The csv_to_array function breaks on commas *and* spaces, but the Tokenfield-based UI is set up so that tags may contain spaces. Fortunately Tokenfield includes a function to break the field into an array for you, so I just used that function instead. (This patch has not been tested.)pull/96/head