Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
Kotlin State & Shared Flows in Action
Search
Mohit S
July 27, 2021
Programming
4
1.4k
Kotlin State & Shared Flows in Action
Shared Flow APIs
Mohit S
July 27, 2021
Tweet
Share
More Decks by Mohit S
See All by Mohit S
Guide to Improving Compose Performance
heyitsmohit
0
280
Building Shared UIs across Platforms with Compose
heyitsmohit
1
670
Building Multiplatform Apps with Compose
heyitsmohit
2
560
Building StateFlows with Jetpack Compose
heyitsmohit
6
1.9k
Building Android Testing Infrastructure
heyitsmohit
1
550
Migrating to Kotlin State & Shared Flows
heyitsmohit
1
820
Using Square Workflow for Android & iOS
heyitsmohit
1
460
Building Android Infrastructure Teams at Scale
heyitsmohit
3
360
Strategies for Migrating to Jetpack Compose
heyitsmohit
2
610
Other Decks in Programming
See All in Programming
The Art of Re-Architecture - Droidcon India 2025
siddroid
0
160
例外処理とどう使い分ける?Result型を使ったエラー設計 #burikaigi
kajitack
16
4.8k
perlをWebAssembly上で動かすと何が嬉しいの??? / Where does Perl-on-Wasm actually make sense?
mackee
0
300
Graviton と Nitro と私
maroon1st
0
160
Context is King? 〜Verifiability時代とコンテキスト設計 / Beyond "Context is King"
rkaga
10
1.5k
re:Invent 2025 のイケてるサービスを紹介する
maroon1st
0
160
AIエージェントの設計で注意するべきポイント6選
har1101
6
3k
0→1 フロントエンド開発 Tips🚀 #レバテックMeetup
bengo4com
0
470
副作用をどこに置くか問題:オブジェクト指向で整理する設計判断ツリー
koxya
1
300
Grafana:建立系統全知視角的捷徑
blueswen
0
280
Flutter On-device AI로 완성하는 오프라인 앱, 박제창 @DevFest INCHEON 2025
itsmedreamwalker
1
190
Findy AI+の開発、運用におけるMCP活用事例
starfish719
0
2.1k
Featured
See All Featured
How GitHub (no longer) Works
holman
316
140k
How to train your dragon (web standard)
notwaldorf
97
6.5k
Distributed Sagas: A Protocol for Coordinating Microservices
caitiem20
333
22k
How to Create Impact in a Changing Tech Landscape [PerfNow 2023]
tammyeverts
55
3.2k
Building Better People: How to give real-time feedback that sticks.
wjessup
370
20k
The Illustrated Guide to Node.js - THAT Conference 2024
reverentgeek
0
220
Sharpening the Axe: The Primacy of Toolmaking
bcantrill
46
2.6k
We Have a Design System, Now What?
morganepeng
54
8k
Building Adaptive Systems
keathley
44
2.9k
A designer walks into a library…
pauljervisheath
210
24k
Fashionably flexible responsive web design (full day workshop)
malarkey
408
66k
A Modern Web Designer's Workflow
chriscoyier
698
190k
Transcript
Mohit Sarveiya Kotlin State & Shared Flows in Action @heyitsmohit
Kotlin State & Shared Flows in Action • State Flow
• Shared Flow • Broadcast Channel vs shared flow • Convert cold streams to shared flows • Buffer Overflow Strategies
Cold vs Hot Flows
What is a cold stream? A cold stream is a
flow that triggers the same code every time it is collected.
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 }
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 } flow.collect { ... }
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 } flow.collect { ... }
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 } flow.collect { ... }
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 } flow.collect { // 2, 3, 4 }
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 } flow.collect { ... } flow.collect { ... }
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 } flow.collect { ... } flow.collect { ... }
Cold Flows val flow = flowOf(1, 2, 3) .map
{ it + 1 } flow.collect { ... } flow.collect { // 2, 3, 4 }
What is a hot stream? A hot stream is a
flow whose active instance exists independently of the presence of collectors.
