Remove Android specific abis from the export preset feature list

[m4gr3d]

The presence of those abis cause them to be included in the set of p_features passed to the gdextension_export_plugin#_export_file(...) method, which caused them to be lumped in the features_wo_arch set.
When trying to find the gdextension library path, we use a predicate with the following logic:

[features_wo_arch, arch_tag](String p_feature) { return features_wo_arch.has(p_feature) || (p_feature == arch_tag); }

For a gdextension config file like the one below, this causes the first android entry (android.armeabi-v7a = ...) to always be returned regardless of archs since it always satisfies the predicate.

[configuration]

entry_symbol = "example_library_init"
compatibility_minimum = 4.1

[libraries]
linux.x86_64 = "res://libgdexample.so"

android.armeabi-v7a = "res://libgdexample.android.template_release.armeabi-v7a.so"
android.arm32 = "res://libgdexample.android.template_release.armeabi-v7a.so"
android.x86 = "res://x86/libgdexample.android.template_release.x86.so"
android.x86_32 = "res://x86/libgdexample.android.template_release.x86.so"
android.x86_64 = "res://libgdexample.android.template_release.x86_64.so"
android.arm64-v8a = "res://libgdexample.android.template_release.arm64-v8a.so"
android.arm64 = "res://libgdexample.android.template_release.arm64-v8a.so"

Fixes #82727

[akien-mga]

These architectures/feature tags are meant to be the Godot-specific ones (arm32, arm64, x86_32, x86_64).

I'm not sure we should start adding all the architectures identifiers which may be used on specific platforms there (e.g. aarch64 on Linux, x86-64 on macOS, x64 on Windows, etc.). Keeping it standardized would be good. Maybe we just shouldn't add Android-specific arch identifiers as feature tags?

[bruvzg]

For a gdextension config file like the one below, this causes the first android entry (android.armeabi-v7a = ...) to always be returned regardless of archs since it always satisfies the predicate.

I would add a separate list of invalid architectures and skip ("unknown architecture" with error message) armeabi-v7a and arm64-v8a altogether. For consistency with other platforms (since everything else in Godot expect arm32 and arm64 instead).

[bruvzg]

Or maybe we should require to always specify architecture, currently it's not the case for macOS and iOS (since binary can be universal), but we can require to specify universal as architecture for these case and consider all entries without arch as invalid.

[m4gr3d]

These architectures/feature tags are meant to be the Godot-specific ones (arm32, arm64, x86_32, x86_64).

I'm not sure we should start adding all the architectures identifiers which may be used on specific platforms there (e.g. aarch64 on Linux, x86-64 on macOS, x64 on Windows, etc.). Keeping it standardized would be good. Maybe we just shouldn't add Android-specific arch identifiers as feature tags?

That could be the approach; we add the Android specific ones automatically in the feature list when the Godot specific arch is selected, so we can either disable that logic or provide an error message to indicate to the user they can't be used in the gdextension config.

[m4gr3d]

For a gdextension config file like the one below, this causes the first android entry (android.armeabi-v7a = ...) to always be returned regardless of archs since it always satisfies the predicate.

I would add a separate list of invalid architectures and skip ("unknown architecture" with error message) armeabi-v7a and arm64-v8a altogether. For consistency with other platforms (since everything else in Godot expect arm32 and arm64 instead).

To @akien-mga's point, the list would need to include the invalid architectures for all supported platforms to be consistent.

[m4gr3d]

I've updated the fix to remove the Android specific abis from the feature list.

[bruvzg]

If it's not used for something Android specific, it seems fine to me.

[akien-mga]

Thanks!

[YuriSizov]

Cherry-picked for 4.1.4.