We have several processes where workflow updates records in an Informix database. The driver we use does not accept non-standard ascii characters (anything 126 or above) but our end users are able to enter those characters on Forms. We've created a "token scrubber" loop that searches for the most common non-standard characters and substitutes with common one's. This works great except now we've had a user submit an em dash character in the text (ascii code 226 128 147). When I try to add it to the token calculator substitute I can paste it in and it looks like and tests like the correct one. But as soon as I save the activity it is automatically converted to the standard hyphen/dash (ascii 045). See attached screenshots of before and after.
Am I doing this wrong, is there a bug, or another way to do this?