From 7360c219f5623e5689deba8805b49dbba6c3bc2a Mon Sep 17 00:00:00 2001 From: =?utf8?q?Roland=20H=C3=A4der?= Date: Tue, 8 Mar 2011 11:31:10 +0000 Subject: [PATCH] Substracting the count of \r\n in chunked-encoded HTTP messages indicates a broken HTTP server implementation, currently only the AnomicHTTPD ( http://anomic.de or http://yacy.net ) is known to me as broken --- .gitattributes | 1 + inc/config-global.php | 2 +- inc/functions.php | 531 ++---------------------------------- inc/http-functions.php | 595 +++++++++++++++++++++++++++++++++++++++++ 4 files changed, 612 insertions(+), 517 deletions(-) create mode 100644 inc/http-functions.php diff --git a/.gitattributes b/.gitattributes index bbf9f9e094..d967244299 100644 --- a/.gitattributes +++ b/.gitattributes @@ -227,6 +227,7 @@ inc/handler.php -text svneol=unset#text/plain inc/header.php -text svneol=unset#text/plain inc/hourly/.htaccess svneol=native#text/plain inc/hourly/hourly_ svneol=native#text/plain +inc/http-functions.php svneol=native#text/plain inc/img/.htaccess -text svneol=unset#text/plain inc/inc-functions.php svneol=native#text/plain inc/install-functions.php -text svneol=unset#text/plain diff --git a/inc/config-global.php b/inc/config-global.php index afe85d4cec..8fa172f431 100644 --- a/inc/config-global.php +++ b/inc/config-global.php @@ -60,7 +60,7 @@ error_reporting(E_ALL | E_STRICT); $PATH = str_replace("\\", '/', substr(dirname(__FILE__), 0, -3)); // Some very important function includes -foreach (array('config','wrapper','template','module','inc','stats') as $inc) { +foreach (array('config','wrapper','template','module','inc','stats','http') as $inc) { include($PATH . 'inc/' . $inc . '-functions.php'); } // END - foreach diff --git a/inc/functions.php b/inc/functions.php index 79e5d8d5a7..119c23891d 100644 --- a/inc/functions.php +++ b/inc/functions.php @@ -40,24 +40,6 @@ if (!defined('__SECURITY')) { die(); } // END - if -// Sends out all headers required for HTTP/1.1 reply -function sendHttpHeaders () { - // Used later - $now = gmdate('D, d M Y H:i:s') . ' GMT'; - - // Send HTTP header - sendHeader('HTTP/1.1 ' . getHttpStatus()); - - // General headers for no caching - sendHeader('Expires: ' . $now); // RFC2616 - Section 14.21 - sendHeader('Last-Modified: ' . $now); - sendHeader('Cache-Control: no-store, no-cache, must-revalidate, pre-check=0, post-check=0, max-age=0'); // HTTP/1.1 - sendHeader('Pragma: no-cache'); // HTTP/1.0 - sendHeader('Connection: Close'); - sendHeader('Content-Type: ' . getContentType() . '; charset=UTF-8'); - sendHeader('Content-Language: ' . getLanguage()); -} - // Init fatal message array function initFatalMessages () { $GLOBALS['fatal_messages'] = array(); @@ -842,377 +824,6 @@ function extractHostnameFromUrl (&$script) { return $host; } -// Send a GET request -function sendGetRequest ($script, $data = array(), $removeHeader = false) { - // Extract hostname and port from script - $host = extractHostnameFromUrl($script); - - // Add data - $body = http_build_query($data, '', '&'); - - // There should be data, else we don't need to extend $script with $body - if (!empty($body)) { - // Do we have a question-mark in the script? - if (strpos($script, '?') === false) { - // No, so first char must be question mark - $body = '?' . $body; - } else { - // Ok, add & - $body = '&' . $body; - } - - // Add script data - $script .= $body; - - // Remove trailed & to make it more conform - if (substr($script, -1, 1) == '&') { - $script = substr($script, 0, -1); - } // END - if - } // END - if - - // Generate GET request header - $request = 'GET /' . trim($script) . ' HTTP/1.1' . getConfig('HTTP_EOL'); - $request .= 'Host: ' . $host . getConfig('HTTP_EOL'); - $request .= 'Referer: ' . getUrl() . '/admin.php' . getConfig('HTTP_EOL'); - if (isConfigEntrySet('FULL_VERSION')) { - $request .