Is it worth switching to Bazel to speed up AOSP builds?

We had a similar debate in our Android team last year. We initially planned for a long migration but reconsidered when we realized the timeline and costs involved. Instead, we went with Incredibuild. It accelerated our builds without requiring us to tear down our current setup. It took just a few weeks to implement, and it handles distributed compilation and shared caching really well—no retraining required!

Dropping a bunch of cash on a high-performance desktop PC helped us slash build times significantly—down to just 7 minutes from 45! Sometimes a hardware upgrade can work wonders without needing a complete system overhaul.

I find Bazel to be a real pain to manage. The setup and upkeep can get complicated. If you go down that road, make sure you have an expert on board to keep things running smoothly. It's amazing when it works, though!

It looks like you're hoping for an easy fix, but let’s be real—there's no magic solution. However, tuning parallel builds and ccache can save you about 30-40% of your build time without all the hassle of a seven-month migration.

Are you doing clean builds for every compile? AOSP can be time-consuming, but if you cache your initial build, subsequent builds should just take minutes. Don't treat it like a quick-and-dirty app in your pipeline!