From 9b307a83ee61a4896227f453603a5991ad5a1f5f Mon Sep 17 00:00:00 2001
From: Andreas Gohr <andi@splitbrain.org>
Date: Wed, 31 Aug 2005 22:55:50 +0200
Subject: [PATCH] HTTP Client added
This patch adds an HTTP client written in PHP. It supports proxy
handling and SSL if PHP was compiled with it.
darcs-hash:20050831205550-7ad00-6dcdff0208d7f18a8ff731febb155d126742c768.gz
---
inc/HTTPClient.php | 393 +++++++++++++++++++++++++++++++++++++++++++++
inc/io.php | 28 ++--
2 files changed, 403 insertions(+), 18 deletions(-)
create mode 100644 inc/HTTPClient.php
diff --git a/inc/HTTPClient.php b/inc/HTTPClient.php
new file mode 100644
index 000000000..8167952c6
--- /dev/null
+++ b/inc/HTTPClient.php
@@ -0,0 +1,393 @@
+<?php
+/**
+ * HTTP Client
+ *
+ * @license GPL 2 (http://www.gnu.org/licenses/gpl.html)
+ * @author Andreas Goetz <cpuidle@gmx.de>
+ */
+
+if(!defined('DOKU_INC')) define('DOKU_INC',realpath(dirname(__FILE__).'/../').'/');
+require_once(DOKU_CONF.'dokuwiki.php');
+
+define('HTTP_NL',"\r\n");
+
+
+/**
+ * Adds DokuWiki specific configs to the HTTP client
+ *
+ * @author Andreas Goetz <cpuidle@gmx.de>
+ */
+class DokuHTTPClient extends HTTPClient {
+
+ /**
+ * Constructor.
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function DokuHTTPClient(){
+ global $conf;
+
+ // call parent constructor
+ $this->HTTPClient();
+
+ // set some values from the config
+ $this->proxy_host = $conf['proxy']['host'];
+ $this->proxy_port = $conf['proxy']['port'];
+ $this->proxy_user = $conf['proxy']['user'];
+ $this->proxy_pass = $conf['proxy']['pass'];
+ $this->proxy_ssl = $conf['proxy']['usessl'];
+ }
+}
+
+/**
+ * This class implements a basic HTTP client
+ *
+ * It supports POST and GET, Proxy usage, basic authentication,
+ * handles cookies and referers. It is based upon the httpclient
+ * function from the VideoDB project.
+ *
+ * @link http://www.splitbrain.org/go/videodb
+ * @author Andreas Goetz <cpuidle@gmx.de>
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+class HTTPClient {
+ //set these if you like
+ var $agent; // User agent
+ var $http; // HTTP version defaults to 1.0
+ var $timeout;
+ var $cookies;
+ var $referer;
+ var $max_redirect;
+ var $max_bodysize; // abort if the response body is bigger than this
+ var $headers;
+ var $debug;
+
+ // don't set these, read on error
+ var $error;
+ var $redirect_count;
+
+ // read these after a successful request
+ var $resp_status;
+ var $resp_body;
+ var $resp_headers;
+
+ // set these to do basic authentication
+ var $user;
+ var $pass;
+
+ // set these if you need to use a proxy
+ var $proxy_host;
+ var $proxy_port;
+ var $proxy_user;
+ var $proxy_pass;
+ var $proxy_ssl; //boolean set to true if your proxy needs SSL
+
+ /**
+ * Constructor.
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function HTTPClient(){
+ $this->agent = 'Mozilla/4.0 (compatible; DokuWiki HTTP Client; '.PHP_OS.')';
+ $this->timeout = 15;
+ $this->cookies = array();
+ $this->referer = '';
+ $this->max_redirect = 3;
+ $this->redirect_count = 0;
+ $this->status = 0;
+ $this->headers = array();
+ $this->http = '1.0';
+ $this->debug = false;
+ $this->max_bodysize = 0;
+ if(extension_loaded('zlib')) $this->headers['Accept-encoding'] = 'gzip';
+ $this->headers['Accept'] = 'text/xml,application/xml,application/xhtml+xml,'.