= 'User-Agent: ' . getTitle() . '/' . getFullVersion() . getConfig('HTTP_EOL'); - } else { - $request .= 'User-Agent: ' . getTitle() . '/' . getConfig('VERSION') . getConfig('HTTP_EOL'); - } - $request .= 'Accept: image/png,image/*;q=0.8,text/plain,text/html,*/*;q=0.5' . getConfig('HTTP_EOL'); - $request .= 'Accept-Charset: UTF-8,*' . getConfig('HTTP_EOL'); - $request .= 'Cache-Control: no-store, no-cache, must-revalidate, post-check=0, pre-check=0' . getConfig('HTTP_EOL'); - $request .= 'Connection: close' . getConfig('HTTP_EOL'); - $request .= getConfig('HTTP_EOL'); - - // Send the raw request - $response = sendRawRequest($host, $request); - - // Should we remove header lines? - if ($removeHeader === true) { - // Okay, remove them - $response = removeHttpHeaderFromResponse($response); - } // END - if - - // Return the result to the caller function - return $response; -} - -// Send a POST request -function sendPostRequest ($script, array $postData, $removeHeader = false) { - // Extract host name from script - $host = extractHostnameFromUrl($script); - - // Construct request body - $body = http_build_query($postData, '', '&'); - - // Generate POST request header - $request = 'POST /' . trim($script) . ' HTTP/1.0' . getConfig('HTTP_EOL'); - $request .= 'Host: ' . $host . getConfig('HTTP_EOL'); - $request .= 'Referer: ' . getUrl() . '/admin.php' . getConfig('HTTP_EOL'); - $request .= 'User-Agent: ' . getTitle() . '/' . getFullVersion() . getConfig('HTTP_EOL'); - $request .= 'Accept: text/plain;q=0.8' . getConfig('HTTP_EOL'); - $request .= 'Accept-Charset: UTF-8,*' . getConfig('HTTP_EOL'); - $request .= 'Cache-Control: no-cache' . getConfig('HTTP_EOL'); - $request .= 'Content-Type: application/x-www-form-urlencoded' . getConfig('HTTP_EOL'); - $request .= 'Content-Length: ' . strlen($body) . getConfig('HTTP_EOL'); - $request .= 'Connection: close' . getConfig('HTTP_EOL'); - $request .= getConfig('HTTP_EOL'); - - // Add body - $request .= $body; - - // Send the raw request - $response = sendRawRequest($host, $request); - - // Should we remove header lines? - if ($removeHeader === true) { - // Okay, remove them - $response = removeHttpHeaderFromResponse($response); - } // END - if - - // Return the result to the caller function - return $response; -} - -// Sends a raw request to another host -function sendRawRequest ($host, $request) { - // Init errno and errdesc with 'all fine' values - $errno = '0'; - $errdesc = ''; - - // Default port is 80 - $port = 80; - - // Initialize array - $response = array('', '', ''); - - // Default is not to use proxy - $useProxy = false; - - // Are proxy settins set? - if (isProxyUsed()) { - // Then use it - $useProxy = true; - } // END - if - - // Load include - loadIncludeOnce('inc/classes/resolver.class.php'); - - // Extract port part from host - $portArray = explode(':', $host); - if (count($portArray) == 2) { - // Extract host and port - $host = $portArray[0]; - $port = $portArray[1]; - } elseif (count($portArray) > 2) { - // This should not happen! - debug_report_bug(__FUNCTION__, __LINE__, 'Invalid ' . $host . '. Please report this to the Mailer-Project team.'); - } - - // Get resolver instance - $resolver = new HostnameResolver(); - - // Open connection - //* DEBUG: */ die('SCRIPT=' . $script); - if ($useProxy === true) { - // Resolve hostname into IP address - $ip = $resolver->resolveHostname(compileRawCode(getProxyHost())); - - // Connect to host through proxy connection - $fp = fsockopen($ip, bigintval(getProxyPort()), $errno, $errdesc, 30); - } else { - // Resolve hostname into IP address - $ip = $resolver->resolveHostname($host); - - // Connect to host directly - $fp = fsockopen($ip, $port, $errno, $errdesc, 30); - } - - // Is there a link? - if (!is_resource($fp)) { - // Failed! - logDebugMessage(__FUNCTION__, __LINE__, $errdesc . ' (' . $errno . ')'); - return $response; - } elseif ((!stream_set_blocking($fp, 