Re: [salve@debian.org: lisp/mail/supercite.el: sc-select-attribution logic is broken (patch included)]

[Glenn Morris <gmorris+emacs@ast.cam.ac.uk>]

Richard Stallman wrote:

> Is there someone here who can evaluate this proposed patch?
> I don't use supercite.

I used to use supercite. I think the complaint is correct, but the
patch

> - --- lisp/mail/supercite.el	2005-04-04 11:19:51.000000000 +0200
> +++ lisp/mail/supercite-dgms.el	2005-05-09 13:30:56.000000000 +0200
> @@ -1174,8 +1174,9 @@
>  	      (setq attribution attrib
>  		    attriblist nil))
>  	     ((listp attrib)
> - -	      (setq attribution (eval attrib)
> - -		    attriblist nil))
> +	      (setq attribution (eval attrib))
> +	      (when attribution
> +		(setq attriblist nil))
>  	     (t (error "%s did not evaluate to a string or list!"
>  		       "sc-attrib-selection-list"))
>  	     )))

is wrong since it leads to an infinite loop when (eval attrib) returns
nil. I suggest the following:

diff -c -c -w -r1.44 supercite.el
*** supercite.el	14 May 2005 11:27:40 -0000	1.44
--- supercite.el	26 May 2005 17:15:39 -0000
***************
*** 1183,1189 ****
  		    attriblist nil))
  	     ((listp attrib)
  	      (setq attribution (eval attrib)
! 		    attriblist nil))
  	     (t (error "%s did not evaluate to a string or list!"
  		       "sc-attrib-selection-list"))
  	     )))
--- 1183,1190 ----
  		    attriblist nil))
  	     ((listp attrib)
  	      (setq attribution (eval attrib)
! 		    attriblist (if (stringp attribution) nil
!                                  (cdr attriblist))))
  	     (t (error "%s did not evaluate to a string or list!"
  		       "sc-attrib-selection-list"))
  	     )))