I might understand. Thank you! 🙂 Is such overhead so large that I should often care about ?

Generally you don’t need to worry about it. I usually only do it when I’m using a function to wrap a single function call or expression. I may also do it if I know I need to return/continue from a passed lambda

I generally do

inline

where it's possible. While there is a not that relevant performance impact, the main reason is that an inline function is independent from

suspend

contexts, thus, when calling your function from a

suspend

fun, you can use other suspend fun's within the lambda if it is inline

Just be sure not to overuse it, as it can increase your binary size if used frequently enough and (I think) make stack traces inaccurate