0)) || (!stream_set_timeout($fp, 1))) { - // Cannot set non-blocking mode or timeout - logDebugMessage(__FUNCTION__, __LINE__, socket_strerror(socket_last_error())); - return $response; - } - - // Do we use proxy? - if ($useProxy === true) { - // Setup proxy tunnel - $response = setupProxyTunnel($host, $port, $fp); - - // If the response is invalid, abort - if ((count($response) == 3) && (empty($response[0])) && (empty($response[1])) && (empty($response[2]))) { - // Invalid response! - logDebugMessage(__FUNCTION__, __LINE__, 'Proxy tunnel not working?'); - return $response; - } // END - if - } // END - if - - // Write request - fwrite($fp, $request); - - // Start counting - $start = microtime(true); - - // Read response - while (!feof($fp)) { - // Get info from stream - $info = stream_get_meta_data($fp); - - // Is it timed out? 15 seconds is a really patient... - if (($info['timed_out'] == true) || (microtime(true) - $start) > 15) { - // Timeout - logDebugMessage(__FUNCTION__, __LINE__, 'Timed out to get data from host ' . $host); - - // Abort here - break; - } // END - if - - // Get line from stream - $line = fgets($fp, 128); - - // Ignore empty lines because of non-blocking mode - if (empty($line)) { - // uslepp a little to avoid 100% CPU load - usleep(10); - - // Skip this - continue; - } // END - if - - // Add it to response - $response[] = $line; - } // END - while - - // Close socket - fclose($fp); - - // Time request if debug-mode is enabled - if (isDebugModeEnabled()) { - // Add debug message... - logDebugMessage(__FUNCTION__, __LINE__, 'Request took ' . (microtime(true) - $start) . ' seconds and returned ' . count($response) . ' line(s).'); - } // END - if - - // Skip first empty lines - $resp = $response; - foreach ($resp as $idx => $line) { - // Trim space away - $line = trim($line); - - // Is this line empty? - if (empty($line)) { - // Then remove it - array_shift($response); - } else { - // Abort on first non-empty line - break; - } - } // END - foreach - - //* DEBUG: */ debugOutput('Request:
'.print_r($request, true).'
'); - //* DEBUG: */ debugOutput('Response:
'.print_r($response, true).'
'); - - // Proxy agent found or something went wrong? - if (!isset($response[0])) { - // No response, maybe timeout - $response = array('', '', ''); - logDebugMessage(__FUNCTION__, __LINE__, 'Invalid empty response array, maybe timed out?'); - } elseif ((substr(strtolower($response[0]), 0, 11) == 'proxy-agent') && ($useProxy === true)) { - // Proxy header detected, so remove two lines - array_shift($response); - array_shift($response); - } // END - if - - // Was the request successfull? - if ((!isInStringIgnoreCase('200 OK', $response[0])) || (empty($response[0]))) { - // Not found / access forbidden - logDebugMessage(__FUNCTION__, __LINE__, 'Unexpected status code ' . $response[0] . ' detected. "200 OK" was expected.'); - $response = array('', '', ''); - } else { - // Check array for chuncked encoding - $response = unchunkHttpResponse($response); - } // END - if - - // Return response - return $response; -} - -// Sets up a proxy tunnel for given hostname and through resource -function setupProxyTunnel ($host, $port, $resource) { - // Initialize array - $response = array('', '', ''); - - // Generate CONNECT request header - $proxyTunnel = 'CONNECT ' . $host . ':' . $port . ' HTTP/1.0' . getConfig('HTTP_EOL'); - $proxyTunnel .= 'Host: ' . $host . getConfig('HTTP_EOL'); - - // Use login data to proxy? (username at least!) - if (getProxyUsername() != '') { - // Add it as well - $encodedAuth = base64_encode(compileRawCode(getProxyUsername()) . ':' . compileRawCode(getProxyPassword())); - $proxyTunnel .= 'Proxy-Authorization: Basic ' . $encodedAuth . getConfig('HTTP_EOL'); - } // END - if - - // Add last new-line - $proxyTunnel .= getConfig('HTTP_EOL'); - //* DEBUG: */ debugOutput('proxyTunnel=
' . $proxyTunnel.'
