Skip to content

GitLab

H
httprl
Commit d481079e authored by mikeytown2's avatar mikeytown2
Browse files
Options

Initial commit.

parents
Hide whitespace changes
Inline Side-by-side
Showing with 486 additions and 0 deletions
httprl.info 0 → 100755
View file @ d481079e
name = HTTP Parallel Request Library
description = Send http requests out in parallel in a blocking or non-blocking manner.
package = Performance and scalability
core = 6.x
httprl.module 0 → 100755
View file @ d481079e
<?php
/**
* @file
* HTTP Parallel Request Library module.
*
*/
/**
* Perform an HTTP request
*
* @see drupal_http_request()
*
* This is a flexible and powerful HTTP client implementation. Correctly
* handles GET, POST, PUT or any other HTTP requests.
*
* @param $url
* A string containing a fully qualified URI.
* @param array $options
* (optional) An array that can have one or more of the following elements:
* - headers: An array containing request headers to send as name/value pairs.
* - method: A string containing the request method. Defaults to 'GET'.
* - data: A string containing the request body, formatted as
* 'param=value&param=value&...'. Defaults to NULL.
* - max_redirects: An integer representing how many times a redirect
* may be followed. Defaults to 3.
* - timeout: A float representing the maximum number of seconds the function
* call may take. The default is 30 seconds. If a timeout occurs, the error
* code is set to the HTTP_REQUEST_TIMEOUT constant.
* - context: A context resource created with stream_context_create().
* - blocking: set to FALSE to make this not care about the returned data.
* @return bool
* return value from httprl_send_request().
*/
function httprl_request($url, array $options = array()) {
global $base_root;
$result = new stdClass();
// Parse the URL and make sure we can handle the schema.
$uri = @parse_url($url);
if (empty($uri)) {
$result->error = 'unable to parse URL';
$result->code = -1001;
return $result;
}
if (!isset($uri['scheme'])) {
$result->error = 'missing schema';
$result->code = -1002;
return $result;
}
// Merge the default options.
$options += array(
'headers' => array(),
'method' => 'GET',
'data' => NULL,
'max_redirects' => 3,
'timeout' => 10.0,
'context' => NULL,
'blocking' => TRUE,
);
// Using http 1.0 so set the connection to be closed by default.
// Sending referrer as well.
$options['headers'] += array(
'Connection' => 'close',
'Referer' => $base_root . request_uri(),
);
// stream_socket_client() requires timeout to be a float.
$options['timeout'] = (float) $options['timeout'];
switch ($uri['scheme']) {
case 'http':
case 'feed':
$port = isset($uri['port']) ? $uri['port'] : 80;
$socket = 'tcp://' . $uri['host'] . ':' . $port;
// RFC 2616: "non-standard ports MUST, default ports MAY be included".
// We don't add the standard port to prevent from breaking rewrite rules
// checking the host that do not take into account the port number.
if (empty($options['headers']['Host'])) {
$options['headers']['Host'] = $uri['host'];
}
if ($port != 80) {
$options['headers']['Host'] .= ':' . $port;
}
break;
case 'https':
// Note: Only works when PHP is compiled with OpenSSL support.
$port = isset($uri['port']) ? $uri['port'] : 443;
$socket = 'ssl://' . $uri['host'] . ':' . $port;
if (empty($options['headers']['Host'])) {
$options['headers']['Host'] = $uri['host'];
}
if ($port != 443) {
$options['headers']['Host'] .= ':' . $port;
}
break;
default:
$result->error = 'invalid schema ' . $uri['scheme'];
$result->code = -1003;
return $result;
}
if (empty($options['context'])) {
$fp = @stream_socket_client($socket, $errno, $errstr, $options['timeout'], STREAM_CLIENT_ASYNC_CONNECT|STREAM_CLIENT_CONNECT);
}
else {
// Create a stream with context. Allows verification of a SSL certificate.
$fp = @stream_socket_client($socket, $errno, $errstr, $options['timeout'], STREAM_CLIENT_ASYNC_CONNECT|STREAM_CLIENT_CONNECT, $options['context']);
}
// Make sure the socket opened properly.
if (!$fp) {
// When a network error occurs, we use a negative number so it does not
// clash with the HTTP status codes.
$result->code = -$errno;
$result->error = trim($errstr) ? trim($errstr) : t('Error opening socket @socket', array('@socket' => $socket));
return $result;
}
// Non blocking stream.
stream_set_blocking($fp, 0);
// Construct the path to act on.
$path = isset($uri['path']) ? $uri['path'] : '/';
if (isset($uri['query'])) {
$path .= '?' . $uri['query'];
}
// Merge the default headers.
$options['headers'] += array(
'User-Agent' => 'Drupal (+http://drupal.org/)',
);
// Only add Content-Length if we actually have any content or if it is a POST
// or PUT request. Some non-standard servers get confused by Content-Length in
// at least HEAD/GET requests, and Squid always requires Content-Length in
// POST/PUT requests.
