Despite this being accidental, I think I'll take advantage of this until

nosuspend

or similar becomes a thing, to prevent accidental bad usage of the APIs I designed (to use

Bundle

on Android with properties on a shared

object

)

I think a clearly kdoc'ed function is worth more than such a hack that might risk being "fixed" soon, people have been known to exclude libs dependencies to get higher versions of others to avoid issues with tons pf alrady written code... but I personally always ctl-q any fun I'm not sure about...

Of course a proper KDoc is important, but I don't see how this would stop working. There's no way a

crossinline

lambda could allow suspension implicitly when you think about it.

Hacks are hacks, even if it would work, the fact that someone else wouldn't understand how the code actually works is already code smell. I found myself looking at source code of well documented functions to reason about some case the devs didn't think about to see if it would work in my situation. Using undocumented/unintended functionality prevents that. You can disagree, but I know alot of people stay away from libs that do too much 'magic'.... 😉

Who said preventing suspension would prevent to document it?

Didn't mean that, but I guess each one with their style and habits 🙂