'); - - // Write request - fwrite($fp, $proxyTunnel); - - // Got response? - if (feof($fp)) { - // No response received - return $response; - } // END - if - - // Read the first line - $resp = trim(fgets($fp, 10240)); - $respArray = explode(' ', $resp); - if ((strtolower($respArray[0]) !== 'http/1.0') || ($respArray[1] != '200')) { - // Invalid response! - return $response; - } // END - if - - // All fine! - return $respArray; -} - -// Check array for chuncked encoding -function unchunkHttpResponse (array $response) { - // Default is not chunked - $isChunked = false; - - // Check if we have chunks - foreach ($response as $line) { - // Make lower-case and trim it - $line = trim(strtolower($line)); - - // Entry found? - if ((strpos($line, 'transfer-encoding') !== false) && (strpos($line, 'chunked') !== false)) { - // Found! - $isChunked = true; - break; - } // END - if - } // END - foreach - - // Is it chunked? - if ($isChunked === true) { - // Good, we still have the HTTP headers in there, so we need to get rid - // of them temporarly - //* DEBUG: */ die('
'.htmlentities(print_r(removeHttpHeaderFromResponse($response), true)).'
'); - $tempResponse = http_chunked_decode(implode('', removeHttpHeaderFromResponse($response))); - - // We got a string back from http_chunked_decode(), so we need to convert it back to an array - //* DEBUG: */ die('tempResponse['.strlen($tempResponse).']=
'.replaceReturnNewLine(htmlentities($tempResponse)).'
'); - - // Re-add the headers - $response = merge_array($GLOBALS['http_headers'], stringToArray("\n", $tempResponse)); - } // END - if - - // Return the unchunked array - return $response; -} - -// Removes HTTP header lines from a response array (e.g. output from sendRequest() ) -function removeHttpHeaderFromResponse (array $response) { - // Save headers for later usage - $GLOBALS['http_headers'] = array(); - - // The first array element has to contain HTTP - if ((isset($response[0])) && (substr(strtoupper($response[0]), 0, 5) == 'HTTP/')) { - // Okay, we have headers, now remove them with a second array - $response2 = $response; - foreach ($response as $line) { - // Remove line - array_shift($response2); - - // Add full line to temporary global array - $GLOBALS['http_headers'][] = $line; - - // Trim it for testing - $lineTest = trim($line); - - // Is this line empty? - if (empty($lineTest)) { - // Then stop here - break; - } // END - if - } // END - foreach - - // Write back the array - $response = $response2; - } // END - if - - // Return the modified response array - return $response; -} - // Taken from www.php.net isInStringIgnoreCase() user comments function isEmailValid ($email) { // Check first part of email address @@ -2663,6 +2274,21 @@ function stringToArray ($delimiter, $string) { return $strArray; } +// Detects the prefix 'mb_' if a multi-byte string is given +function detectMultiBytePrefix ($str) { + // Default is without multi-byte + $mbPrefix = ''; + + // Detect multi-byte (strictly) + if (mb_detect_encoding($str, 'auto', true) !== false) { + // With multi-byte encoded string + $mbPrefix = 'mb_'; + } // END - if + + // Return the prefix + return $mbPrefix; +} + //----------------------------------------------------------------------------- // Automatically re-created functions, all taken from user comments on www.php.net //----------------------------------------------------------------------------- @@ -2675,132 +2301,5 @@ if (!function_exists('html_entity_decode')) { } } // END - if -if (!function_exists('http_build_query')) { - // Taken from documentation on www.php.net, credits to Marco K. (Germany) and some light mods by R.Haeder - function http_build_query($data, $prefix = '', $sep = '', $key = '') { - $ret = array(); - foreach ((array) $data as $k => $v) { - if (is_int($k) && $prefix != null) { - $k = urlencode($prefix . $k); - } // END - if - - if ((!empty($key)) || ($key === 0)) { - $k = $key . '[' . urlencode($k) . ']'; - } // END - if - - if (is_array($v) || is_object($v)) { - array_push($ret, http_build_query($v, '', $sep, $k)); - } else { - array_push($ret, $k.'='.urlencode($v)); - } - } // END - foreach - - if (empty($sep)) { - $sep = ini_get('arg_separator.output'); - } // END - if - - return implode($sep, $ret); - } -} // END - if - -if (!function_exists('http_chunked_decode')) { - /** - * dechunk an http 'transfer-encoding: chunked' message. - * - * @param $chunk The encoded message - * @return $dechunk The decoded message. If $chunk wasn't encoded properly debug_report_bug() is being called - * @author Marques Johansson - * @link http://php.net/manual/en/function.http-chunked-decode.php#89786 - */ - function http_chunked_decode ($chunk) { - // Init some variables - $offset = 0; - $len = mb_strlen($chunk); - $dechunk = ''; - - // Walk through all chunks - while ($offset < $len) { - // Where does the \r\n begin? - $lineEndAt = mb_strpos($chunk, getConfig('HTTP_EOL'), $offset); - - /* DEBUG: * - print 'lineEndAt['.__LINE__.']='.$lineEndAt.'
