It would be nice if IDE would know it's a conflate...

# announcements

[coder82]

It would be nice if IDE would know it's a conflated channel and not put the suspend arrow indicator when calling send() on it... maybe it was better a specific type than an initialization parameter

[gildor]

It’s not nice, because send() is still suspend function, even if it invoking very fast

[coder82]

is it? but the documentation says it won't suspend and offer returns always true

[gildor]

if you know that it’s conflated channel, just offer(), if for some reason you want to avoid suspend indicator

[coder82]

true

[gildor]

it won’t suspend

it’s just return immediately, but it still suspend function and cannot be called out of suspending lambda

[coder82]

ok

[Matej Drobnič]

Let me steal this thread with similar question: Does that mean that offer and send are competely identical with conflated channel or is send still slightly more expensive since it needs to drag whole coroutine machinery with it?

[gildor]

Yes, offer is a bit more efficient because doesn’t create suspend point as send

[Matej Drobnič]

so basically we should alwys use offer on conflated channels?

[Matej Drobnič]

Or is there a reason to use send?

[gildor]

I wouldn’t say that “we should”

[gildor]

reason to use send

If you want cancellation/suspend point

[gildor]

or just to make it more future proof, because essentially you must be 100% sure that it’s conflated channel