EF> The problem came up when a colleague edited a file with emacs, which seems EF> to allow a non-terminated final line, and the shell script interpreting EF> that file ignored that line, which lead to interesting results and was EF> quite hard to spot.
EAMCS: Eighty Modes And Continuing Surprises
(setq-default require-final-newline t)
in .emacs to the rescue, IIRC. (I eventually went back to straight vi(1).)
Martin
