Lihang Xu

> This works for devices whose natural orientation is potrait but may have the opposite effect for devices which are landscape natural orientation In the docs: https://developer.android.com/training/camera2/camera-preview-large-screens#camera-orientation, it says that...

Using orientation seems a better solution. I also find a page on detecting the natural orientation: http://www.java2s.com/example/android/user-interface/get-device-natural-orientation.html

I've updated the code to use orientation.

不敢恭维，找了一个标题内容，觉得没有机器翻译的好。而且我提的计算机辅助翻译主要还是人来翻，不是全自动机器翻译。 原文： >Unpacking a Sequence into Separate Variables 谷歌翻译： > 将序列拆分为单独的变量 社区翻译： > 解压序列赋值给多个变量 另外还有和人邮出版的翻译版对比： 

Since react-native-camera is deprecated in favor of react-native-vision-camera, maybe we should turn to vision-camera and its frame processor plugins like [vision-camera-dynamsoft-barcode-reader](https://github.com/xulihang/vision-camera-dynamsoft-barcode-reader) and [vision-camera-code-scanner](https://github.com/rodgomesc/vision-camera-code-scanner)

Turned to html to image in the end.

I use react-native-svg to draw the overlay and pass a scan region argument to the frame processor to crop the image before processing. You can find an example here: https://github.com/tony-xlh/react-native-mrz-scanner/blob/main/src/screens/Scanner.tsx

JavaFX may not support Tibetan input methods. If there is a way to input without this input method, it may work.

I found that it is possible to stack "a" with SHIFT key pressed using the Himalaya input method.

Which JavaFX program are you using? You can replace ImageTrans's jre with its jre to have a try.