Concurrency in Haskell with Streamly

on YouTube ↗52m

" summary

Claims a practical recipe for fast, reliable concurrency in Haskell using Streamly streams plus MVars, TVars, STM and ReaderT IO, shows code on GitHub and slides, demonstrates how async/link propagates exceptions, how STM retry prevents invalid updates, and how producer/consumer patterns avoid MVar deadlocks. Examples include merging concurrent API calls (Hacker News) with parallel maps and converting between streams and TVars; the implemented examples run and expose where race conditions or livelock can still occur.

AI-written summary · about

" chapters (33)

1. 0:00Intro
2. 2:11Concurrency & parallelism: definitions
3. 3:43Concurrency & parallelism: basics
4. 4:38Scope of the talk
5. 6:12Terminology
6. 10:02Example of why sharing is hard
7. 12:21Solutions to race conditions non-exhaustive Don't just read and write RAM Wily nily in a multi-threaded program
8. 14:02How does Haskell's purity prevent race condit
9. 14:29Limitations of pure code for concurrency
10. 15:04Haskell threads
11. 16:06Async (a thin abstraction over threads)
12. 16:32Exception propagation using async
13. 17:04Three tools for architecting the impure parts of concurrent Haskell program
14. 17:54MVar mutual exclusion of reading and writing
15. 18:14Reading an MVar
16. 19:11Writing to an MVar
17. 21:26Deadlocks, synchronization, and MVars
18. 22:02Examples of MVar deadlocks
19. 22:54When is MVar code deadlock-free?
20. 23:41What happens when MVar code has deadlock
21. 25:55Practical applications
22. 29:18Software transactional memory vs. MVars
23. 30:20Deadlocks and race conditions using STM
24. 31:10STM example
25. 33:05TVar race condition example
26. 33:56Using TVars for synchronization
27. 36:58Types of streams
28. 38:13Stream types
29. 40:13Creating a stream using an MVar
30. 41:19Creating a TVar using a stream
31. 42:23Creating a stream using a TVar
32. 43:01Circular dependencies: streams, MVars, and
33. 45:25Limitations & possible solutions

" related talks

• 38mAccidentally ConcurrentEvan Czaplicki · Code Mesh 2015
• 32mPRACTICAL REACTIVE STREAMS WITH MONIXJacek Kunicki · Scalar 2018
• 42mLearn you Akka Streams for great Good!Heiko Seeberger · Scala Days Copenhagen 2017
• 27mAsync/Await: Thinking beyond CombineRitesh Gupta · iOS Conf SG 2024
• 48mAsynchronous streams in direct style with and without macrosPhilipp Haller · Scala Days Lausanne 2019
• 46mStreams, effects and beautiful folds, a winning trilogyEric torreborre · Scala Days Copenhagen 2017
• 25mFunctional effect and streaming systems in ClojureLéo Noel · re:Clojure 2021
• 38mConcurrency doesn,t have to be hard: Kotlin Coroutines and ChannelsJag Saund · Android Makers 2019