$content_length = strlen($options['data']);
if ($content_length > 0 || $options['method'] == 'POST' || $options['method'] == 'PUT') {
$options['headers']['Content-Length'] = $content_length;
}
// If the server URL has a user then attempt to use basic authentication.
if (isset($uri['user'])) {
$options['headers']['Authorization'] = 'Basic ' . base64_encode($uri['user'] . (!empty($uri['pass']) ? ":" . $uri['pass'] : ''));
}
// If the database prefix is being used by SimpleTest to run the tests in a copied
// database then set the user-agent header to the database prefix so that any
// calls to other Drupal pages will run the SimpleTest prefixed database. The
// user-agent is used to ensure that multiple testing sessions running at the
// same time won't interfere with each other as they would if the database
// prefix were stored statically in a file or database variable.
$test_info = &$GLOBALS['drupal_test_info'];
if (!empty($test_info['test_run_id'])) {
$options['headers']['User-Agent'] = drupal_generate_test_ua($test_info['test_run_id']);
}
$request = $options['method'] . ' ' . $path . " HTTP/1.0\r\n";
foreach ($options['headers'] as $name => $value) {
$request .= $name . ': ' . trim($value) . "\r\n";
}
$request .= "\r\n" . $options['data'];
$result->request = $request;
return httprl_send_request($fp, $url, $request, $options);
}
/**
* Perform an HTTP request; does not wait for reply & you never will get it
* back.
*
* @see drupal_http_request()
*
* This is a flexible and powerful HTTP client implementation. Correctly
* handles GET, POST, PUT or any other HTTP requests.
*
* @param $fp
* (optional) A file pointer.
* @param $request
* (optional) A string containing the request headers to send to the server.
* @param $timeout
* (optional) An interger holding the stream timeout value.
* @return bool
* TRUE if function worked as planed.
*/
function httprl_send_request($fp = NULL, $url = '', $request = '', $options = '') {
static $responses = array();
static $streams = array();
static $timeout = 0;
static $counter = 0;
// Store data in a static and exit.
if (!empty($fp)) {
$streams[$counter] = $fp;
$timeout = max($options['timeout'], $timeout);
$result = new stdClass();
$result->url = $url;
$result->request = $request;
$result->status = 'in progress';
$result->data = '';
$result->options = $options;
$result->chunk_size = 1024;
$responses[$counter] = $result;
$counter++;
return TRUE;
}
// Exit if there is nothing to do.
if (empty($responses)) {
return FALSE;
}
timer_start(__FUNCTION__);
// Run the loop as long as we have a stream to write to.
while (!empty($streams)) {
// Set the read and write vars to the streams var.
$read = $write = $streams;
$except = array();
// Do some voodoo and open all streams at once.
$n = stream_select($read, $write, $except, $timeout);
// We have some streams to write to.
if (!empty($n)) {
// Readable sockets either have data for us, or are failed connection attempts.
foreach ($read as $r) {
$id = array_search($r, $streams);
// Do not read from the non blocking sockets.
if (empty($responses[$id]->options['blocking'])) {
continue;
}
// Read socket.
$chunk = fread($r, $responses[$id]->chunk_size);
$responses[$id]->data .= $chunk;
// Get stream data.
$info = stream_get_meta_data($r);
// See if the headers have come back yet.
if (empty($responses[$id]->headers) &&
( strpos($responses[$id]->data, "\r\n\r\n")
|| strpos($responses[$id]->data, "\n\n")
|| strpos($responses[$id]->data, "\r\r")
)
) {
httprl_parse_data($responses[$id]);
if (!empty($responses[$id]->headers)) {
// Stream timed out, close connection.
if (isset($responses[$id]->error) && $responses[$id]->error == 'request timed out') {
fclose($r);
unset($streams[$id]);
$responses[$id]->status = 'Done.';
continue;
}
// Sream was a redirect, close this connection; redirect is being followed.
if (!empty($responses[$id]->redirect_url)) {
fclose($r);
unset($streams[$id]);
$responses[$id]->status = 'Done.';
continue;
}
}
// Now that we have the headers, increase the chunk size.
$responses[$id]->chunk_size = 32768;
}
$alive = !$info['eof'] && !$info['timed_out'] && strlen($chunk);
if (!$alive) {
if ($responses[$id]->status == 'in progress') {
$responses[$id]->status = 'Failed to connect.';
}
else {
$responses[$id]->status = 'Done.';
}
fclose($r);
unset($streams[$id]);
}
else {
$responses[$id]->status = 'Reading data';
}
}
// Write to each stream if it is available.
foreach ($write as $w) {
$id = array_search($w, $streams);
if (isset($streams[$id]) && $responses[$id]->status == 'in progress') {
$len = strlen($responses[$id]->request);
$bytes = fwrite($w, $responses[$id]->request, $len);
if ($bytes >= $len) {
$responses[$id]->status = "Request sent, waiting for response.";
}
else {
$responses[$id]->request = substr($responses[$id]->request, $bytes);
}
if (empty($responses[$id]->options['blocking'])) {
fclose($w);
unset($streams[$id]);
$responses[$id]->status = 'Non-Blocking request sent out. Not waiting for the response.';
}
}
}
}
else {
break;
}
}
// Put redirects into the correct request url.
