:mega: :mega: :mega: kotlinx.coroutines 1.6.0-RC3 is here! Mostly to fix the `Flow.collect` broken ...

# coroutines

[Dmitry Khalanskiy [JB]]

📣 📣 📣 kotlinx.coroutines 1.6.0-RC3 is here!

Mostly to fix the Flow.collect broken in RC2.

Full changelog: https://github.com/Kotlin/kotlinx.coroutines/releases/tag/1.6.0-RC3

[louiscad]

Can we get a native-mt version of it?

[Dmitry Khalanskiy [JB]]

We won't be able to publish it soon, sorry.

[spierce7]

no native-mt is blocking from upgrading to Kotlin 1.6

[Dmitry Khalanskiy [JB]]

Scott, this may be true in your case, but in general, the need for native-mt is reduced in 1.6.0-RC now that the main coroutines artifact supports Native multithreading with the new memory model: https://github.com/Kotlin/kotlinx.coroutines/issues/2914

[spierce7]

the need for native-mt is reduced

@Dmitry Khalanskiy [JB] Is the new memory model safe to use in production?

[Vsevolod Tolstopyatov [JB]]

This is a tricky question.

This is an experimental technology with all known issues and limitation described here.
It’s worth getting familiar with this document and decide for yourself whether it is safe for you or not.

What I can say is that this new MM is our bet -- all internal tests are ran mostly with new MM, coroutines are now abandoning native-mt in the favour of new MM and so on, so it’s not a one-shot experiment that we will throw away.

[spierce7]

Until the memory model is safe to use in production, I’m forced to rely on native-mt