modules/boost/boost.module

<?php
// $Id: boost.module,v 1.3 2006/11/27 16:34:49 arto Exp $

/**
 * @file
 * Provides static page caching for Drupal.
 */

//////////////////////////////////////////////////////////////////////////////
// BOOST SETTINGS

define('BOOST_PATH',                 dirname(__FILE__));
define('BOOST_FRONTPAGE',            drupal_get_normal_path(variable_get('site_frontpage', 'node')));

define('BOOST_ENABLED',              variable_get('boost', CACHE_DISABLED));
define('BOOST_FILE_PATH',            variable_get('boost_file_path', 'cache'));
define('BOOST_FILE_EXTENSION',       variable_get('boost_file_extension', '.html'));
define('BOOST_CACHEABILITY_OPTION',  variable_get('boost_cacheability_option', 0));
define('BOOST_CACHEABILITY_PAGES',   variable_get('boost_cacheability_pages', ''));
define('BOOST_FETCH_METHOD',         variable_get('boost_fetch_method', 'php'));
define('BOOST_PRE_PROCESS_FUNCTION', variable_get('boost_pre_process_function', ''));
define('BOOST_POST_UPDATE_COMMAND',  variable_get('boost_post_update_command', ''));
define('BOOST_CRON_LIMIT',           variable_get('boost_cron_limit', 100));

// This cookie is set for all logged-in users, so that they can be excluded
// from caching (or, in the future, get a user-specific cached page):
define('BOOST_COOKIE',               variable_get('boost_cookie', 'DRUPAL_UID'));

// This line is appended to the generated static files; it is very useful
// for troubleshooting (e.g. determining whether one got the dynamic or
// static version):
define('BOOST_BANNER',               variable_get('boost_banner', "<!-- Page cached by Boost at %date -->\n"));

// This is needed since the $user object is already destructed in _boost_ob_handler():
define('BOOST_USER_ID',              $GLOBALS['user']->uid);

//////////////////////////////////////////////////////////////////////////////
// BOOST INCLUDES

require_once BOOST_PATH . '/boost.helpers.inc';
require_once BOOST_PATH . '/boost.api.inc';

//////////////////////////////////////////////////////////////////////////////
// DRUPAL API HOOKS

/**
 * Implementation of hook_help(). Provides online user help.
 */
function boost_help($section) {
  switch ($section) {
    case 'admin/modules#name':
      return t('boost');
    case 'admin/modules#description':
      return t('Provides a performance and scalability boost through caching Drupal pages as static HTML files.');
    case 'admin/help#boost':
      $file = drupal_get_path('module', 'boost') . '/README.txt';
      if (file_exists($file))
        return '<pre>' . implode("\n", array_slice(explode("\n", @file_get_contents($file)), 2)) . '</pre>';
      break;
    case 'admin/settings/boost':
      return '<p>' . '</p>'; // TODO: add help text.
  }
}

/**
 * Implementation of hook_perm(). Defines user permissions.
 */
function boost_perm() {
  return array('administer cache');
}

/**
 * Implementation of hook_menu(). Defines menu items and page callbacks.
 */
function boost_menu($may_cache) {
  $access = user_access('administer cache');
  $items = array();
  if ($may_cache) {
    // TODO: define menu actions for cache administration.
  }
  return $items;
}

/**
 * Implementation of hook_init(). Performs page setup tasks.
 */
function boost_init() {
  // Stop right here unless we're being called for an ordinary page request
  if (strpos($_SERVER['PHP_SELF'], 'index.php') === FALSE)
    return;

  // TODO: check interaction with other modules that use ob_start(); this
  // may have to be moved to an earlier stage of the page request.
  if (!variable_get('cache', CACHE_DISABLED) && BOOST_ENABLED) {
    // We only support GET requests by anonymous visitors:
    global $user;
    if (empty($user->uid) && $_SERVER['REQUEST_METHOD'] == 'GET') {
      // Make sure no query string (in addition to ?q=) was set, and that
      // the page is cacheable according to our current configuration:
      if (count($_GET) == 1 && boost_is_cacheable($_GET['q']))
        ob_start('_boost_ob_handler');
    }
  }

  // Executed when saving Drupal's settings:
  if (!empty($_POST['edit']) && $_GET['q'] == 'admin/settings') {
    // Forcibly disable Drupal's built-in SQL caching to prevent any conflicts of interest:
    variable_set('cache', CACHE_DISABLED);

    // TODO: handle 'offline' site maintenance settings.

