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
Taming the Monorepo Beast: How We Made AndroidX...
Search
Aurimas Liutikas
September 26, 2024
Technology
140
2
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
Taming the Monorepo Beast: How We Made AndroidX Library Releases At Scale More Bearable
Talk by Aurimas Liutikas (Google) at DPE Summit SF 2024
Aurimas Liutikas
September 26, 2024
More Decks by Aurimas Liutikas
See All by Aurimas Liutikas
Moments When Things Go Wrong
aurimas
3
200
The Benevolent Gradle Overlord: Keeping Order
aurimas
2
160
Gradle: Your Build, Your Rules
aurimas
2
100
APIs: How Hard Can They Be?
aurimas
1
74
Gradle: The Build System That Loves To Hate You
aurimas
4
3.1k
Ultimate Iteration Speeds with Gradle Configuration Cache
aurimas
2
230
More Effective Testing on Android Devices
aurimas
1
120
Gradle under the microscope
aurimas
2
220
Chasing the Speed of Gradle Builds: a history of Androidx build
aurimas
1
170
Other Decks in Technology
See All in Technology
Pavlokで始める電撃駆動開発
sgrsn
0
170
新しい SLO が良い感じにハマっている話
z63d
4
2.1k
Eight Engineering Unit 紹介資料
sansan33
PRO
3
8.1k
Bill One 開発エンジニア 紹介資料
sansan33
PRO
7
19k
【Google Cloud Next Tokyo'26】Gemini Enterprise と Oracle AI Database で実現する、業務データ活用を実現する AI エージェント実装
shisyu_gaku
0
210
AI時代の強いチームの作り方
yuukiyo
23
15k
Cursor Meetup Sapporo - Cursor物語 続編
cocacola917
0
120
TypeScript入門 2026
recruitengineers
PRO
1
170
ウォーターフォール開発案件のPMとしてAI活用を模索している話
hatahata021
3
310
AIエージェントの知識表現と推論に なぜグラフが使われるのか - 記号的AIの復権とニューラルAIとの統合
yohei1126
1
270
【CEDEC2026】コードレビュー支援ツール開発から学ぶ:LLMを用いた業務システムの実践的な運用設計と誤出力対策
cygames
PRO
0
510
Goでデータパイプラインを作ろう
sansantech
PRO
0
210
Featured
See All Featured
Reflections from 52 weeks, 52 projects
jeffersonlam
356
21k
Why You Should Never Use an ORM
jnunemaker
PRO
61
9.9k
First, design no harm
axbom
PRO
2
1.2k
The World Runs on Bad Software
bkeepers
PRO
72
12k
Prompt Engineering for Job Search
mfonobong
0
390
Helping Users Find Their Own Way: Creating Modern Search Experiences
danielanewman
31
3.3k
Claude Code のすすめ
schroneko
67
230k
Applied NLP in the Age of Generative AI
inesmontani
PRO
4
2.4k
What the history of the web can teach us about the future of AI
inesmontani
PRO
1
650
Color Theory Basics | Prateek | Gurzu
gurzu
0
410
From π to Pie charts
rasagy
0
250
Introduction to Domain-Driven Design and Collaborative software design
baasie
1
920
Transcript
Taming the Monorepo Beast How We Made AndroidX Library Releases
At Scale More Bearable Aurimas Liutikas / AndroidX @ Google @
[email protected]
What is androidx? Collection of widely used open-source libraries Main
Android API surface Broad range of functionality Developed by dozens of separate teams
Early Days Releases (com.android.support)
Early Days Releases Monolith releases a few times a year
21.0.3 21.0.2 21.0.0
Early Days Releases Uniform version with major based on Android
API level 21.0.3 21.0.2 21.0.0 22.2.1 22.2.0 22.1.1 22.1.0 22.0.0 23.4.0 23.3.0 23.2.1 23.2.0 23.1.1 23.1.0 23.0.1 23.0.0
Early Days Releases Eventually added alpha/beta 21.0.3 21.0.2 21.0.0 22.2.1
22.2.0 22.1.1 22.1.0 22.0.0 23.4.0 23.3.0 23.2.1 23.2.0 23.1.1 23.1.0 23.0.1 23.0.0 24.2.1 24.2.0 24.1.1 24.1.0 24.0.0 24.0.0-beta1 24.0.0-alpha2 24.0.0-alpha1 25.4.0 25.3.1 25.3.0 25.2.0 25.1.1 25.1.0 25.0.1 25.0.0 26.1.0 26.0.2 26.0.1 26.0.0 26.0.0-beta2 26.0.0-beta1 26.0.0-alpha1
No Longer Scaling (circa 2018) 60 libraries from ~13 teams
9 pages of release instructions Some libraries were monoliths of their own Bug fix in one library → 59 released with no changes Forced releases to align with Android OS releases
08 May 2018 Hello World, androidx Generated by Google AI
Image Generator
The New Plan Split and restructure libraries • À la
carte library adoption • Clear maven coordinate to package mapping ◦ androidx.foo:foo-bar →androidx.foo.bar Generated by Google AI Image Generator
The New Plan Adopt per library semantic versioning
The New Plan Adopt per library semantic versioning MAJOR.MINOR.PATCH
The New Plan Adopt per library semantic versioning MAJOR.MINOR.PATCH
The New Plan Adopt per library semantic versioning MAJOR.MINOR.PATCH
The New Plan Adopt per library semantic versioning MAJOR.MINOR.PATCH-(alpha|beta|rc)
The New Plan Depend on the lowest possible versions of
other libraries androidx.foo:foo:1.0.0 androidx.foo:foo:1.0.1 androidx.foo:foo:1.1.0 androidx.foo:foo:1.2.0 androidx.foo:foo:1.3.0 androidx.foo:foo:1.0.0 androidx.foo:foo:1.0.1 androidx.foo:foo:1.1.0 androidx.foo:foo:1.2.0 androidx.foo:foo:1.3.0 vs
Let’s execute the new plan!
Welp, this is a lot of manual work
Interconnectedness
Interconnectedness
Pinned vs Unpinned api(project(“:foo:foo-bar”)) vs api(“androidx.foo:foo-bar:1.1.0”)
Can we move to self-service?
Optional Release Trains (controlled chaos)
None
None
None
androidx-main vs release branches alpha & beta from androidx-main •
regular development rc & stable from androidx-foo-release • cherry-pick only
How Has It Gone?
Scale of androidx androidx migration Jetpad
Scale of androidx androidx migration Jetpad
Scale of androidx Jetpad
MISSION ACCOMPLISHED?
Current Pain Points Painful bugfix releases
Current Pain Points Painful bugfix releases Difficult build infrastructure backports
to release branches
Current Pain Points Painful bugfix releases Difficult build infrastructure backports
to release branches Kotlin Multiplatform artifact explosion
That’s all Folks!