<KT-75400>: crash the (Kotlin/JVM) compiler with t...

# compiler

[Youssef Shoaib [MOD]]

KT-75400: crash the (Kotlin/JVM) compiler with this one weird trick!

fun bar() {
  foo { return }
}

inline fun foo(block: () -> Nothing) {
  try {
    block()
  } finally {
    
  }
}

Any status update on JVM using IR-inlining? IR-inlining seems to be more robust

[udalov]

IR inlining for JVM will most likely be discontinued

[Youssef Shoaib [MOD]]

Oh! I thought it had some advantages wrt better optimizations (and ofc, sharing the work between platforms), but was it a bigger effort than expected or something? I trust in your judgement, I'm just curious because it seemed a natural evolution after K2

[udalov]

Yes, it turned out to be a bigger effort, whose gains are not immediately clear because we still have to support bytecode inliner indefinitely anyway