And on the contrary, (string-match "\\s-" "\n") => nil That seems backwards to me. More speculatively, I wonder if we shouldn't special-case ?\n and have *both* of those evaluate to nil. The use case is that in a line- oriented parser for shell scripts (specifically version.sh), I'd like to write (looking-at #r"\(\w+\)=\(\S-\(.*\S-\)?\)?\s-*$") The problem is that if the value assigned is the empty string (not whitespace), the first \S- sucks up the newline, and then the rest of the next line! _______________________________________________ XEmacs-Beta mailing list XEmacs-Beta(a)xemacs.org http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta