I have checked some changes to semantic-complete.el into CVS which
will hopefully solve this problem. Consider the new implementation to
have some experimental usability enhancements. ;)
Now, for this case, if you keep pressing TAB or SPC, it will focus
(highlight) the exact matches. Once you make it through the cycle, it
will expand the list of completions to all the other items with
As before, if you just type stuff in, it will start completing past
For the "not unique" problem, the focus feature will now kick in as
it should have before, and allow you to select from multiple items
with the same name.
Please give it a try and let me know how it goes.
>>> Stephen Berman <Stephen.Berman@...> seems to think that:
>On Tue, 27 Mar 2007 16:36:35 +0200 Stephen Berman <Stephen.Berman@...> wrote:
>> Do this:
>> 1. emacs -Q
>> 2. M-: (load-file "/<path to>/cedet/common/cedet.el")
>> 3. M-: (semantic-load-enable-code-helpers)
>> 4. C-x C-f /<path to>/emacs/lisp/calendar/calendar.el
>> 5. C-c , j and at the prompt type `calendar' and the RET. You get the
>> message "[Not Unique]", although `calendar' occurs as a defun in this
>> file (and if found e.g. by imenu). If you type TAB you get completion
>> candidates for "calendar-" but cannot just complete to `calendar'.
>> I have found this problem only with this word in this file. In the
>> same file, if instead `calendar' you type `diary' at the prompt, you
>> correctly get "[Complete]", as well as completion candidates for
>> "diary-" with TAB.
>I have done some testing and conclude that the above problem is due to
>calendar.el containing both (defun calendar ...) and (provide
>'calendar). In general, if file X.el contains either more than one
>(def* X), where def* can be defun, defvar, defcustom, defface, and
>perhaps others -- but defgroup and defalias do not count --, or one
>(def* X) and a (provide 'X), then semantic-complete-jump-local will
>signal "[Not Unique]" and start cycling through all definitions that
>begin with X, i.e., also all X-* definitions. I guess this isn't
>really a bug after all, but it is a great inconvenience in the case of
>libraries like calendar, which contain dozens of calendar-*
>definitions. It would be nice if semantic-complete-jump-local could
>in such cases cycle through just the X definitions, ignoring X-*.
>This SF.net email is sponsored by DB2 Express
>Download DB2 Express C - the FREE version of DB2 express and take
>control of your XML. No limits. Just data. Click to get it now.
>Cedet-devel mailing list
Eric Ludlam: zappo@..., eric@...
Home: http://www.ludlam.net Siege: http://www.siege-engine.com
Emacs: http://cedet.sourceforge.net GNU: http://www.gnu.org