Hot Streams State Flow Shared Flow
State Flow View View Model
State Flow View View Model
State Flow View View Model State
State Flow sealed class UiState { data class Error(
val exception: Throwable ): UiState() }
State Flow sealed class UiState { data class Success(
val data: Data ): UiState() data class Error( val exception: Throwable ): UiState() }
State Flow sealed class UiState { data class Success(
val data: Data ): UiState() data class Error( val exception: Throwable ): UiState() }
State Flow val uiState = MutableStateFlow()
val uiState = MutableStateFlow( UiState.Success(Data()) ) State Flow
State Flow val uiState = MutableStateFlow( . .. ) uiState.emit(
UIState.Success(Data()) )
State Flow val uiState = MutableStateFlow( . .. ) uiState.value
= UIState.Success(Data())
State Flow val uiState = MutableStateFlow( . .. ) uiState.collect
{ ... } Latest value is received
State Flow val uiState = MutableStateFlow( . .. ) uiState.collect
{ ... } uiState.collect { ... } Latest value is received
State Flow Conflation val uiState = MutableStateFlow( . .. )
uiState.value = UIState.Success( .. . ) uiState.value = UIState.Error( .. . ) Conflate
State Flow Conflation val uiState = MutableStateFlow( . .. )
uiState.value = UIState.Success( .. . ) uiState.value = UIState.Error( .. . ) uiState.collect { ... } Error
State Flow vs Live Data State Flow Live Data Default
Value Unsubscribe (Stopped State)
State Flow vs Live Data State Flow Live Data Default
Value Unsubscribe (Stopped State)
State Flow vs Live Data State Flow Live Data Default
Value Unsubscribe (Stopped State)
State Flow Summary • How to setup state flow •
Emit and collect • State Flow vs Live Data
Shared Flow
Shared Flow
Shared Flow Consumer 1 Consumer 2
Shared Flow Consumer 1 Consumer 2
Shared Flow Consumer 1 Consumer 2 Event Event
Shared Flow Consumer 1 Consumer 2 Replay Replay
Shared Flow Buffer
Shared Flow val flow = MutableSharedFlow < > ()
Shared Flow val flow = MutableSharedFlow<String>()
Shared Flow val flow = MutableSharedFlow<String>()
Shared Flow val flow = MutableSharedFlow<String>() launch { flow.collect {
} }
Shared Flow val flow = MutableSharedFlow<String>() launch { flow.emit("Event
1") } launch { flow.collect { } }
Shared Flow val flow = MutableSharedFlow<String>() launch { flow.emit("Event
1") } launch { flow.collect { } } Event 1
Shared Flow val flow = MutableSharedFlow<String>() launch { flow.emit("Event
1”) } launch { delay(2000); flow.collect { } }
Shared Flow val flow = MutableSharedFlow<String>() launch { delay(2000);
flow.collect { } } launch { flow.emit("Event 1”) }
Shared Flow val flow = MutableSharedFlow<String>() launch { delay(2000);
flow.collect { } } launch { flow.emit("Event 1”) }
Shared Flow val flow = MutableSharedFlow<String>() launch { delay(2000);
flow.collect { } } No value is received launch { flow.emit("Event 1”) }
Shared Flow val flow = MutableSharedFlow<String>() launch { flow.emit("Event
1”) } launch { delay(2000); flow.collect { } } Replay
Shared Flow val flow = MutableSharedFlow<String>(replay = 1) launch
{ flow.emit("Event 1”) } launch { delay(2000); flow.collect { } }
Shared Flow val flow = MutableSharedFlow<String>(replay = 1) launch
{ delay(2000); flow.collect { } } launch { flow.emit("Event 1”) } Event 1
val flow = MutableSharedFlow<String>(replay = 1) Shared Flow State
Flow launch { flow.collect { } } launch { flow.subscriptionCount.value }
Shared Flow val flow = MutableSharedFlow<String>(replay = 1) 1 Subscriber
launch { flow.collect { } } launch { flow.subscriptionCount.value }
Shared Flow val flow = MutableSharedFlow<String>(replay = 1) Shared Flow
does not complete normally launch { flow.collect { } }
Cold Flows val flow = flowOf(1, 2, 3) flow
.onCompletion { } .collect { ... } Flow completes normally
Shared Flow val flow = MutableSharedFlow<String>(replay = 1) val job