-offset['.__LINE__.']='.$offset.'
-len='.$len.'
-next[offset]=
'.replaceReturnNewLine(htmlentities(mb_substr($chunk, $offset, 10))).'
'; - /* DEBUG: */ - - // Get next hex-coded chunk length - $chunkLenHex = mb_substr($chunk, $offset, ($lineEndAt - $offset)); - - /* DEBUG: * - print 'chunkLenHex['.__LINE__.']='.replaceReturnNewLine(htmlentities($chunkLenHex)).'
-'; - /* DEBUG: */ - - // Validation if it is hexadecimal - if (!isHexadecimal($chunkLenHex)) { - // Please help debugging this - //* DEBUG: */ die('ABORT:chunkLenHex=
'.replaceReturnNewLine(htmlentities($chunkLenHex)).'
'); - debug_report_bug(__FUNCTION__, __LINE__, 'Value ' . $chunkLenHex . ' is not properly chunk encoded.'); - - // This won't be reached - return $chunk; - } // END - if - - // Position of next chunk is right after \r\n - $offset = $offset + strlen($chunkLenHex) + strlen(getConfig('HTTP_EOL')); - $chunkLen = hexdec(rtrim($chunkLenHex, getConfig('HTTP_EOL'))); - - /* DEBUG: * - print 'chunkLen='.$chunkLen.'
-offset['.__LINE__.']='.$offset.'
'; - /* DEBUG: */ - - // Moved out for debugging - $next = mb_substr($chunk, $offset, $chunkLen); - //* DEBUG: */ print 'next=
'.replaceReturnNewLine(htmlentities($next)).'
'; - - // Count occurrences of \r\n - $count = mb_substr_count($next, getConfig('HTTP_EOL')); - - // Correct it because we need to subtract occurrences of \r\n - $chunkLen = hexdec(rtrim($chunkLenHex, getConfig('HTTP_EOL'))) - ($count * strlen(getConfig('HTTP_EOL'))); - - $dechunk .= mb_substr($chunk, $offset, $chunkLen); - - /* DEBUG: * - print('offset['.__LINE__.']='.$offset.'
-lineEndAt['.__LINE__.']='.$lineEndAt.'
-len='.$len.'
-count='.$count.'
-chunkLen='.$chunkLen.'
-chunkLenHex='.$chunkLenHex.'
-dechunk=
'.replaceReturnNewLine(htmlentities($dechunk)).'
-chunk=
'.replaceReturnNewLine(htmlentities($chunk)).'
'); - /* DEBUG: */ - - // Is $offset + $chunkLen larger than or equal $len? - if (($offset + $chunkLen) >= $len) { - // Then stop processing here - break; - } // END - if - - // Calculate next offset of chunk - $offset = mb_strpos($chunk, getConfig('HTTP_EOL'), $offset + $chunkLen) + 2; - - /* DEBUG: * - print('offset['.__LINE__.']='.$offset.'
-next[100]=
'.replaceReturnNewLine(htmlentities(mb_substr($chunk, $offset, 100))).'