+ 'text/html,text/plain,image/png,image/jpeg,image/gif,*/*';
+ $this->headers['Accept-Language'] = 'en-us';
+ }
+
+
+ /**
+ * Simple function to do a GET request
+ *
+ * Returns the wanted page or false on an error;
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function get($url){
+ if(!$this->sendRequest($url)) return false;
+ if($this->status != 200) return false;
+ return $this->resp_body;
+ }
+
+ /**
+ * Simple function to do a POST request
+ *
+ * Returns the resulting page or false on an error;
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function post($url,$data){
+ if(!$this->sendRequest($url,$data,'POST')) return false;
+ if($this->status != 200) return false;
+ return $this->resp_body;
+ }
+
+ /**
+ * Do an HTTP request
+ *
+ * @author Andreas Goetz <cpuidle@gmx.de>
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function sendRequest($url,$data=array(),$method='GET'){
+ $this->error = '';
+
+ // parse URL into bits
+ $uri = parse_url($url);
+ $server = $uri['host'];
+ $path = $uri['path'];
+ if(empty($path)) $path = '/';
+ if(!empty($uri['query'])) $path .= '?'.$uri['query'];
+ $port = $uri['port'];
+ if($uri['user']) $this->user = $uri['user'];
+ if($uri['pass']) $this->user = $uri['pass'];
+
+ // proxy setup
+ if($this->proxy_host){
+ $request_url = $url;
+ $server = $config['proxy_host'];
+ $port = $config['proxy_port'];
+ if (empty($port)) $port = 8080;
+ }else{
+ $request_url = $path;
+ $server = $server;
+ if (empty($port)) $port = ($uri['scheme'] == 'https') ? 443 : 80;
+ }
+
+ // add SSL stream prefix if needed - needs SSL support in PHP
+ if($port == 443 || $this->proxy_ssl) $server = 'ssl://'.$server;
+
+ // prepare headers
+ $headers = $this->headers;
+ $headers['Host'] = $uri['host'];
+ $headers['User-Agent'] = $this->agent;
+ $headers['Referer'] = $this->referer;
+ $headers['Connection'] = 'Close';
+ if($method == 'POST'){
+ $post = _postEncode($data);
+ $headers['Content-Type'] = 'application/x-www-form-urlencoded';
+ $headers['Content-Length'] = strlen($post);
+ }
+ if($this->user) {
+ $headers['Authorization'] = 'BASIC '.base64_encode($this->user.':'.$this->pass);
+ }
+ if($this->proxy_user) {
+ $headers['Proxy-Authorization'] = 'BASIC '.base64_encode($this->proxy_user.':'.$this->proxy_pass);
+ }
+
+ // open socket
+ $socket = @fsockopen($server,$port,$errno, $errstr, $this->timeout);
+ if (!$socket){
+ $this->error = "Could not connect to $server:$port\n$errstr ($errno)";
+ return $false;
+ }
+
+ // build request
+ $request = "$method $request_url HTTP/".$this->http.HTTP_NL;
+ $request .= $this->_buildHeaders($headers);
+ $request .= $this->_getCookies();
+ $request .= HTTP_NL;
+ $request .= $post;
+
+ $this->_debug('request',$request);
+
+ // send request
+ fputs($socket, $request);
+
+ // read headers from socket
+ $r_headers = '';
+ do{
+ $r_headers .= fread($socket,1); #FIXME read full lines here?
+ }while(!preg_match('/\r\n\r\n$/',$r_headers));
+
+ //read body (with chunked encoding if needed)
+ $r_body = '';
+ if(preg_match('/transfer\-(en)?coding:\s+chunked\r\n/i',$r_header)){
+ do {
+ unset($chunk_size);
+ do {
+ $byte = fread($socket,1);
+ $chunk_size .= $byte;
+ } while (preg_match('/[a-zA-Z0-9]/',$byte)); // read chunksize including \r
+
+ $byte = fread($socket,1); // readtrailing \n
+ $chunk_size = hexdec($chunk_size);
+ $this_chunk = fread($socket,$chunk_size);
+ $r_body .= $this_chunk;
+ if ($chunk_size) $byte = fread($socket,2); // read trailing \r\n
+
+ if($this->max_bodysize && strlen($r_body) > $this->max_bodysize){
+ $this->error = 'Allowed response size exceeded';
+ return false;
+ }
+ } while ($chunk_size);
+ }else{
+ // read entire socket
+ while (!feof($socket)) {
+ $r_body .= fread($socket,4096);
+ if($this->max_bodysize && strlen($r_body) > $this->max_bodysize){
+ $this->error = 'Allowed response size exceeded';
+ return false;
+ }
+ }
+ }
+
+ // close socket
+ $status = socket_get_status($socket);
+ fclose($socket);
+
+ $this->_debug('response headers',$r_headers);
+
+ // check for timeout
+ if ($status['timed_out']){
+ $this->error = "Connection timed out";
+ return false;
+ }
+
+ // get Status
+ if (!preg_match('/^HTTP\/(\d\.\d)\s*(\d+).*?\n/', $r_headers, $m)) {
+ $this->error = 'Server returned bad answer';
+ return false;
+ }
+ $this->status = $m[2];
+
+ // handle headers and cookies
+ $this->resp_headers = $this->_parseHeaders($r_headers);
+ if(isset($this->resp_headers['set-cookie'])){
+ foreach ($this->resp_headers['set-cookie'] as $c){
+ list($key, $value, $foo) = split('=', $cookie);
+ $this->cookies[$key] = $value;
+ }
+ }
+
+ $this->_debug('Object headers',$this->resp_headers);
+
+ // check server status code to follow redirect
+ if($this->status == 301 || $this->status == 302 ){
+ if (empty($this->resp_headers['location'])){
+ $this->error = 'Redirect but no Location Header found';
+ return false;
+ }elseif($this->redirect_count == $this->max_redirect){
+ $this->error = 'Maximum number of redirects exceeded';
+ return false;
+ }else{
+ $this->redirect_count++;
+ $this->referer = $url;
+ if (!preg_match('/^http/i', $this->resp_headers['location'])){
+ $this->resp_headers['location'] = $uri['scheme'].'://'.$uri['host'].
+ $this->resp_headers['location'];
+ }
+ // perform redirected request, always via GET (required by RFC)
+ return $this->_sendRequest($this->resp_headers['location'],array(),'GET');
+ }
+ }
+
+ // decode gzip if needed
+ if($this->resp_headers['content-encoding'] == 'gzip'){
+ $this->resp_body = gzinflate(substr($r_body, 10));
+ }else{
+ $this->resp_body = $r_body;
+ }
+
+ $this->_debug('response body',$this->resp_body);
+ $this->redirect_count = 0;
+ return true;
+ }
+
+ /**
+ * print debug info
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function _debug($info,$var){
+ if(!$this->debug) return;
+ print '<b>'.$info.'</b><br />';
+ ob_start();
+ print_r($var);
+ $content = htmlspecialchars(ob_get_contents());
+ ob_end_clean();
+ print '<pre>'.$content.'</pre>';
+ }
+
+ /**
+ * convert given header string to Header array
+ *
+ * All Keys are lowercased.
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function _parseHeaders($string){
+ $headers = array();
+ $lines = explode("\n",$string);
+ foreach($lines as $line){
+ list($key,$val) = explode(':',$line,2);
+ $key = strtolower(trim($key));
+ $val = trim($val);
+ if(empty($val)) continue;
+ if(isset($headers[$key])){
+ if(is_array($headers[$key])){
+ $headers[$key][] = $val;
+ }else{
+ $headers[$key] = array($headers[$key],$val);
+ }
+ }else{
+ $headers[$key] = $val;
+ }
+ }
+ return $headers;
+ }
+
+ /**
+ * convert given header array to header string
+ *
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function _buildHeaders($headers){
+ $string = '';
+ foreach($headers as $key => $value){
+ if(empty($value)) continue;
+ $string .= $key.': '.$value.HTTP_NL;
+ }
+ return $string;
+ }
+
+ /**
+ * get cookies as http header string
+ *
+ * @author Andreas Goetz <cpuidle@gmx.de>
+ */
+ function _getCookies(){
+ foreach ($this->cookies as $key => $val){
+ if ($headers) $headers .= '; ';
+ $headers .= $key.'='.$val;
+ }
+
+ if ($headers) $headers = "Cookie: $headers".HTTP_NL;
+ return $headers;
+ }
+
+ /**
+ * Encode data for posting
+ *
+ * @todo handle mixed encoding for file upoads
+ * @author Andreas Gohr <andi@splitbrain.org>
+ */
+ function _postEncode($data){
+ foreach($params as $key => $val){
+ if($url) $url .= '&';
+ $url .= $key.'='.urlencode($val);
+ }
+ return $url;
+ }
+}
+
+//Setup VIM: ex: et ts=4 enc=utf-8 :
diff --git a/inc/io.php b/inc/io.php
index 5cf50af8f..c25bd8975 100644
--- a/inc/io.php
+++ b/inc/io.php
@@ -8,6 +8,7 @@
if(!defined('DOKU_INC')) define('DOKU_INC',realpath(dirname(__FILE__).'/../').'/');
require_once(DOKU_INC.'inc/common.php');
+ require_once(DOKU_INC.'inc/HTTPClient.php');
/**
* Removes empty directories
@@ -261,30 +262,21 @@ function io_mkdir_ftp($dir){
* downloads a file from the net and saves it to the given location
*
* @author Andreas Gohr <andi@splitbrain.org>
- * @todo Add size limit
*/
function io_download($url,$file){
- $fp = @fopen($url,"rb");
- if(!$fp) return false;
+ $http = new DokuHTTPClient();
+ $http->max_bodysize = 2*1024*1024; //max. 2MB
+ $http->timeout = 25; //max. 25 sec
- $kb = 0;
- $now = time();
+ $data = $http->get($url);
+ if(!$data) return false;
- while(!feof($fp)){
- if($kb++ > 2048 || (time() - $now) > 45){
- //abort on 2 MB and timeout on 45 sec
- return false;
- }
- $cont.= fread($fp,1024);
- }
+ $fp = @fopen($file,"w");
+ if(!$fp) return false;
+ fwrite($fp,$data);
fclose($fp);
-
- $fp2 = @fopen($file,"w");
- if(!$fp2) return false;
- fwrite($fp2,$cont);
- fclose($fp2);
return true;
-}
+}
/**
* Runs an external command and returns it's output as string
--
GitLab