"beet ls" not understanding searches with Japanese in the query (Windows)

[Michael Ward]

Hi, love this project, thanks for the work you put into it.

I'm able to import music with Japanese in it fine.  The filenames come out fine even with Japanese, and "beet ls" lists the songs in my Cygwin terminal with the Japanese.  The problem is, when my beet ls query contains Japanese.

$ beet ls -a the voice  <-- searching against English portion (gets a proper hit, and the result shows correct Japanese)

平原綾香 - The Voice

$ beet ls -a 平原綾香 <-- searching against Japanese portion (same artist as above, but no result)

<<no result>>

I have "terminal_encoding: utf8" explicitly set in config.yaml (without this, all Japanese is garbled in the output).  That seems to make *output* good, so I'm guessing there's an encoding *input* issue (given that the issue is only when I provide Japanese as input), but I'm not sure what my options are.  Any ideas?

Windows 10 with Python 3.7.  My "system-level" locale for "non-Unicode programs" is set to English.  In the menu to change this in Control Panel, there is a curious, "Beta: use Unicode UTF-8 for worldwide language support," but Googling about this indicates that it can be problematic (as it's still in beta, I suppose).  Wonder if that would work? (not in a scenario where I can easily reboot at the moment to check...).