----:---:---:---:---:---:---:---:---
-'); - /* DEBUG: */ - } // END - while - - // Return de-chunked string - return $dechunk; - } -} // END - if - // [EOF] ?> diff --git a/inc/http-functions.php b/inc/http-functions.php new file mode 100644 index 0000000000..b0cbca319d --- /dev/null +++ b/inc/http-functions.php @@ -0,0 +1,595 @@ + 2) { + // This should not happen! + debug_report_bug(__FUNCTION__, __LINE__, 'Invalid ' . $host . '. Please report this to the Mailer-Project team.'); + } + + // Get resolver instance + $resolver = new HostnameResolver(); + + // Open connection + //* DEBUG: */ die('SCRIPT=' . $script); + if ($useProxy === true) { + // Resolve hostname into IP address + $ip = $resolver->resolveHostname(compileRawCode(getProxyHost())); + + // Connect to host through proxy connection + $fp = fsockopen($ip, bigintval(getProxyPort()), $errno, $errdesc, 30); + } else { + // Resolve hostname into IP address + $ip = $resolver->resolveHostname($host); + + // Connect to host directly + $fp = fsockopen($ip, $port, $errno, $errdesc, 30); + } + + // Is there a link? + if (!is_resource($fp)) { + // Failed! + logDebugMessage(__FUNCTION__, __LINE__, $errdesc . ' (' . $errno . ')'); + return $response; + } elseif ((!stream_set_blocking($fp, 0)) || (!stream_set_timeout($fp, 1))) { + // Cannot set non-blocking mode or timeout + logDebugMessage(__FUNCTION__, __LINE__, socket_strerror(socket_last_error())); + return $response; + } + + // Do we use proxy? + if ($useProxy === true) { + // Setup proxy tunnel + $response = setupProxyTunnel($host, $port, $fp); + + // If the response is invalid, abort + if ((count($response) == 3) && (empty($response[0])) && (empty($response[1])) && (empty($response[2]))) { + // Invalid response! + logDebugMessage(__FUNCTION__, __LINE__, 'Proxy tunnel not working?'); + return $response; + } // END - if + } // END - if + + // Write request + fwrite($fp, $request); + + // Start counting + $start = microtime(true); + + // Read response + while (!feof($fp)) { + // Get info from stream + $info = stream_get_meta_data($fp); + + // Is it timed out? 15 seconds is a really patient... + if (($info['timed_out'] == true) || (microtime(true) - $start) > 15) { + // Timeout + logDebugMessage(__FUNCTION__, __LINE__, 'Timed out to get data from host ' . $host); + + // Abort here + break; + } // END - if + + // Get line from stream + $line = fgets($fp, 128); + + // Ignore empty lines because of non-blocking mode + if (empty($line)) { + // uslepp a little to avoid 100% CPU load + usleep(10); + + // Skip this + continue; + } // END - if + + // Check for broken HTTP implementations + if (substr(strtolower($line), 0, 7) == 'server:') { + // Anomic (see http://anomic.de, http://yacy.net) is currently broken + $GLOBALS['is_http_server_broken'] = in_array(trim(substr(strtolower($line), 7)), array('anomichttpd')); + } // END - if + + // Add it to response + $response[] = $line; + } // END - while + + // Close socket + fclose($fp); + + // Time request if debug-mode is enabled + if (isDebugModeEnabled()) { + // Add debug message... + logDebugMessage(__FUNCTION__, __LINE__, 'Request took ' . (microtime(true) - $start) . ' seconds and returned ' . count($response) . ' line(s).'); + } // END - if + + // Skip first empty lines + $resp = $response; + foreach ($resp as $idx => $line) { + // Trim space away + $line = trim($line); + + // Is this line empty? + if (empty($line)) { + // Then remove it + array_shift($response); + } else { + // Abort on first non-empty line + break; + } + } // END - foreach + + //* DEBUG: */ debugOutput('Request:
'.print_r($request, true).'
'); + //* DEBUG: */ debugOutput('Response:
'.print_r($response, true).'
'); + + // Proxy agent found or something went wrong? + if (!isset($response[0])) { + // No response, maybe timeout + $response = array('', '', ''); + logDebugMessage(__FUNCTION__, __LINE__, 'Invalid empty response array, maybe timed out?'); + } elseif ((substr(strtolower($response[0]), 0, 11) == 'proxy-agent') && ($useProxy === true)) { + // Proxy header detected, so remove two lines + array_shift($response); + array_shift($response); + } // END - if + + // Was the request successfull? + if ((!isInStringIgnoreCase('200 OK', $response[0])) || (empty($response[0]))) { + // Not found / access forbidden + logDebugMessage(__FUNCTION__, __LINE__, 'Unexpected status code ' . $response[0] . ' detected. "200 OK" was expected.'); + $response = array('', '', ''); + } else { + // Check array for chuncked encoding + $response = unchunkHttpResponse($response); + } // END - if + + // Return response + return $response; +} + +// Sets up a proxy tunnel for given hostname and through resource +function setupProxyTunnel ($host, $port, $resource) { + // Initialize array + $response = array('', '', ''); + + // Generate CONNECT request header + $proxyTunnel = 'CONNECT ' . $host . ':' . $port . ' HTTP/1.0' . getConfig('HTTP_EOL'); + $proxyTunnel .= 'Host: ' . $host . getConfig('HTTP_EOL'); + + // Use login data to proxy? (username at least!) + if (getProxyUsername() != '') { + // Add it as well + $encodedAuth = base64_encode(compileRawCode(getProxyUsername()) . ':' . compileRawCode(getProxyPassword())); + $proxyTunnel .= 'Proxy-Authorization: Basic ' . $encodedAuth . getConfig('HTTP_EOL'); + } // END - if + + // Add last new-line + $proxyTunnel .= getConfig('HTTP_EOL'); + //* DEBUG: */ debugOutput('proxyTunnel=
' . $proxyTunnel.'