$redirects = array();
foreach ($responses as $id => $result) {
if (!empty($result->redirect_url)) {
$redirects[$result->url] = array(
'redirect_url' => $result->redirect_url,
'redirect_code' => $result->redirect_code,
'id' => $id,
);
unset($responses[$id]);
}
}
if (!empty($redirects)) {
foreach ($redirects as $orginal_url => $info) {
foreach ($responses as $id => &$result) {
if ($result->url == $info['redirect_url']) {
$result->redirect_url = $info['redirect_url'];
$result->redirect_code = $info['redirect_code'];
$result->url = $orginal_url;
unset($redirects[$orginal_url]);
unset($info);
break;
}
}
}
}
// Set output.
$output = array();
foreach ($responses as $id => &$result) {
$output[$result->url] = $result;
}
// Free memory.
$responses = array();
$streams = array();
$timeout = 0;
$counter = 0;
return $output;
}
/**
* Extract the header and meta data from the http data stream.
*
* @see drupal_http_request()
*
* @param $result
* An object from httprl_send_request.
*/
function httprl_parse_data(&$result) {
// If in non blocking mode, skip.
if (empty($result->options['blocking'])) {
continue;
}
// If the headers are already parsed, skip.
if (!empty($result->headers)) {
continue;
}
// Parse response headers from the response body.
// Be tolerant of malformed HTTP responses that separate header and body with
// \n\n or \r\r instead of \r\n\r\n.
$response = $result->data;
list($response, $result->data) = preg_split("/\r\n\r\n|\n\n|\r\r/", $response, 2);
$response = preg_split("/\r\n|\n|\r/", $response);
// Parse the response status line.
list($protocol, $code, $status_message) = explode(' ', trim(array_shift($response)), 3);
$result->protocol = $protocol;
$result->status_message = $status_message;
$result->headers = array();
// Parse the response headers.
while ($line = trim(array_shift($response))) {
list($name, $value) = explode(':', $line, 2);
$name = strtolower($name);
if (isset($result->headers[$name]) && $name == 'set-cookie') {
// RFC 2109: the Set-Cookie response header comprises the token Set-
// Cookie:, followed by a comma-separated list of one or more cookies.
$result->headers[$name] .= ',' . trim($value);
}
else {
$result->headers[$name] = trim($value);
}
}
$responses = array(
100 => 'Continue',
101 => 'Switching Protocols',
200 => 'OK',
201 => 'Created',
202 => 'Accepted',
203 => 'Non-Authoritative Information',
204 => 'No Content',
205 => 'Reset Content',
206 => 'Partial Content',
300 => 'Multiple Choices',
301 => 'Moved Permanently',
302 => 'Found',
303 => 'See Other',
304 => 'Not Modified',
305 => 'Use Proxy',
307 => 'Temporary Redirect',
400 => 'Bad Request',
401 => 'Unauthorized',
402 => 'Payment Required',
403 => 'Forbidden',
404 => 'Not Found',
405 => 'Method Not Allowed',
406 => 'Not Acceptable',
407 => 'Proxy Authentication Required',
408 => 'Request Time-out',
409 => 'Conflict',
410 => 'Gone',
411 => 'Length Required',
412 => 'Precondition Failed',
413 => 'Request Entity Too Large',
414 => 'Request-URI Too Large',
415 => 'Unsupported Media Type',
416 => 'Requested range not satisfiable',
417 => 'Expectation Failed',
500 => 'Internal Server Error',
501 => 'Not Implemented',
502 => 'Bad Gateway',
503 => 'Service Unavailable',
504 => 'Gateway Time-out',
505 => 'HTTP Version not supported',
);
// RFC 2616 states that all unknown HTTP codes must be treated the same as the
// base code in their class.
if (!isset($responses[$code])) {
$code = floor($code / 100) * 100;
}
$result->code = $code;
switch ($code) {
case 200: // OK
case 304: // Not modified
break;
case 301: // Moved permanently
case 302: // Moved temporarily
case 307: // Moved temporarily
$location = $result->headers['location'];
$result->options['timeout'] -= timer_read(__FUNCTION__) / 1000;
if ($result->options['timeout'] <= 0) {
$result->code = HTTP_REQUEST_TIMEOUT;
$result->error = 'request timed out';
}
elseif ($result->options['max_redirects']) {
// Redirect to the new location.
// TODO: Parse cookies from header and send them in the redirect request.
$result->options['max_redirects']--;
$result->options['headers']['Referer'] = $result->url;
httprl_request($location, $result->options);
$result->redirect_code = $code;
}
$result->redirect_url = $location;
break;
default:
$result->error = $status_message;
}
}
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment