You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, when you build a specific ABI for Android, we still tell CMake to build all the ABIs we support. This makes the build three times slower than it needs to be when you first build for a specific target ABI. This should be configured in Apps/Playground/node_modules/@babylonjs/react-native/android/build.gradle. An example of what I believe to be the correct way to configure this can be found here: https://stackoverflow.com/questions/55479984/pass-architecture-information-from-gradlew-command-to-cmake/55486008#55486008
The text was updated successfully, but these errors were encountered:
I need a bit of clarification here. In BN CI, a parameter has been added to only build arm64 https://github.com/BabylonJS/BabylonNative/blob/88d09c881d81d0ae4889098c2aa9da53e0d929db/.github/jobs/android.yml#L32
For BRN, we can do the same but it needs some modification in gulp.js to deduce or forward the ABI parameter.
And when using npm run android command, I'm not sure it can deduce the type of ABI needed when starting the build process (emulator or arm64 connected phone)
So, to sumarize it, what are the use cases for limiting the ABIs built?
Does it make sense to have prebuilt binaries for some BN libraries instead?
This issue is pretty old, but I believe the primary intent here was just to improve the dev inner loop (specifically local builds from Android Studio). I think my expectation would be:
Building from Android Studio: respect the currently selected cpu architecture
Building via npm command: build all ABIs, or maybe require a new argument to select an ABI if we thought that would be useful
Building via gulp / build pipeline: build all ABIs
So I think really this issue is just about the first bullet, and I think the stack overflow solution linked above achieves that (use the project ABI is specified, otherwise fall back to all ABIs).
Currently, when you build a specific ABI for Android, we still tell CMake to build all the ABIs we support. This makes the build three times slower than it needs to be when you first build for a specific target ABI. This should be configured in Apps/Playground/node_modules/@babylonjs/react-native/android/build.gradle. An example of what I believe to be the correct way to configure this can be found here: https://stackoverflow.com/questions/55479984/pass-architecture-information-from-gradlew-command-to-cmake/55486008#55486008
The text was updated successfully, but these errors were encountered: