Currently the default en-US dictionary used by spellchecker is using a legacy quirk to prevent it flagging other non-Latin languages as misspelled. The dictionary file was encoded in ISO-8859-1 and marked as such. Then the spellchecker will skip feed...
