Re: [Phpwiki-talk] Site Map code

[Arno H. <aho...@xm...>]

Malcom,

> Just for kicks, I added a "Site Map" function to the Wiki, which takes =
a
> given starting page and does a breadth-first-search of the entire Wiki
> using that page as the root. The result is remarkably readable (for my
> Wiki, at least). Have a look at:
>
> http://ccls1.cse.unsw.edu.au/~malcolmr/nomicwiki/index.php?map=3DNomicW=
iki

Looks nice :o)  Post the code please.

> 1) I want to render it more like a standard Wiki page. In particular, I
>    want to make all the links active. Is there an appropriate function
>    to output the HTML code for a link?

In stdlib.php there are two functions called LinkExistingWikiWord() and=20
LinkUnknownWikiWord(). I guess this is what you're looking for.
Also, you can check the existence of a page with IsWikiPage().

Problem with sitemaps like these is, that they are quite expensive to=20
compute. If your wiki is small and has view visitors it doesn't matter mu=
ch.
But my public wiki at senseis.xmp.net gets mirrored via wget at least twi=
ce=20
a week. And those !&@(? mirror every link they get, even the edit page an=
d
similar stuff. They also would mirror the sitemap for every page. So=20
basically each mirroring would be a small DoS attack on your wiki. That's=
=20
why Ward's c2.com wiki uses the referer header to get rid of robots.

> 2) I am using ExtractWikiPageLinks to get a list of outgoing links from
>    a page, but this also returns things I don't want: outside links,
>    disabled links, and text of the form '[[...]'. What is the best way
>    to filter these out?

I think there's a bug in 1.1.9 which causes this behaviour. Try the lates=
t=20
nightly build instead. Or replace it with the following function

   function ExtractWikiPageLinks($content)
   {
      global $WikiNameRegexp;

      $wikilinks =3D array();
      $numlines =3D count($content);
      for($l =3D 0; $l < $numlines; $l++)
      {
         // remove escaped '['
         $line =3D str_replace('[[', ' ', $content[$l]);

         // bracket links (only type wiki-* is of interest)
         $numBracketLinks =3D=20
preg_match_all("/\[\s*([^\]|]+\|)?\s*(.+?)\s*\]/", $line, $brktlinks);
         for ($i =3D 0; $i < $numBracketLinks; $i++) {
            $link =3D ParseAndLink($brktlinks[0][$i]);
            if (preg_match("#^wiki#", $link['type']))
               $wikilinks[$brktlinks[2][$i]] =3D 1;

            $brktlink =3D preg_quote($brktlinks[0][$i]);
            $line =3D preg_replace("|$brktlink|", '', $line);
         }

         // BumpyText old-style wiki links
         if (preg_match_all("/!?$WikiNameRegexp/", $line, $link)) {
            for ($i =3D 0; isset($link[0][$i]); $i++) {
               if($link[0][$i][0] <> '!')
                  $wikilinks[$link[0][$i]] =3D 1;
            }
         }
      }
      return $wikilinks;
   }     =20


This should return only wiki internal links.

/Arno