UsingLindentWithEmacs < Training

---++ Using external code beautifiers with Emacs

The following Lisp code is a convenient wrapper to the =Lindent=
script provided with the Linux kernel.  For it to work correctly,
the =Lindent= script has to be somewhere on the path.  Otherwise
modify the lindent-command variable to use an absolute pathname
for =Lindent=.

<verbatim>
(defvar lindent-command "Lindent"
  "Command to call to reformat C source code using Linux CodingStyle.")

(defun dzu-lindent-region (start end replace)
  "Call Lindent on region.  Specifying a prefix arg replaces the region."
  (interactive (list (region-beginning) (region-end)
           current-prefix-arg))

  (unless c-indentation-style
    (error "Not editing a C file"))
  ;; Lindent presumes linux style so adjust if neccessary                                              
  (unless (eq c-indentation-style 'linux)
    (c-set-style "linux"))
  (shell-command-on-region start end lindent-command nil replace)
  ;; When we beautify only partially at an indent level we have to add that                            
  ;; manually afterwards.  Fortunately cc-mode can take care of that.                                  
  (when replace
    (indent-region start end nil)))
</verbatim>

You could for example bind this to the =f8= key:

<verbatim>
(global-set-key [f8] 'dzu-lindent-region)
</verbatim>

To use it, simply mark the piece of code in question (between mark and
point) and call it (e.g. press =f8= when you have bound it to that
key).  Without a prefix argument, the beautified code will be
displayed in the echo area so you can visually compare the original
with the resulting code.

Providing a prefix argument with =C-u= before pressing =f8= replaces
the original code with the output of =Lindent=.  Note that it is *ok*
only to indent chunks from a source file.  In this case the
indentation from before the region is used for the result - regardless
if it was correct or not.

-- Main.DetlevZundel - 11 Sep 2007