Re: [Emacs-diffs] fix/bug-2034 21904f5: Support mode line constructs for 'mode-name' in c-mode (bug#2034)

Re: [Emacs-diffs] fix/bug-2034 21904f5: Support mode line constructs for 'mode-name' in c-mode (bug#2034)

[Stefan Monnier]

> +  (unless c-modeline-flags
> +    (setq mode-name (list "" mode-name (list 'c-modeline-display-flags
> +					     'c-modeline-flags))))
> +  (setq c-modeline-flags
> +	(format "/%s%s%s%s%s"
> +		(if c-block-comment-flag "*" "/")
> +		(if c-electric-flag "l" "")
> +		(if (and c-electric-flag c-auto-newline)
> +		    "a" "")
> +		(if c-hungry-delete-key "h" "")
> +		(if (and
> +		     ;; (cc-)subword might not be loaded.
> +		     (boundp 'c-subword-mode)
> +		     (symbol-value 'c-subword-mode))
> +		    ;; FIXME: subword-mode already comes with its
> +		    ;; own lighter!
> +		    "w"
> +		  "")))

At that point, why use `format` here instead of using mode-line
constructs for everything, as in:

    (setq c-mode-line-flags
          '("/"
            (c-block-comment-flag "*" "/")
            ...))

Then c-update-modeline doesn't need to be called repeatedly since the
refresh is done automatically by force-mode-line-update.


        Stefan

Re: [Emacs-diffs] fix/bug-2034 21904f5: Support mode line constructs for 'mode-name' in c-mode (bug#2034)

[Phil Sainty]

On 2018-07-04 09:03, Stefan Monnier wrote:
> At that point, why use `format` here instead of using mode-line
> constructs for everything, as in:
> 
>     (setq c-mode-line-flags
>           '("/"
>             (c-block-comment-flag "*" "/")
>             ...))
> 
> Then c-update-modeline doesn't need to be called repeatedly since the
> refresh is done automatically by force-mode-line-update.

That's true.  It would mean that a bit more work is happening per 
modeline
render, vs only (I think?) calculating the flags when they change; but
assuming that's nothing to worry about, it sounds sensible to me too.


-Phil

Re: [Emacs-diffs] fix/bug-2034 21904f5: Support mode line constructs for 'mode-name' in c-mode (bug#2034)

[Phil Sainty]

Grepping for uses of c-update-modeline I found this related comment in
lisp/progmodes/antlr-mode.el:

(define-derived-mode antlr-mode prog-mode
   ;; FIXME: Since it uses cc-mode, it bumps into c-update-modeline's
   ;; limitation to mode-name being a string.
   ;; '("Antlr." (:eval (cadr (assq antlr-language 
antlr-language-alist))))
   "Antlr"

I'll test the commented variant with my other changes in place, but it's 
my
understanding that the NAME argument to `define-derived-mode' *should* 
be a
string, even if the mode's body then sets `mode-name' explicitly to 
something
else.

It's certainly documented that NAME is a string; and if it's not a 
string,
then IIRC you run into problems like antlr-mode-hook's docstring 
containing
the full mode line construct rather than the formatted name.

So I expect to use the following in the mode body:
(setq mode-name '("Antlr." (:eval (cadr (assq antlr-language 
antlr-language-alist)))))


-Phil