Re: Case problem with menus

[David Kastrup]

Richard Stallman <address@hidden> writes:

>     Cough, cough.  What is the deal exactly with the symbol names, anyway?
>     Where does their case get externally visible?
>
> If you use lookup-key or define-key.
>
>       And why is lowercasing
>     them desirable?  I am afraid that I still don't understand the problem
>     that is presumably addressed by the lowercasing.
>
> For consistency in naming of these symbols.  Please don't think of
> these symbols as an internal implementation detail.  Users do access
> these symbols, and normally they are lower-case.  When people see a
> menu item `Greek' they can assume the symbol is `greek'.
>
> Please try this:
>
> (defsubst easy-menu-intern (s)
>   (if (stringp s)
>       (let ((copy (copy-sequence s))
>           (pos 0)
>           found)
>       (while (setq found (string-match "\\<\\sw" copy pos))
>         (setq pos (match-end 0))
>         (unless (= (upcase (aref copy found))
>                    (downcase (aref copy found)))
>           (aset copy found
>                 (if (= (upcase (aref copy found))
>                        (aref copy found))
>                     (downcase (aref copy found))
>                   (upcase (aref copy found))))))
>        (intern copy))
>     s))

I don't like the implications for menu names that contain words not
starting with uppercase.  Wouldn't it be possible to deal with this in
a manner similarly to generate-new-buffer-name?  By appending <2> and
so on until we have a unique symbol in the structure where the menu
gets added?  Yes, it would not allow one to guess without looking
whether <greek> means "Greek" or "greek" in an ambiguous menu setup,
but to be honest, if <Greek> means "greek" this is not significatly
more helpful.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum