WG: AW: HTML_DocumentDependencies / Net_URL problem

Hi Laurent, thanks again for your answer - yes I understand what you mean ;-) ... and sorry for the next "but" ... If I follow your suggestion, I would have to parse the url myself. Ok, no problem but it would be exactly what Net_URL::__construct() does - except of the one thing that Net_URL always uses the url of the current script to get values for missing parts in incomplete URLs. My current solution is quite easy and works fine, but it's "not nice": I just set $_SERVER["HTTP_HOST"], $_SERVER['PHP_SELF'],.. to the values of the URL I've opened and everything works fine.. So currently I see three possibilities and I really don't know ... 1) Extending Net_URL to accept a second parameter $url = new Net_URL('./test.html', 'http://www.remotehost.tld/folder/'); So I can tell the class what to use as defaults instead of the current script. (Would be about 5 lines in Net_URL::__construct() ;-) 2) Parse the URL myself. (Copy and Paste Net_URL and add the five lines for the default values... or write something around Net_URL that would be very similar to Net_URLs constructor function..) 3) Use my current solution. Imho the only clean way is candidate No. 1. If I'm right: May I reopen the feature request? No. 2 stands in conflict to oop guidelines and No. 3 is - in german I would say "totaler Pfusch", my English dictionary says "botch". If I'm on the wrong path, please push me someone back to the right one :-) Chris -----Ursprüngliche Nachricht----- Von: Laurent Laville [mailto:pear@laurent-laville.org] Gesendet: Mittwoch, 25. Januar 2006 18:48 An: Christoph Lütjen - webEdition Software GmbH Cc: pear-dev@lists.php.net Betreff: Re: AW: HTML_DocumentDependencies / Net_URL problem Christoph Lütjen - webEdition Software GmbH a écrit : > > Exactly that's my problem. "º!äøðsë > P@³¶!ëhttp://host2/example/text.html" contains e.g. > > <a href="./anotherfile.html">text</a> > > This Link refers to http://host2/example/anotherfile.html and > not to > something relative to the script that's currently running. It's relative to > the opened file (using HTTP_Request($uri) and not to the PHP Script that has > opened it. > > > About relative path, i think we could have (security) problem , if we could navigate into remote (host2) directories. So solve a relative "./anotherfile.html" which should give http://host2/example/anotherfile.html will be impossible. (hope i'm not wrong) But you can set and replace values you expect to have. //host1/parser.php <?php require_once 'Net/URL.php'; require_once 'HTTP/Request.php'; $uri = 'http://host2/example/text.html'; $href_doc1 = new Net_URL($uri); var_dump($href_doc1); $host1_path = dirname($href_doc1->path); var_dump($host1_path); $req = &new HTTP_Request($uri); $req->setMethod(HTTP_REQUEST_METHOD_GET); $req->sendRequest(); $response1 = $req->getResponseBody(); $dom = new DOMDocument(); @$dom->loadHTML($response1); $tags = $dom->getElementsByTagName('a'); foreach ($tags as $tag) { $dest = $tag->getAttribute('href'); $href_doc = new Net_URL($dest); var_dump($href_doc); $host2_dest = $href_doc1->protocol .'://'. $href_doc1->host . $host1_path; if (substr($href_doc->url, 0, 1) !== '/') { $host2_dest .= '/'; } $host2_dest .= $href_doc->url; $tag = array( 'host' => $href_doc1->host, 'url' => $host2_dest ); var_dump($tag); // [1] } ?> Sorry for the draft and lack of explains, but i'm pretty sure you can understand what i means. with a //host2/example/text.html doc that have 2 links <a href="/link.html">text 1</a> <a href="./anotherfile.html">text 2</a> In [1] you will have such output : host => host2 url => http://host2/example/link.html then host => host2 url => http://host2/example/./anotherfile.html > Hoping I'm not simply too stupid... ;-) > > There are no stupid question, but answers may be stupid (hope it will be not my case ;-) ) Laurent