I understand the reason.

I found some source and other IO functions which read the source code from file. On my courses, these scripts are stored with encoding UTF-8, but the default chinese encoding for Chinese Traditional windows users are BIG5. Similar issues might be occurred for other non-English courses.

The solution is to add the argument encoding = "UTF-8", in my case, to these functions or connections. More generally, providing the default encoding or a hook before parsing might be a more robust solution.

Best,

你好，请问能否是在哪里这么设置？

在代码后面加上encoding='UTF-8'就可以了，如下图