Breaking the build

on YouTube ↗41m

" summary

Companies report massive scale problems: one codebase has 500 modules and another had 40–50 minute builds before changes. Teams measure time-to-merge and build medians, use Gradle Profiler for repeatable benchmarks, Graph Assert to find high betweenness modules, and Foundry or custom convention plugins to generate modules. Common-module bottlenecks get class-by-class extraction, plugins get published as binaries to avoid build churn, and some teams migrate to Bazel or watch the Declarative Gradle EAP for long‑term gains.

AI-written summary · about

" chapters (16)

1. 0:00Welcome
2. 2:28Android build 2025
3. 5:46Dedicate resources to builds
4. 8:53Build performance metrics
5. 12:20Modularization strategies
6. 16:57Gradle plugins and static analysis
7. 18:00Bazel on Android and iOS
8. 23:56Bazel migration worthiness
9. 25:08Custom Gradle plugins
10. 27:19Declarative Gradle future
11. 28:22Gradle abstraction debate
12. 29:25Bottleneck modules breakdown
13. 31:33Gradle best practices
14. 35:08Monorepo vs polyrepo builds
15. 38:00Platform team challenges
16. 40:52Q&A

" related talks

• 36mImproving Gradle BuildsMohit Sarveiya · Android Worldwide 2022
• 40mNail your Gradle build timeJosef Raska · droidcon Berlin 2021
• 37mThe Build ShruggedInaki Villar · Droidcon Vietnam 2019
• 35mGradle recipes for reducing your build timesAdam Ahmed · droidcon London 2022
• 44mGradle: The Build System That Loves To Hate YouAurimas Liutikas · droidcon London 2024
• 39mWorking with Alternative Build SystemsGautam Korlam · droidcon SF 2019
• 35mGradle + CI vs. Modulezilla: Winning the WarTom Belk, Dom Leone · Android Summit 2019
• 37mModularization - flatten your graph and get the real benefitsJosef Raska · droidcon London 2022