'); + + // Write request + fwrite($fp, $proxyTunnel); + + // Got response? + if (feof($fp)) { + // No response received + return $response; + } // END - if + + // Read the first line + $resp = trim(fgets($fp, 10240)); + $respArray = explode(' ', $resp); + if ((strtolower($respArray[0]) !== 'http/1.0') || ($respArray[1] != '200')) { + // Invalid response! + return $response; + } // END - if + + // All fine! + return $respArray; +} + +// Check array for chuncked encoding +function unchunkHttpResponse (array $response) { + // Default is not chunked + $isChunked = false; + + // Check if we have chunks + foreach ($response as $line) { + // Make lower-case and trim it + $line = trim(strtolower($line)); + + // Entry found? + if ((strpos($line, 'transfer-encoding') !== false) && (strpos($line, 'chunked') !== false)) { + // Found! + $isChunked = true; + break; + } // END - if + } // END - foreach + + // Is it chunked? + if ($isChunked === true) { + // Good, we still have the HTTP headers in there, so we need to get rid + // of them temporarly + //* DEBUG: */ die('
'.htmlentities(print_r(removeHttpHeaderFromResponse($response), true)).'
'); + $tempResponse = http_chunked_decode(implode('', removeHttpHeaderFromResponse($response))); + + // We got a string back from http_chunked_decode(), so we need to convert it back to an array + //* DEBUG: */ die('tempResponse['.strlen($tempResponse).']=
'.replaceReturnNewLine(htmlentities($tempResponse)).'
'); + + // Re-add the headers + $response = merge_array($GLOBALS['http_headers'], stringToArray("\n", $tempResponse)); + } // END - if + + // Return the unchunked array + return $response; +} + +// Removes HTTP header lines from a response array (e.g. output from sendRequest() ) +function removeHttpHeaderFromResponse (array $response) { + // Save headers for later usage + $GLOBALS['http_headers'] = array(); + + // The first array element has to contain HTTP + if ((isset($response[0])) && (substr(strtoupper($response[0]), 0, 5) == 'HTTP/')) { + // Okay, we have headers, now remove them with a second array + $response2 = $response; + foreach ($response as $line) { + // Remove line + array_shift($response2); + + // Add full line to temporary global array + $GLOBALS['http_headers'][] = $line; + + // Trim it for testing + $lineTest = trim($line); + + // Is this line empty? + if (empty($lineTest)) { + // Then stop here + break; + } // END - if + } // END - foreach + + // Write back the array + $response = $response2; + } // END - if + + // Return the modified response array + return $response; +} + +// Returns the flag if a broken HTTP server implementation was detected +function isBrokenHttpServerImplentation () { + // Determine it + $isBroken = ((isset($GLOBALS['is_http_server_broken'])) && ($GLOBALS['is_http_server_broken'] === true)); + + // ... and return it + return $isBroken; +} + +//----------------------------------------------------------------------------- +// Automatically re-created functions, all taken from user comments on www.php.net +//----------------------------------------------------------------------------- + +if (!function_exists('http_build_query')) { + // Taken from documentation on www.php.net, credits to Marco K. (Germany) and some light mods by R.Haeder + function http_build_query($data, $prefix = '', $sep = '', $key = '') { + $ret = array(); + foreach ((array) $data as $k => $v) { + if (is_int($k) && $prefix != null) { + $k = urlencode($prefix . $k); + } // END - if + + if ((!empty($key)) || ($key === 0)) { + $k = $key . '[' . urlencode($k) . ']'; + } // END - if + + if (is_array($v) || is_object($v)) { + array_push($ret, http_build_query($v, '', $sep, $k)); + } else { + array_push($ret, $k.'='.urlencode($v)); + } + } // END - foreach + + if (empty($sep)) { + $sep = ini_get('arg_separator.output'); + } // END - if + + return implode($sep, $ret); + } +} // END - if + +if (!function_exists('http_chunked_decode')) { + /** + * dechunk an HTTP 'transfer-encoding: chunked' message. + * + * @param $chunk The encoded message + * @return $dechunk The decoded message. If $chunk wasn't encoded properly debug_report_bug() is being called + * @author Marques Johansson + * @link http://php.net/manual/en/function.http-chunked-decode.php#89786 + */ + function http_chunked_decode ($chunk) { + // Detect multi-byte encoding + $mbPrefix = detectMultiBytePrefix($chunk); + //* DEBUG: */ print 'mbPrefix=' . $mbPrefix . '
'; + + // Init some variables + $offset = 0; + $len = call_user_func_array($mbPrefix . 'strlen', array(($chunk))); + $dechunk = ''; + + // Walk through all chunks + while ($offset < $len) { + // Where does the \r\n begin? + $lineEndAt = call_user_func_array($mbPrefix . 'strpos', array($chunk, getConfig('HTTP_EOL'), $offset)); + + /* DEBUG: * + print 'lineEndAt['.__LINE__.']='.$lineEndAt.'
