Hi, is this discrepency between Emacs and XEmacs date routines a feature? I'm not sure what limits on the elisp time structure is so the result of passing year 1904 to `encode-time' is perhaps undefined. It would be nice if XEmacs `encode-time' behaved the same as in Emacs, of course. (`parse-time-string' and `date-time' are found in recent Gnuses but they aren't needed to repeat the problem.) Emacs 19.34, 20.3, 20.3.9: (parse-time-string "Sun, 23 May 1904 00:19:03 -0500") (3 19 0 23 5 1904 0 nil -18000) (encode-time 3 19 0 23 5 1904 0 nil -18000) (-31593 8391) (date-to-time "Sun, 23 May 1904 00:19:03 -0500") (-31593 8391) (format-time-string "%d-%b" (date-to-time "Sun, 23 May 1904 00:19:03 -0500")) "23-May" XEmacs 20.4: (parse-time-string "Sun, 23 May 1904 00:19:03 -0500") (3 19 0 23 5 1904 0 nil -18000) (encode-time 3 19 0 23 5 1904 0 nil -18000) (33943 8391) ;; Huh? Compare with Emacs (date-to-time "Sun, 23 May 1904 00:19:03 -0500") (33943 8391) (format-time-string "%d-%b" (date-to-time "Sun, 23 May 1904 00:19:03 -0500")) Signaling: (error "Invalid time specification") format-time-string("%d-%b" (33943 8391)) XEmacs 21.2.10: (parse-time-string "Sun, 23 May 1904 00:19:03 -0500") (3 19 0 23 5 1904 0 nil -18000) (encode-time 3 19 0 23 5 1904 0 nil -18000) Signaling: (error "Specified time is not representable") encode-time(3 19 0 23 5 1904 0 nil -18000)