    $old = variable_get('boost', '');
    if (!empty($_POST['edit']['boost'])) {
      // Ensure the cache directory exists or can be created
      file_check_directory($_POST['edit']['boost_file_path'], FILE_CREATE_DIRECTORY, 'boost_file_path');
    }
    else if (!empty($old)) { // the cache was previously enabled
      if (boost_cache_expire_all())
        drupal_set_message('Static cache files deleted.');
    }
  }
}

/**
 * Implementation of hook_exit(). Performs cleanup tasks.
 *
 * For POST requests by anonymous visitors, this adds a dummy query string
 * to any URL being redirected to using drupal_goto().
 *
 * This is pretty much a hack that assumes a bit too much familiarity with
 * what happens under the hood of the Drupal core function drupal_goto().
 *
 * It's necessary, though, in order for any session messages set on form
 * submission to actually show up on the next page if that page has been
 * cached by Boost.
 */
function boost_exit($destination = NULL) {
  // Check that hook_exit() was invoked by drupal_goto() for a POST request:
  if (!empty($destination) && $_SERVER['REQUEST_METHOD'] == 'POST') {

    // Check that we're dealing with an anonymous visitor. and that some
    // session messages have actually been set during this page request:
    global $user;
    if (empty($user->uid) && ($messages = drupal_set_message())) {

      // Check that the page we're redirecting to has been cached by Boost
      // and really necessitates special handling:
      extract(parse_url($destination));
      $path = ($path == base_path() ? '' : substr($path, strlen(base_path())));
      if (boost_is_cached($path) && empty($query)) {
        // FIXME: call any remaining exit hooks since we're about to terminate.

        // Add a query string to ensure we don't serve a static copy of
        // the page we're redirecting to, which would prevent the session
        // messages from showing up:
        $destination = url($path, 't=' . time(), $fragment, TRUE);

        // Do what drupal_goto() would do if we were to return to it:
        exit(header('Location: ' . $destination));
      }
    }
  }
}

/**
 * Implementation of hook_form_alter(). Performs alterations before a form
 * is rendered.
 */
function boost_form_alter($form_id, &$form) {
  // Alter Drupal's settings form by hiding the default cache enabled/disabled control (which will now always default to CACHE_DISABLED), and add our own control instead.
  if ($form_id == 'system_settings_form') {
    require_once BOOST_PATH . '/boost.admin.inc';
    $form['cache'] = boost_system_settings_form($form['cache']);
  }
}

/**
 * Implementation of hook_cron(). Performs periodic actions.
 */
function boost_cron() {
  if (!BOOST_ENABLED) return;

  if (boost_cache_expire_all()) {
    watchdog('boost', t('Expired stale files from static page cache.'), WATCHDOG_NOTICE);
  }
}

/**
 * Implementation of hook_comment(). Acts on comment modification.
 */
function boost_comment($comment, $op) {
  if (!BOOST_ENABLED) return;

  switch ($op) {
    case 'insert':
    case 'update':
      // Expire the relevant node page from the static page cache to prevent serving stale content:
      if (!empty($comment['nid']))
        boost_cache_expire('node/' . $comment['nid'], TRUE);
      break;
  }
}

/**
 * Implementation of hook_nodeapi(). Acts on nodes defined by other modules.
 */
function boost_nodeapi(&$node, $op, $teaser = NULL, $page = NULL) {
  if (!BOOST_ENABLED) return;

  switch ($op) {
    case 'insert':
    case 'update':
    case 'delete':
      // Expire all relevant node pages from the static page cache to prevent serving stale content:
      if (!empty($node->nid))
        boost_cache_expire('node/' . $node->nid, TRUE);
      break;
  }
}

