If I compile with

1.0.4-eap-62

, does my compiler change each time JetBrains update the EAP ? My project was building fine yesterday but now I get a compile error on a large string

Error:(4, 1) org.jetbrains.kotlin.codegen.CompilationException: Back-end (JVM) Internal error: Failed to generate property Playlist_Raw_Json
Cause: java.lang.IllegalArgumentException
File being compiled and position: (4,1) in /Users/francois/dev/code/LostContext-App/app/src/main/java/com/lostincontext/data/playlist/JSON.kt
PsiElement: val Playlist_Raw_Json = """

(I have a fairly large string since I am writing a first prototype where I have a raw JSON in the code before I switch to a live API call in the next iteration)

no, we don't update already existing uploaded binaries

@cy If your service's SLA is response time but you generally have low throughput, then short pauses allow you to guarantee response time at the cost of more overall time spent on GC.

Have a 5 second pause once every 2 hours with a response time SLA of 50ms means that you miss your SLA once every 2 hours.

If instead you have a 10ms pause every second, you end up paying way more time doing GC, but you never miss your SLA.

@zoltu so you have to set pause limit to something in the middle (200ms by default)

@cy That was my thought as well but since that was the only way I could explain this break, it is very weird then

@francois it could be caused by the incremental compilation

don't you use any CI ?

it happens for a clean build as well 😕 I don't use a CI since it is a small side project I am starting. I should set one up when I have the time, but I would still need to be able to compile from my machine 🙂

there are lots of financial trading shops or video game server-type apps that really need low pause times

or any web site that's tuned heavily for latency

e.g. an entire google websearch takes ~200msec or less, so if you take a 100msec GC whilst processing that request you just increased your latency by a large amount

and when sites slow down people use them less, so that directly costs you $$$

this is why azul can make money by charging high prices for their jvm, when other jvms are free

Im creating a data object with a list as a parameter. I attempted to create an additional setter for the property (so I could some extra processing to other variables,) and I'm getting an error saying that there is already a public final version of setPropertyName. Is this intended/some sort of protection? The same setup has worked with other types of properties

if you have a property named

propertyName

, it will have a compiler-generated setter called

setPropertyName

. you cannot define your own method with the same name and parameter list, no matter what the property type is

also having multiple setters for the same property will likely lead to maintenance problems

perhaps there is a more kotlin-like way to accomplish what I am doing, then. One of my other properties is internal and is directly related to the one I am setting. I would like the other property to update as soon as the one I am currently working with updates. Would this be possible to do in a data class?

you can define additional properties in the body of the data class; they won’t be used in the generated equals/hashCode/toString though

Interesting; Thank you!

Hello - I've been having some troubles with a library I'm working on and compatibility with the Kotlin runtime.

The Kotlin runtime includes the IntelliJ annotations: https://github.com/JetBrains/kotlin/tree/master/core/runtime.jvm/src/org/jetbrains/annotations

This makes it difficult for my pure java library to also include them.

Anyone who uses both gets duplicate class definition errors.

It looks like I'm headed towards just removing the annotations from my java lib because it's not worth the trouble, but I'm open to any better suggestions.

any chance you can just include JSR 305?

I can look into that, sure.

yeah I mean, I've configured intelliJ to use it, and the kotlin plugin and intelliJ seem to respect it