This suffers from limitations in the other fallback code, namely that Hans and Hant are not tried because everything gets collapsed into Hani. The specific way that this should be handled depends on the order of system locales on some platforms.

I went ahead and added a workaround that checks if some lang fallback includes the Hant script, and adds that also for Hani. This should be locale-dependent in practice but this makes ambiguous han characters at least show up for now.

One of the limitations of this is that it only functions when postScriptName is present on at least one of the <font... elements in a fallback family. This is most of the language and script fallback families on common devices, but not all of them; on my device this misses the fallback families for Hebrew.

I'd be happy to test this on my device if there's a procedure for that, but I don't have the expertise to work out what's happening very quickly, and so could only give a rubber stamping review.

In which case, it should probably be given by Chad.

I'd be happy to test this on my device if there's a procedure for that, but I don't have the expertise to work out what's happening very quickly, and so could only give a rubber stamping review.

In which case, it should probably be given by Chad.

There is https://github.com/xorgy/xilem/tree/andtaffy if you have a local parley hooked up in the Cargo.toml. I primarily test this stuff on desktop (you just need to swap the "fontconfig/mod.rs" path for "android.rs" in your target's cfg of mod system;... and then you need to set $ANDROID_ROOT to a directory which has the equivalent structure (adb pull /system/etc/fonts.xml and /system/fonts/ into a dir, and point $ANDROID_ROOT TO THE system dir that you create). On a real Android device, $ANDROID_ROOT is set in the environment.