/**
 * Implementation of hook_taxonomy(). Acts on taxonomy changes.
 */
function boost_taxonomy($op, $type, $term = NULL) {
  if (!BOOST_ENABLED) return;

  switch ($op) {
    case 'insert':
    case 'update':
    case 'delete':
      // TODO: Expire all relevant taxonomy pages from the static page cache to prevent serving stale content.
      break;
  }
}

/**
 * Implementation of hook_user(). Acts on user account actions.
 */
function boost_user($op, &$edit, &$account, $category = NULL) {
  if (!BOOST_ENABLED) return;

  global $user;
  switch ($op) {
    case 'login':
      // Set special cookie to prevent logged-in users getting served pages from the static page cache.
      $expires = ini_get('session.cookie_lifetime');
      $expires = (!empty($expires) && is_numeric($expires) ? time() + (int)$expires : 0);
      setcookie(BOOST_COOKIE, $user->uid, $expires, ini_get('session.cookie_path'), ini_get('session.cookie_domain'), ini_get('session.cookie_secure') == '1');
      break;
    case 'logout':
      setcookie(BOOST_COOKIE, FALSE, time() - 86400, ini_get('session.cookie_path'), ini_get('session.cookie_domain'), ini_get('session.cookie_secure') == '1');
      break;
    case 'insert':
      // TODO: create user-specific cache directory.
      break;
    case 'delete':
      // Expire the relevant user page from the static page cache to prevent serving stale content:
      if (!empty($account->uid))
        boost_cache_expire('user/' . $account->uid);
      // TODO: recursively delete user-specific cache directory.
      break;
  }
}

/**
 * Implementation of hook_settings(). Declares administrative settings for a module.
 *
 * @deprecated in Drupal 5.0.
 */
function boost_settings() {
  require_once BOOST_PATH . '/boost.admin.inc';
  return boost_settings_form();
}

//////////////////////////////////////////////////////////////////////////////
// OUTPUT BUFFERING CALLBACK

/**
 * PHP output buffering callback.
 *
 * NOTE: objects have already been destructed so $user is not available.
 */
function _boost_ob_handler($buffer) {
  // Ensure we're in the correct working directory, since some web servers (e.g. Apache) mess this up here.
  chdir(dirname($_SERVER['SCRIPT_FILENAME']));

  // Check the currently set content type; at present we can't deal with anything else than HTML.
  if (_boost_get_content_type() == 'text/html') {
    if (strlen($buffer) > 0) { // Sanity check
      boost_cache_set($_GET['q'], $buffer);
    }
  }

  // Allow the page request to finish up normally
  return $buffer;
}

/**
 * Determines the MIME content type of the current page response based on
 * the currently set Content-Type HTTP header.
 *
 * This should normally return the string 'text/html' unless another module
 * has overridden the content type.
 */
function _boost_get_content_type($default = NULL) {
  static $regex = '/^Content-Type:\s*([\w\d\/\-]+)/i';

  // The last Content-Type header is the one that counts:
  $headers = preg_grep($regex, explode("\n", drupal_set_header()));
  if (!empty($headers) && preg_match($regex, array_pop($headers), $matches))
    return $matches[1]; // found it

