Neil Woods <[EMAIL PROTECTED]> wrote: [...]
> A quick fix would be to change the test as indicated: > > (defconst apt-utils-completing-read-hashtable-p > (and (not apt-utils-xemacs-p) > (or > ;; Next released version after 21.3 will support this > (and > (>= emacs-major-version 21) > (>= emacs-minor-version 5)) ; <----- was 4 > (>= emacs-major-version 22) > ;; As will the current pretest > (string-match "\\..*\\..*\\." emacs-version))) > "Non-nil if `completing-read' supports hash table as input.") > > or simply to take out the major-version test < 22, on the likely > assumption that the next major release of Emacs will be version 22 (and > not 21.5). > > Hope that helps. Thanks Neil, I tried your first solution, and it took care of the problem, but somehow font-locking was gone. I just checked the author's website, and followed up to the package's latest version: https://alioth.debian.org/projects/mph-emacs-pkgs which has the following version of the defconst: (defconst apt-utils-completing-read-hashtable-p ;; I think this is a valid way to check this feature... (condition-case nil (or (all-completions "" (make-hash-table)) t) (error nil)) "Non-nil if `completing-read' supports hash table as input.") This seems to be working normally again. Cheers, -- Sebastian P. Luque _______________________________________________ Help-gnu-emacs mailing list [email protected] http://lists.gnu.org/mailman/listinfo/help-gnu-emacs