= launch { flow.collect { } } job.cancel()
Shared Flow val flow = MutableSharedFlow<String>(replay = 1) val job
= launch { flow.onCompletion { }.collect { } } job.cancel() Flow completes exceptionally
Shared Flow Summary • Setup • Replay and emit •
Cancellation
Broadcast Channel vs Shared Flow
Broadcast Channel Shared Flow Channel APIs Replay Buffer Closed
val channel = BroadcastChannel<Int>(10)
val channel = BroadcastChannel<Int>(10) channel.send( ... )
val channel = BroadcastChannel<Int>(10) channel.send( ... ) channel.close()
val stateFlow = MutableSharedFlow() stateFlow.emit( ... )
Broadcast Channel Shared Flow Channel APIs Replay Buffer Closed
val stateFlow = MutableSharedFlow(replay = 2) stateFlow.emit( ... )
Broadcast Channel Shared Flow Channel APIs Replay Buffer Closed
val channel = BroadcastChannel<Int>(capacity = 10) channel.send( ... )
val stateFlow = MutableSharedFlow( replay = 2, extraBufferCapacity = 10
) stateFlow.emit( ... )
Broadcast Channel Shared Flow Channel APIs Replay Buffer Closed
val channel = BroadcastChannel<Int>(10) channel.send( ... ) channel.close()
val stateFlow = MutableSharedFlow( replay = 2, extraBufferCapacity = 10
) stateFlow.emit( ... )
Broadcast Channel Shared Flow Channel APIs Replay Buffer Closed
Broadcast Channel Shared Flow
val channel = BroadcastChannel<Int>(capacity)
val channel = BroadcastChannel<Int>(capacity) val flow = MutableSharedFlow<String>(extraBufferCapacity)
channel.send( ... ) channel.trySend( ... ) flow.emit( ... ) flow.tryEmit(
... )
Broadcast Channel Shared Flow Channel APIs Replay Buffer Closed
Flow Sharing Strategies
Cold Flow Hot Flow Convert
Sharing Policies • While Subscribed • Eagerly • Lazily
Sharing Policies flow
Sharing Policies flow.shareIn( )
Sharing Policies flow.shareIn( externalScope, )
Sharing Policies flow.shareIn( externalScope, replay = 1, )
Sharing Policies flow.shareIn( externalScope, replay = 1, started = SharingStarted.WhileSubscribed()
)
Sharing Policies val sharedFlow = flow.shareIn( externalScope, replay = 1,
started = SharingStarted.WhileSubscribed() )
While Subscribed • Active as long as external scope is
alive • Remains as long as there are collectors.
Properties Active as long as external scope is alive
Properties Active as long as external scope is alive sharedFlow.collect
{ } Subscriber
flow.shareIn( externalScope, replay = 1, started = SharingStarted.WhileSubscribed() ) Properties
Active as long as external scope is alive
Properties Active as long as external scope is alive sharedFlow.collect
{ } externalScope.cancel()
Properties Active as long as external scope is alive sharedFlow.collect
{ } externalScope.cancel() Complete Exceptionally
Properties Remains as long as there are collectors.
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…)
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…) val job = launch { sharedFlow.onCompletion { }.collect { } }
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…) val job = launch { sharedFlow.onCompletion { }.collect { } } job.cancel()
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…) val job = launch { sharedFlow.onCompletion { }.collect { } } job.cancel()
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…) val job = launch { sharedFlow.onCompletion { }.collect { } } job.cancel()
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…) val job1 = launch { sharedFlow.collect { } } val job2 = launch { sharedFlow.collect { } }
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…) job1.cancel() val job2 = launch { sharedFlow.collect { } }
Properties Remains as long as there are collectors. val sharedFlow
= flow.onCompletion { }.shareIn(…) job1.cancel() val job2 = launch { sharedFlow.collect { } } Remain Active
Properties • Active as long as external scope is alive
• Remains as long as there are collectors.
