same_locals_1_stack_item_frame(@69,Object[#24])
    same_locals_1_stack_item_frame(@71,Object[#5])
    same_locals_1_stack_item_frame(@87,Object[#24])
    same_locals_1_stack_item_frame(@89,Object[#5])
    append_frame(@126,Object[#76])
    chop_frame(@132,1)


	at com.app.package.ui.MyActivity.onCreate(MyActivity.kt:36)
	at android.app.Activity.performCreate(Activity.java:7136)
	at android.app.Activity.performCreate(Activity.java:7127)

The 36th line of

MyActivity

is the

priceSubmission

observer Is this Kotlin's IR backend issue? Thanks.

I was a little surprised to see that the

useIR

flag became a no-op with 1.5. Was this intentional? I’m aware that the new IR backend will be the default with 1.5. But I assumed setting it to false will go back to the old backend (rather than using another property). https://github.com/JetBrains/kotlin/blob/1.5.0-M1/compiler/cli/src/org/jetbrains/kotlin/cli/jvm/jvmArguments.kt#L195-L200

When trying out new IR, should we add

kotlinOptions.useIR = true

for every module we have? New IR works fine, no issues so far 👌

Not gonna lie, it’s quite discouraging if bugs reported for the IR backend over a month ago are targeted to be fixed in

1.5.20

rather than the

1.5.0

release https://youtrack.jetbrains.com/issue/KT-44891

Is JVM IR on by default in 1.5.0-M2?

In that case I found https://youtrack.jetbrains.com/issue/KT-45760

With new backend IR, should kotlinx serialization library be updated to specific version? Getting this crash when run build with IR enabled (Kotlin 1.4.31, Kotlinx serialization 1.0.1)

Hello, just ran a quick benchmark on a Android Project, Kotlin 1.4.30, and it seems

useIR

does slow down the build a bit:

useIR: false, clean & rebuild:
1. 41s
2. 41s
3. 40s

useIR: true, clean & rebuild:
1. 44s
2. 43s
3. 43s

Granted, it is only marginal slower since it is also only a relatively small project (single module) - but will it affect larger projects proportionally? Out of curiosity I also tried with

compose: true

and was rather disappointed:

useIR: true, compose: true, clean & rebuild:
1. 53s
2. 52s
3. 52s

I honestly expected the new Backend to be if anything, then faster than the existing one?

I'm trying to test a fix in 1.5.0 m2 to see if this fixes an issue with IR compiler BUT I can't seem to run it because I use hilt in this project. I just added hilt last week. Hope I don't have to revert it. Searching in kotlinlang slack and on google hasn't gotten me any promising results. Am I missing something simple to just enable hilt working with 1.5+?

Not sure if it's a bug in the new backend or in MockK, but it sure is a difference in behavior: This test fails with the new backend and works with the old one.

Found a bug with contravariance and intersection types, filed with a repro sample here: https://youtrack.jetbrains.com/issue/KT-45945 https://youtrack.jetbrains.com/issue/KT-45945

We’re running into a crash at runtime with Kotlin 1.5.0-M2 IR enabled. This doesn’t happen with M1, so it’s a new regression. It’s actually also the first runtime crash I’m running into with the IR backend. All other problems were compile time problems. https://youtrack.jetbrains.com/issue/KT-45967

We are having a failure on our CI, running UI tests, only after enabling the useIR option in our 🤖 project. The failure is about a custom rule failing with this

protected val activity: T by lazy {
activityRule.activity
} must not be null

This is happening somehow else. also, if it helps,

@JvmField
@Rule
var activityRule: SomeActivityTestRule<out T> =

This is blocking me from updating to 1.5 and it doesn't even have a target fix version 😑 I reported it over a month ago.

I’ve been working on a compiler plugin that hooks in via

AnalysisHandlerExtension

, it just generates some extra files and nothing too complicated. It works fine on the old backend, but when IR is enabled it fails somewhere deep in the compiler for reasons that aren’t clear to me. Looking to see if I should file a bug or if there’s something I’m doing wrong using this API.

Caused by: java.lang.AssertionError: IR backend shouldn't call KotlinTypeMapper.mapType

Repo is here: https://github.com/ZacSweers/auto-service-kt, stacktrace in thread

Hi, I get an internal compiler error to my project right after I updated to 1.5.0 from 1.4.3. The problem seems to be when trying to invoke an extension function with some lambdas as parameters. I opened an issue here https://youtrack.jetbrains.com/issue/KT-46557

I get

IllegalStateException: IrTypeAliasSymbol expected: Unbound public symbol for public kotlinx.coroutines/CancellationException|null[0]

compiling KMM module for Kotlin/Native with Kotlin 1.5: https://youtrack.jetbrains.com/issue/KT-46697

Hello, I found that after 1.5.0 upgrade lambdas are compiled as method, like:

private static final void setupStrictMode$lambda-2(com.vinted.MDApplication, android.os.strictmode.Violation);
  private static final void safePenaltyListener$lambda-3(com.vinted.MDApplication, android.os.strictmode.Violation);

Is it a new compiler feature, or JVM 8 support?

This is probably a shot in the dark, but hoping to get some guidance on where to start. I have a plugin into a android map app written in java but the plugin is 100% kotlin. After updating to kotlin 1.5, everything worked fine in debug but in release with obfuscation, got an abstract method error crash. Long story short, setting useoldjvmbackend to true fixes the issue. It would probably be hard to provide a sample repo for this scenario since it's a plugin to a java map app,but I could see someslight differences in the mapping.txt file when using the old jvm backend and the new one. Production wise, can move forward with using 1.5 but sucks to not be able to realize all the improvements the team is making with the new IR.

I see this difference with the way garbage collection is handled after jvm-ir-backend change. In this code,

println(weakStateRef.get())

used to return

null

prior to upgrading. Now, it is not cleared when `System.gc()`is called. Is this expected?

data class State(val stateKey: String? = null)

fun func1(newState: State) {
    // do nothing
}

fun func2() {
    val weakStateRef = WeakReference(State("Foo"))
    func1(weakStateRef.get()!!)
    System.gc()
    println(weakStateRef.get()) // State(stateKey="Foo")
}

Hi, I’m trying multimodule with kotlinspring but i had an error like this. Help plz…. I'm having this error when I click gradle compileKotlin or gradle build.

hi all! not sure if I'm asking in the right channel, but has anyone come across this issue and knows what could cause it? happens when I try running tests.

# A fatal error has been detected by the Java Runtime Environment:
#
#  SIGSEGV (0xb) at pc=0x00007f2734cd8aac, pid=19064, tid=19263
#
# JRE version: OpenJDK Runtime Environment Temurin-17.0.2+8 (17.0.2+8) (build 17.0.2+8)
# Java VM: OpenJDK 64-Bit Server VM Temurin-17.0.2+8 (17.0.2+8, mixed mode, sharing, tiered, compressed oops, compressed class ptrs, g1 gc, linux-amd64)
# Problematic frame:
# V  [libjvm.so+0x8d8aac]  jni_SetStaticIntField+0x8c

FYI

archived the channel