TLDR: It's become obvious I need to dig deeper into Android app development.
My first smartphone (2014) was an LG930 Android. I chose Android because I wanted to write apps for it. I still carry an Android, and I've been publishing Android versions of my libraries since 2020, but somehow I never got around to writing and installing my own apps.
This goes against my stated philosophy of not publishing software I can't test. I knew people were using my libraries in Android apps, and I trusted they would inform me if there were serious issues. Apparently I was mistaken.
Back in October I finally created a very simple Android test app for my Jolt JNI library, and it seemed to work, so I relaxed. Then last week I wrote a more thorough test and found a major problem with my library that would cause any realistic Android app to fail: JVM code could invoke native C++ code, but if native code tried to call back into the JVM, the app would usually crash.
I hastily published a new release (v3.8.0) that solves the callback issue. Mea culpa.
I strongly suspect my older Libbulletjme library has an analogous bug, likely going back to 2020. But in order to be sure (and to test the fix, when I write one) I'll need an Android test app for Libbulletjme ASAP.
Meanwhile, all this tinkering in the Android SDK has reminded me how cool Androids are and re-awakened my interest in writing apps. That may sound like a digression on a digression, but I expect it will actually lead me back to the open-source project that originally got me interested in Jolt Physics, namely the JMonkeyEngine project.