Hi Marco,

On 2/23/07, Marco Baringer mb@bese.it wrote:

"Bill Clementson" billclem@gmail.com writes:

...

--- cl-indent-orig.el 2007-02-16 13:49:06.000000000 -0800 +++ cl-indent-mods.el 2007-02-16 13:51:39.000000000 -0800 @@ -214,15 +214,17 @@ ((null method) (when (null (cdr path)) ;; (package prefix was stripped off above)

•              (cond ((string-match "\\`def"
  

•                                   function)
  

•              (cond ((and (string-match "\\`def"
  

•                                        function)
  

•                          (not (string-match "\\`default-"
  

•                                             function)))
                      (setq tentative-defun t))
                     ((string-match
  

wouldn't (not (string-match "\[^:]def" function)) be better?

I don't think that works for Zach's specific example. Here is his code (indented with the standard cl-indent):

(defclass field () ((name :initarg :name :accessor name) (default-value :initarg :default-value :accessor default-value)) (:default-initargs :name nil :default-value nil))

Using my modification to cl-indent, the indentation becomes:

(defclass field () ((name :initarg :name :accessor name) (default-value :initarg :default-value :accessor default-value)) (:default-initargs :name nil :default-value nil))

So, the indentation of both the default-value form and the :default-initargs form is adjusted (which is what Zach was after, I think).

I said it was a "yucky patch" because it just fixes the indentation for this particular scenario. Often, indentation of "def..." forms is handled oddly by cl-indent because of the in-built supposition that (unless otherwise defined) anything that begins with "def" is a definition form. A better patch might be to provide a defcustom emacs variable that could be used for any "def..." forms that aren't to be indented as definition forms and to check that variable before indenting code as a definition form.

-- Bill Clementson