  return $default;
}

//////////////////////////////////////////////////////////////////////////////
1	<?php
2	// $Id: boost.module,v 1.3 2006/11/27 16:34:49 arto Exp $
3
4	/**
5	* @file
6	* Provides static page caching for Drupal.
7	*/
8
9	//////////////////////////////////////////////////////////////////////////////
10	// BOOST SETTINGS
11
12	define('BOOST_PATH', dirname(__FILE__));
13	define('BOOST_FRONTPAGE', drupal_get_normal_path(variable_get('site_frontpage', 'node')));
14
15	define('BOOST_ENABLED', variable_get('boost', CACHE_DISABLED));
16	define('BOOST_FILE_PATH', variable_get('boost_file_path', 'cache'));
17	define('BOOST_FILE_EXTENSION', variable_get('boost_file_extension', '.html'));
18	define('BOOST_CACHEABILITY_OPTION', variable_get('boost_cacheability_option', 0));
19	define('BOOST_CACHEABILITY_PAGES', variable_get('boost_cacheability_pages', ''));
20	define('BOOST_FETCH_METHOD', variable_get('boost_fetch_method', 'php'));
21	define('BOOST_PRE_PROCESS_FUNCTION', variable_get('boost_pre_process_function', ''));
22	define('BOOST_POST_UPDATE_COMMAND', variable_get('boost_post_update_command', ''));
23	define('BOOST_CRON_LIMIT', variable_get('boost_cron_limit', 100));
24
25	// This cookie is set for all logged-in users, so that they can be excluded
26	// from caching (or, in the future, get a user-specific cached page):
27	define('BOOST_COOKIE', variable_get('boost_cookie', 'DRUPAL_UID'));
28
29	// This line is appended to the generated static files; it is very useful
30	// for troubleshooting (e.g. determining whether one got the dynamic or
31	// static version):
32	define('BOOST_BANNER', variable_get('boost_banner', "<!-- Page cached by Boost at %date -->\n"));
33
34	// This is needed since the $user object is already destructed in _boost_ob_handler():
35	define('BOOST_USER_ID', $GLOBALS['user']->uid);
36
37	//////////////////////////////////////////////////////////////////////////////
38	// BOOST INCLUDES
39
40	require_once BOOST_PATH . '/boost.helpers.inc';
41	require_once BOOST_PATH . '/boost.api.inc';
42
43	//////////////////////////////////////////////////////////////////////////////
44	// DRUPAL API HOOKS
45
46	/**
47	* Implementation of hook_help(). Provides online user help.
48	*/
49	function boost_help($section) {
50	switch ($section) {
51	case 'admin/modules#name':
52	return t('boost');
53	case 'admin/modules#description':
54	return t('Provides a performance and scalability boost through caching Drupal pages as static HTML files.');
55	case 'admin/help#boost':
56	$file = drupal_get_path('module', 'boost') . '/README.txt';
57	if (file_exists($file))
58	return '<pre>' . implode("\n", array_slice(explode("\n", @file_get_contents($file)), 2)) . '</pre>';
59	break;
60	case 'admin/settings/boost':
61	return '<p>' . '</p>'; // TODO: add help text.
62	}
63	}
64
65	/**
66	* Implementation of hook_perm(). Defines user permissions.
67	*/
68	function boost_perm() {
69	return array('administer cache');
70	}
71
72	/**
73	* Implementation of hook_menu(). Defines menu items and page callbacks.
74	*/
75	function boost_menu($may_cache) {
76	$access = user_access('administer cache');
77	$items = array();
78	if ($may_cache) {
79	// TODO: define menu actions for cache administration.
80	}
81	return $items;
82	}
83
84	/**
85	* Implementation of hook_init(). Performs page setup tasks.
86	*/
87	function boost_init() {
88	// Stop right here unless we're being called for an ordinary page request
89	if (strpos($_SERVER['PHP_SELF'], 'index.php') === FALSE)
90	return;
91
92	// TODO: check interaction with other modules that use ob_start(); this
93	// may have to be moved to an earlier stage of the page request.
94	if (!variable_get('cache', CACHE_DISABLED) && BOOST_ENABLED) {
95	// We only support GET requests by anonymous visitors:
96	global $user;
97	if (empty($user->uid) && $_SERVER['REQUEST_METHOD'] == 'GET') {
