GNU bug report logs - #15219
Emacs Lisp mode and Lisp mode font-locking

Previous Next

To reply to this bug, email your comments to 15219 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Bozhidar Batsov <bozhidar.batsov <at> gmail.com>
To: bug-gnu-emacs <at> gnu.org
Subject: Emacs Lisp mode and Lisp mode font-locking
Date: Fri, 30 Aug 2013 12:32:05 +0300

[Message part 1 (text/plain, inline)]

I've noticed something odd about the font-locking in Emacs Lisp mode and Lisp mode - keyword args are highlighted using the font-lock-builtin-face and constructs like &optional are highlighted using font-lock-type-face. I guess this was was done way back and hasn't been updated in a while, but I think it might a good idea to revise this. Pretty sure those font faces are intended for different usage. I think it would be great if all Emacs programming modes used the built-in font-lock faces consistently, so that the meaning of certain faces doesn't change from mode to mode. I guess that the two modes might also start using the font-lock-built-in face to highlight their core functions (like car, cdr, mapcar, mapc, etc) - as Clojure mode does. Personally I feel that uses of the keyword face for things that are not special forms (like the when macro) should be replaced with uses of the built-in face.
I've posted this to devel a while back (http://lists.gnu.org/archive/html/emacs-devel/2013-07/msg00592.html) and Michael Heerdegen suggested reporting it as a bug as well.


-- 
Cheers,
Bozhidar

[Message part 2 (text/html, inline)]

Message #8 received at 15219 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Monnier <monnier <at> iro.umontreal.ca>
To: Bozhidar Batsov <bozhidar.batsov <at> gmail.com>
Cc: 15219 <at> debbugs.gnu.org
Subject: Re: bug#15219: Emacs Lisp mode and Lisp mode font-locking
Date: Fri, 30 Aug 2013 08:45:15 -0400

> I've noticed something odd about the font-locking in Emacs Lisp mode and
> Lisp mode - keyword args are highlighted using the font-lock-builtin-face
> and constructs like &optional are highlighted using
> font-lock-type-face. I guess this was was done way back and hasn't been
> updated in a while, but I think it might a good idea to revise this. Pretty
> sure those font faces are intended for different usage. I think it would be
> great if all Emacs programming modes used the built-in font-lock faces
> consistently, so that the meaning of certain faces doesn't change from mode
> to mode. I guess that the two modes might also start using the
> font-lock-built-in face to highlight their core functions (like car, cdr,
> mapcar, mapc, etc) - as Clojure mode does. Personally I feel that uses of
> the keyword face for things that are not special forms (like the when macro)
> should be replaced with uses of the built-in face.

Which things are special forms and which things are macros is
somewhat arbitrary.  The important thing to know about them is that
they're not functions, so their arguments may not be evaluated in the
usual way.
Also macros are used to introduce new syntactic constructs: in normal
programming languages, something like `when' would be highlighted as
a `keyword', not as a `builtin'.

I'm not completely sure what the `builtin' is used for, usually.
What information is it meant to provide?  I mean, do you consider
`mapcar' "builtin" because it's fast?  Because it's implemented in C?
Because it's predefined (no need for any kind of autoloading/require)?
Is `car' builtin because "if you didn't have it, you'd have to invent
it" (IOW can't be implemented in Elisp)?


        Stefan

Message #11 received at 15219 <at> debbugs.gnu.org (full text, mbox):

From: Bozhidar Batsov <bozhidar.batsov <at> gmail.com>
To: Stefan Monnier <monnier <at> iro.umontreal.ca>
Cc: 15219 <at> debbugs.gnu.org
Subject: Re: bug#15219: Emacs Lisp mode and Lisp mode font-locking
Date: Fri, 30 Aug 2013 16:23:37 +0300

[Message part 1 (text/plain, inline)]

On Friday, August 30, 2013 at 3:45 PM, Stefan Monnier wrote:
> > I've noticed something odd about the font-locking in Emacs Lisp mode and
> > Lisp mode - keyword args are highlighted using the font-lock-builtin-face
> > and constructs like &optional are highlighted using
> > font-lock-type-face. I guess this was was done way back and hasn't been
> > updated in a while, but I think it might a good idea to revise this. Pretty
> > sure those font faces are intended for different usage. I think it would be
> > great if all Emacs programming modes used the built-in font-lock faces
> > consistently, so that the meaning of certain faces doesn't change from mode
> > to mode. I guess that the two modes might also start using the
> > font-lock-built-in face to highlight their core functions (like car, cdr,
> > mapcar, mapc, etc) - as Clojure mode does. Personally I feel that uses of
> > the keyword face for things that are not special forms (like the when macro)
> > should be replaced with uses of the built-in face.
> > 
> 
> 
> Which things are special forms and which things are macros is
> somewhat arbitrary. The important thing to know about them is that
> they're not functions, so their arguments may not be evaluated in the
> usual way.
> 
> 

Fair point. I guess it makes sense to some extend to consider some macros to be "keywords". 
> Also macros are used to introduce new syntactic constructs: in normal
> programming languages, something like `when' would be highlighted as
> a `keyword', not as a `builtin'.
> 
> I'm not completely sure what the `builtin' is used for, usually.
> What information is it meant to provide? 
> 
> 

Generally the idea behind fontifying built-ins (AFAIK) is to remind programmers that something is part of the core of the language and it'd be unwise to redefine or shadow it. For instance in Clojure mode all methods from the clojure.core namespace are font-locked with the built-in face.
> I mean, do you consider
> `mapcar' "builtin" because it's fast? Because it's implemented in C?
> Because it's predefined (no need for any kind of autoloading/require)?
> Is `car' builtin because "if you didn't have it, you'd have to invent
> it" (IOW can't be implemented in Elisp)?
> 
Normally built-in functions would be those that came with the standard library of the language and are pretty generic in nature.  I guess in Emacs Lisp that would be functions like mapcar, mapc, vector, car, cdr - the building blocks on which everything else rests upon. I guess there are not that much of those and they haven't changed much in a long while. And as the Emacs manual itself states "font-lock-builtin-face is for the names of built-in functions". 
> 
> 
> Stefan
Anyways, I'd be happy to get just a fix for the keywords and &optional (and related) faces, since those are definitely off.

[Message part 2 (text/html, inline)]

Message #14 received at 15219 <at> debbugs.gnu.org (full text, mbox):

From: Lars Ingebrigtsen <larsi <at> gnus.org>
To: Bozhidar Batsov <bozhidar.batsov <at> gmail.com>
Cc: Stefan Monnier <monnier <at> iro.umontreal.ca>, 15219 <at> debbugs.gnu.org
Subject: Re: bug#15219: Emacs Lisp mode and Lisp mode font-locking
Date: Wed, 30 Oct 2019 14:30:15 +0100

Bozhidar Batsov <bozhidar.batsov <at> gmail.com> writes:

> Anyways, I'd be happy to get just a fix for the keywords and &optional (and
> related) faces, since those are definitely off.

The complaint is that :foobar below uses font-lock-builtin-face and
&optional uses font-lock-type-face in emacs-lisp-mode.

(defun foo (bar &optional zot)
  (list :foobar bar))

This is defined in lisp-cl-font-lock-keywords in lisp-mode.el, and I
agree that separate faces should probably be used for those -- keywords
aren't builtins and &optional isn't a type.

:keywords are constants, I guess, but we use font-lock-constant-face for
catch/throw and the like, and I don't think we have anything that
vaguely makes sense for &optional.

Would adding new faces for these make sense?

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.