#!/usr/bin/gosh
;A scheme program to parse LeFigaro headlines and generate RSS 0.91 output.
;Copyright (c) 2004 Junichi Uekawa
;Distributed under GPL version 2 or later.

(use rfc.http)
(use gauche.charconv)
(use srfi-11)
(use srfi-13)

(let*
    (
     (html 
      (ces-convert
       (let-values (((code head body) (http-get "www.lefigaro.fr" "/index.html"))) body)
       "iso-8859-1" "utf-8"))
     (my-regexp (string->regexp "<A CLASS=TIT[^ ]* HREF=\"([^\"]*)\">([^<]*)<BR></A>.*?<SPAN CLASS=TXT-UNE>.*?<A HREF=\"[^\"]*\" CLASS=TXT-UNE>([^<]*)</A>"))
     (match #f)
     )

  (print "<?xml version=\"1.0\" encoding=\"UTF-8\"?>
<!DOCTYPE rss PUBLIC \"-//Netscape Communications//DTD RSS 0.91//EN\"
            \"http://my.netscape.com/publish/formats/rss-0.91.dtd\">
<rss version=\"0.91\">
<channel>
<title>Le Figaro</title>
<link>http://www.lefigaro.fr/</link>
<description>RSS feed for Le Figaro</description>
<language>fr</language>
")

  (do ()
      ((eqv? (eqv? html #f) #t)) ;; if false, execute commands.
    (set! match (rxmatch my-regexp html))
    (if match
	(rxmatch-let 
	 match 
	 (#f aref headline text)
	 (print
	  (format "
    <item>
      <title>~s</title>
      <link>http://www.lefigaro.fr~s</link>
      <description>
~s
      </description>
    </item>
"
		  headline aref text))
	 (set! html (match 'after)))
	(set! html #f)))

  (print "
</channel>
</rss>
"))