98	// Make sure no query string (in addition to ?q=) was set, and that
99	// the page is cacheable according to our current configuration:
100	if (count($_GET) == 1 && boost_is_cacheable($_GET['q']))
101	ob_start('_boost_ob_handler');
102	}
103	}
104
105	// Executed when saving Drupal's settings:
106	if (!empty($_POST['edit']) && $_GET['q'] == 'admin/settings') {
107	// Forcibly disable Drupal's built-in SQL caching to prevent any conflicts of interest:
108	variable_set('cache', CACHE_DISABLED);
109
110	// TODO: handle 'offline' site maintenance settings.
111
112	$old = variable_get('boost', '');
113	if (!empty($_POST['edit']['boost'])) {
114	// Ensure the cache directory exists or can be created
115	file_check_directory($_POST['edit']['boost_file_path'], FILE_CREATE_DIRECTORY, 'boost_file_path');
116	}
117	else if (!empty($old)) { // the cache was previously enabled
118	if (boost_cache_expire_all())
119	drupal_set_message('Static cache files deleted.');
120	}
121	}
122	}
123
124	/**
125	* Implementation of hook_exit(). Performs cleanup tasks.
126	*
127	* For POST requests by anonymous visitors, this adds a dummy query string
128	* to any URL being redirected to using drupal_goto().
129	*
130	* This is pretty much a hack that assumes a bit too much familiarity with
131	* what happens under the hood of the Drupal core function drupal_goto().
132	*
133	* It's necessary, though, in order for any session messages set on form
134	* submission to actually show up on the next page if that page has been
135	* cached by Boost.
136	*/
137	function boost_exit($destination = NULL) {
138	// Check that hook_exit() was invoked by drupal_goto() for a POST request:
139	if (!empty($destination) && $_SERVER['REQUEST_METHOD'] == 'POST') {
140
141	// Check that we're dealing with an anonymous visitor. and that some
142	// session messages have actually been set during this page request:
143	global $user;
144	if (empty($user->uid) && ($messages = drupal_set_message())) {
145
146	// Check that the page we're redirecting to has been cached by Boost
147	// and really necessitates special handling:
148	extract(parse_url($destination));
149	$path = ($path == base_path() ? '' : substr($path, strlen(base_path())));
150	if (boost_is_cached($path) && empty($query)) {
151	// FIXME: call any remaining exit hooks since we're about to terminate.
152
153	// Add a query string to ensure we don't serve a static copy of
154	// the page we're redirecting to, which would prevent the session
155	// messages from showing up:
156	$destination = url($path, 't=' . time(), $fragment, TRUE);
157
158	// Do what drupal_goto() would do if we were to return to it:
159	exit(header('Location: ' . $destination));
160	}
161	}
162	}
163	}
164
165	/**
166	* Implementation of hook_form_alter(). Performs alterations before a form
167	* is rendered.
168	*/
169	function boost_form_alter($form_id, &$form) {
170	// Alter Drupal's settings form by hiding the default cache enabled/disabled control (which will now always default to CACHE_DISABLED), and add our own control instead.
171	if ($form_id == 'system_settings_form') {
172	require_once BOOST_PATH . '/boost.admin.inc';
173	$form['cache'] = boost_system_settings_form($form['cache']);
174	}
175	}
176
177	/**
178	* Implementation of hook_cron(). Performs periodic actions.
179	*/
180	function boost_cron() {
181	if (!BOOST_ENABLED) return;
182
183	if (boost_cache_expire_all()) {
184	watchdog('boost', t('Expired stale files from static page cache.'), WATCHDOG_NOTICE);
185	}
186	}
187
188	/**
189	* Implementation of hook_comment(). Acts on comment modification.
190	*/
191	function boost_comment($comment, $op) {
192	if (!BOOST_ENABLED) return;
193
194	switch ($op) {
195	case 'insert':
196	case 'update':
197	// Expire the relevant node page from the static page cache to prevent serving stale content:
198	if (!empty($comment['nid']))
199	boost_cache_expire('node/' . $comment['nid'], TRUE);
