Special characters disappear if CJK characters follow up.
Using Korean IME (or any other symbol input) type unicode symbol such as ☆, §, ※, # (notice that this is different from # or § which are typed directly from keyboards).
Type full-width character (CJK in this case)
Unicode character previously typed disappears
Expected result
A unicode symbol should remain as is.
Actual result
A unicode symbol gets deleted due to text sanitizing or login with character width calculation.(my guess)
CJK uses IME and using general keyup events will mess up various functionality that depends on it. For example auto-complete upon ‘keyup’ will not work with CJK.
§ in this case but the problem is more general I think.
You can see that alphabet does not delete the preceding §, but Korean characters do (as if the text is being replaced)