Sharing Policies • While Subscribed • Eagerly • Lazily
Eagerly flow.shareIn( externalScope, replay = 1, started = SharingStarted.Eagerly() )
Eagerly Start producer eagerly and never stop flow .onStart {
println("ON START") } .shareIn( ... started = SharingStarted.Eagerly)
Eagerly Start producer eagerly and never stop flow .onStart {
println("ON START") } .shareIn( ... started = SharingStarted.Eagerly)
Eagerly Start producer eagerly and never stop flow .onStart {
println("ON START") } .shareIn( ... started = SharingStarted.Eagerly) ON START
Eagerly Start producer eagerly and never stop flow .onComplete {
println("ON COMPLETE”) } .shareIn( ... started = SharingStarted.Eagerly)
Eagerly Start producer eagerly and never stop flow .onComplete {
println("ON COMPLETE”) } .shareIn( ... started = SharingStarted.Eagerly) externalScope.cancel()
Eagerly Start producer eagerly and never stop flow .onComplete {
println("ON COMPLETE”) } .shareIn( ... started = SharingStarted.Eagerly) Never stops externalScope.cancel()
Eagerly Start producer eagerly and never stop
Sharing Policies • While Subscribed • Eagerly • Lazily
Lazily Start sharing after the first subscriber appears and never
stop
Lazily flow.shareIn( externalScope, replay = 1, started = SharingStarted.Lazily )
Lazily flow .onStart { println("ON START") } .shareIn(…,started = SharingStarted.Lazily)
Lazily flow .onStart { println("ON START") } .shareIn(…,started = SharingStarted.Lazily)
launch { sharedFlow.collect { } }
Lazily flow .onStart { println("ON START") } .shareIn(…,started = SharingStarted.Lazily)
launch { sharedFlow.collect { } } "ON START"
Lazily flow .onCompletion { println("COMPLETE") } .shareIn(…,started = SharingStarted.Lazily)
flow .onCompletion { println("COMPLETE") } .shareIn(externalScope,…,started = SharingStarted.Lazily) Lazily
flow .onCompletion { println("COMPLETE") } .shareIn(externalScope,…,started = SharingStarted.Lazily) Lazily externalScope.cancel()
flow .onCompletion { println("COMPLETE") } .shareIn(externalScope,…,started = SharingStarted.Lazily) Lazily Never
stops externalScope.cancel()
Lazily Start sharing after the first subscriber appears and never
stop
Sharing Policies • While Subscribed • Active while there are
active subscribers. • Eagerly • Start producer eagerly and never stop • Lazily • Start after the first subscriber appears and never stop
Buffer Overflow Strategies
Shared Flow Buffer
Shared Flow Producer Consumer
Shared Flow Producer Consumer Generating events fast
Shared Flow Producer Consumer Listening to events with
delay
Shared Flow Producer Consumer
Shared Flow Producer Consumer
Shared Flow Producer Consumer What happens when it is
full?
Buffering Overflow Strategies • Suspend • Drop oldest • Drop
latest
Shared Flow Producer Consumer Suspend
Buffering Overflow Strategies val flow = MutableSharedFlow<String>( extraBufferCapacity = 2,
onBufferOverflow = BufferOverflow.SUSPEND ) Buffer + Replay Count
Buffering Overflow Strategies val flow = MutableSharedFlow<String>( extraBufferCapacity = 2,
onBufferOverflow = BufferOverflow.SUSPEND )
Buffering Overflow Strategies launch { flow.emit("Event 1") flow.emit("Event 2") flow.emit("Event
3") } Suspend
Buffering Overflow Strategies • Suspend • Drop oldest • Drop
latest
Shared Flow Producer Consumer Drop Oldest
Shared Flow Producer Consumer Drop latest
Buffering Overflow Strategies • Suspend • Drop oldest • Drop
latest
Kotlin State & Shared Flows in Action • State Flow
• Shared Flow • Broadcast Channel vs shared flow • Convert cold streams to shared flows • Buffer Overflow Strategies
https: // codingwithmohit.com/coroutines/learning-shared-and-state-flows-with-tests/ Coding with Mohit
Thank You! www.codingwithmohit.com @heyitsmohit