200	break;
201	}
202	}
203
204	/**
205	* Implementation of hook_nodeapi(). Acts on nodes defined by other modules.
206	*/
207	function boost_nodeapi(&$node, $op, $teaser = NULL, $page = NULL) {
208	if (!BOOST_ENABLED) return;
209
210	switch ($op) {
211	case 'insert':
212	case 'update':
213	case 'delete':
214	// Expire all relevant node pages from the static page cache to prevent serving stale content:
215	if (!empty($node->nid))
216	boost_cache_expire('node/' . $node->nid, TRUE);
217	break;
218	}
219	}
220
221	/**
222	* Implementation of hook_taxonomy(). Acts on taxonomy changes.
223	*/
224	function boost_taxonomy($op, $type, $term = NULL) {
225	if (!BOOST_ENABLED) return;
226
227	switch ($op) {
228	case 'insert':
229	case 'update':
230	case 'delete':
231	// TODO: Expire all relevant taxonomy pages from the static page cache to prevent serving stale content.
232	break;
233	}
234	}
235
236	/**
237	* Implementation of hook_user(). Acts on user account actions.
238	*/
239	function boost_user($op, &$edit, &$account, $category = NULL) {
240	if (!BOOST_ENABLED) return;
241
242	global $user;
243	switch ($op) {
244	case 'login':
245	// Set special cookie to prevent logged-in users getting served pages from the static page cache.
246	$expires = ini_get('session.cookie_lifetime');
247	$expires = (!empty($expires) && is_numeric($expires) ? time() + (int)$expires : 0);
248	setcookie(BOOST_COOKIE, $user->uid, $expires, ini_get('session.cookie_path'), ini_get('session.cookie_domain'), ini_get('session.cookie_secure') == '1');
249	break;
250	case 'logout':
251	setcookie(BOOST_COOKIE, FALSE, time() - 86400, ini_get('session.cookie_path'), ini_get('session.cookie_domain'), ini_get('session.cookie_secure') == '1');
252	break;
253	case 'insert':
254	// TODO: create user-specific cache directory.
255	break;
256	case 'delete':
257	// Expire the relevant user page from the static page cache to prevent serving stale content:
258	if (!empty($account->uid))
259	boost_cache_expire('user/' . $account->uid);
260	// TODO: recursively delete user-specific cache directory.
261	break;
262	}
263	}
264
265	/**
266	* Implementation of hook_settings(). Declares administrative settings for a module.
267	*
268	* @deprecated in Drupal 5.0.
269	*/
270	function boost_settings() {
271	require_once BOOST_PATH . '/boost.admin.inc';
272	return boost_settings_form();
273	}
274
275	//////////////////////////////////////////////////////////////////////////////
276	// OUTPUT BUFFERING CALLBACK
277
278	/**
279	* PHP output buffering callback.
280	*
281	* NOTE: objects have already been destructed so $user is not available.
282	*/
283	function _boost_ob_handler($buffer) {
284	// Ensure we're in the correct working directory, since some web servers (e.g. Apache) mess this up here.
285	chdir(dirname($_SERVER['SCRIPT_FILENAME']));
286
287	// Check the currently set content type; at present we can't deal with anything else than HTML.
288	if (_boost_get_content_type() == 'text/html') {
289	if (strlen($buffer) > 0) { // Sanity check
290	boost_cache_set($_GET['q'], $buffer);
291	}
292	}
293
294	// Allow the page request to finish up normally
295	return $buffer;
296	}
297
298	/**
299	* Determines the MIME content type of the current page response based on
300	* the currently set Content-Type HTTP header.
301	*
302	* This should normally return the string 'text/html' unless another module
303	* has overridden the content type.
304	*/
305	function _boost_get_content_type($default = NULL) {
306	static $regex = '/^Content-Type:\s*([\w\d\/\-]+)/i';
307
308	// The last Content-Type header is the one that counts:
309	$headers = preg_grep($regex, explode("\n", drupal_set_header()));
310	if (!empty($headers) && preg_match($regex, array_pop($headers), $matches))
311	return $matches[1]; // found it
312
313	return $default;
314	}
315
316	//////////////////////////////////////////////////////////////////////////////