I believe that this is a documentation issue, not a code bug.

The CONVERT function with the filename character set should already do the trick:

SET NAMES utf8;

SELECT CONVERT(CONVERT('測試資料' USING filename) USING binary),

CONVERT(_filename '@6e2c@8a66@8cc7@6599@5eab' USING utf8);

It will return the conversion:

@6e2c@8a66@8cc7@6599	測試資料庫

Note: these strings are from the description. The hexadecimal string had 5 encoded 5-byte characters, while the UTF-8 string only has 4 characters corresponding to the 4 first 5-byte characters.

Ok, that covers the SQL level part, but I still think we should have some binary tool, maybe a "myls", to be able to do the conversion on the command line without having to fall back to mysql -e 'SELECT CONVERT ...'

Shouldn't be too hard to create a simple wrapper binary around the strconvert() function?

I've documented the filename charset in https://mariadb.com/kb/en/library/identifier-to-file-name-mapping/

Let's keep this issue open until a tool is created. But it should be a Task, not a Bug.

Is it a simple wrapper command that calls an existing function?
So I'll work on this for my first contribution.

example:
terminal$ myls "測試資料"
@6e2c@8a66@8cc7@6599
terminal$ myls -r "@6e2c@8a66@8cc7@6599@5eab"
測試資料庫