Previous to Ben’s unicode-internal work, if split-char was available on any emacs, it never gave nil for a valid character. And so a certain amount of code was written with that assumption. My code has been particularly impacted by this change; see the error provoked with C-u C-x = for any remotely exotic Unicode character for an example. To clarify what I mean, this code: (mapcar #'split-char (list ?\U000130bA (max-char) (- (max-char) #xff))) Gives (nil nil nil) on current unicode-internal XEmacs, and the following on a current GNU Emacs: ((unicode 1 48 186) (eight-bit 255) (emacs 63 255 0)) My working plan at the moment is to emulate GNU’s behaviour in those situations where we currently give nil. Thoughts? Details of the GNU build: GNU Emacs 27.0.50 (build 2, i386-apple-darwin10.8.0, NS appkit-1038.36 Version 10.6.8 (Build 10K549)) of 2018-04-29 built on bonbon Repository revision: bf12cc6c7830937b4c1a3df46294e74bea749a5f -- ‘As I sat looking up at the Guinness ad, I could never figure out / How your man stayed up on the surfboard after forty pints of stout’ (C. Moore)