I'd rather believed that XEmacs had user-defined reader macros. Turns out it doesn't, but it does defined two extra reader macros: #+ and #-
miniblog.
Related Posts
Interesting discussion of inefficiencies in elisp bytecode during function calls: https://www.xemacs.org/Architecting-XEmacs/faster-elisp.html
(From the XEmacs site, but entirely applicable to GNU Emacs and Remacs too I think)
TIL XEmacs has more datatypes! For example, a keymap is an opaque type rather than a list:
I've been reading XEmacs source code today. It's amazing how much the C code resembes the GNU version (but with more derogatory comments).