Kotlin Multiplatform multithreading old and new way

on YouTube ↗42m

" summary

Kotlin Multiplatform concurrency historically forces Kotlin/Native code to use freezing and custom dispatchers to avoid frozen-exceptions; the talk shows a working HTTP client built on iOS URLSession that uses Workers plus an atomic-reference byte buffer and explicit clearing to avoid leaks, and explains detached object graphs, channels/flows and CompletableDeferred for lambda/callback bridging. The claimed payoff: with the new memory model (enable kotlin.native.binary.memoryModel=experimental) freezing is removed, default dispatchers work on iOS and multi-threaded GC fixes the main failure modes.

AI-written summary · about

" chapters (12)

1. 0:00Introduction
2. 1:30Kotlin Multiplatform overview
3. 5:26Old concurrency approach
4. 10:48Thread safety in Kotlin Native
5. 14:41Native tools for concurrency
6. 23:39Detached object graph
7. 25:26Combining coroutines with flows
8. 28:22New memory model introduction
9. 32:18Migrating to new model
10. 34:51Benefits of new model
11. 37:42Flows in Kotlin Multiplatform
12. 40:54Summary and Q&A

" related talks

• 58mKotlin Multiplatform: Memory management old and new wayAnna Zharkova · Android Worldwide 2022
• 44mKotlin/Native Concurrency for Mobile MultiplatformKevin Galligan · Conference for Kotliners 2020
• 39mKotlin Native Concurrency ExplainedKevin Galligan · droidcon Toronto 2019
• 39mKotlin Native Concurrency ExplainedKevin Galligan · KotlinConf 2019
• 46mKotlin/Native Concurrency ModelNikolay Igotti · KotlinConf 2018
• 41mCoroutines for Kotlin Multiplatform in practiseChristian Melchior · droidcon Berlin 2021
• 36mKotlin Multiplatform at five yearsRussell Wolf · droidcon New York 2022
• 26mKotlin/Native for Multiplatform DevelopmentEnrique López Mañas · Conference for Kotliners 2019