The change in question is to ignore casing on the token resolution strategy. While innocuous enough, we implemented it forgetting why we went down the strict-casing route in the past.
On a positive note, we've implemented another request, expanding to (try and) cover all the special Unicode special character ranges. While you still can't directly copy and paste a symbol like ◆, and expect output, you can however, quote the symbol like so "◆", and it will work. This is part of the Unicode fix, and we've made it a cleaner process, going back to it after all this time.
The function json.get now treats non-Json strings as a single entry JSON array, to support backward compatibility, and provide continuity in macro execution.
Lastly, all reported issues since the last release have been addressed on here.