+offset['.__LINE__.']='.$offset.'
+len='.$len.'
+next[offset,10]=
'.replaceReturnNewLine(htmlentities(call_user_func_array($mbPrefix . 'substr', array($chunk, $offset, 10)))).'
'; + /* DEBUG: */ + + // Get next hex-coded chunk length + $chunkLenHex = call_user_func_array($mbPrefix . 'substr', array($chunk, $offset, ($lineEndAt - $offset))); + + /* DEBUG: * + print 'chunkLenHex['.__LINE__.']='.replaceReturnNewLine(htmlentities($chunkLenHex)).'
+'; + /* DEBUG: */ + + // Validation if it is hexadecimal + if (!isHexadecimal($chunkLenHex)) { + // Please help debugging this + //* DEBUG: */ die('ABORT:chunkLenHex=
'.replaceReturnNewLine(htmlentities($chunkLenHex)).'
'); + debug_report_bug(__FUNCTION__, __LINE__, 'Value ' . $chunkLenHex . ' is no valid hexa-decimal string.'); + + // This won't be reached + return $chunk; + } // END - if + + // Position of next chunk is right after \r\n + $offset = $offset + strlen($chunkLenHex) + strlen(getConfig('HTTP_EOL')); + $chunkLen = hexdec(rtrim($chunkLenHex, getConfig('HTTP_EOL'))); + + /* DEBUG: * + print 'chunkLen='.$chunkLen.'
+offset['.__LINE__.']='.$offset.'
'; + /* DEBUG: */ + + // Moved out for debugging + $next = call_user_func_array($mbPrefix . 'substr', array($chunk, $offset, $chunkLen)); + //* DEBUG: */ print 'next=
'.replaceReturnNewLine(htmlentities($next)).'
'; + + /* + * Hack for e.g. YaCy HTTPDaemon (Anomic Server), this HTTP server + * is currently (revision 7567) broken and does not include the \r\n + * characters when it does sent "chunked" messages. + */ + $count = 0; + if (isBrokenHttpServerImplentation()) { + // Count occurrences of \r\n + $count = call_user_func_array($mbPrefix . 'substr_count', array($next, getConfig('HTTP_EOL'))); + } // END - if + + // Correct it because we need to subtract occurrences of \r\n + $chunkLen = hexdec(rtrim($chunkLenHex, getConfig('HTTP_EOL'))) - ($count * strlen(getConfig('HTTP_EOL'))); + + // Add next chunk to $dechunk + $dechunk .= call_user_func_array($mbPrefix . 'substr', array($chunk, $offset, $chunkLen)); + + /* DEBUG: * + print('offset['.__LINE__.']='.$offset.'
+lineEndAt['.__LINE__.']='.$lineEndAt.'
+len='.$len.'
+count='.$count.'
+chunkLen='.$chunkLen.'
+chunkLenHex='.$chunkLenHex.'
+dechunk=
'.replaceReturnNewLine(htmlentities($dechunk)).'
+chunk=
'.replaceReturnNewLine(htmlentities($chunk)).'
'); + /* DEBUG: */ + + // Is $offset + $chunkLen larger than or equal $len? + if (($offset + $chunkLen) >= $len) { + // Then stop processing here + break; + } // END - if + + // Calculate next offset of chunk + $offset = call_user_func_array($mbPrefix . 'strpos', array($chunk, getConfig('HTTP_EOL'), $offset + $chunkLen)) + 2; + + /* DEBUG: * + print('offset['.__LINE__.']='.$offset.'
+next[100]=
'.replaceReturnNewLine(htmlentities(call_user_func_array($mbPrefix . 'substr', array($chunk, $offset, 100)))).'
+---:---:---:---:---:---:---:---:---
+'); + /* DEBUG: */ + } // END - while + + // Return de-chunked string + return $dechunk; + } +} // END - if + +// [EOF] +?> -- 2.30.2