Non-UTF-8 encoded Chinese characters in subprocess output on Windows cause a crash.

[tianzj]

In codewiki's str_replace_editor.py, the .decode() calls on subprocess output crash when encountering GBK-encoded Chinese characters on Windows.

[0xZKc0de]

I’m seeing a similar issue on Windows: when CodeWiki reads or writes files using Python’s open() without explicitly specifying encoding="utf-8", it can fail with UnicodeEncodeError or UnicodeDecodeError if the generated documentation contains emojis or other non-ASCII characters.

A possible fix is to use:

open(path, "r", encoding="utf-8", errors="replace")
open(path, "w", encoding="utf-8")