<?php

/**
 * @file
 * Installation file for Link Checker module.
 */

/**
 * Defines the minimum MySQL version number this module supports.
 *
 * Previous versions do not support subselects, but for performance reasons
 * the module makes use of them whenever possible. PgSQL do not have this
 * kind of limitations in the versions supported by Drupal 6.
 */
define('LINKCHECKER_MINIMUM_MYSQL', '5.0');

/**
 * Implementation of hook_install().
 */
function linkchecker_install() {
  drupal_install_schema('linkchecker');

  $linkchecker_default_impersonate_user = user_load(1);
  variable_set('linkchecker_impersonate_user', $linkchecker_default_impersonate_user->name);
}

/**
 * Implementation of hook_uninstall().
 */
function linkchecker_uninstall() {
  drupal_uninstall_schema('linkchecker');

  variable_del('linkchecker_action_status_code_301');
  variable_del('linkchecker_action_status_code_404');
  variable_del('linkchecker_check_connections_max');
  variable_del('linkchecker_check_library');
  variable_del('linkchecker_check_links_interval');
  variable_del('linkchecker_check_useragent');
  variable_del('linkchecker_cleanup_links_last');
  variable_del('linkchecker_disable_link_check_for_urls');
  variable_del('linkchecker_extract_from_a');
  variable_del('linkchecker_extract_from_audio');
  variable_del('linkchecker_extract_from_embed');
  variable_del('linkchecker_extract_from_iframe');
  variable_del('linkchecker_extract_from_img');
  variable_del('linkchecker_extract_from_object');
  variable_del('linkchecker_extract_from_source');
  variable_del('linkchecker_extract_from_video');
  variable_del('linkchecker_filter_blacklist');
  variable_del('linkchecker_fqdn_only');
  variable_del('linkchecker_ignore_response_codes');
  variable_del('linkchecker_impersonate_user');
  variable_del('linkchecker_scan_blocks');
  variable_del('linkchecker_scan_comments');
  variable_del('linkchecker_scan_nodetypes');
}

/**
 * Implementation of hook_schema().
 */
function linkchecker_schema() {

  $schema['linkchecker_links'] = array(
    'description' => 'Stores all links.',
    'fields' => array(
      'lid'  => array(
        'type' => 'serial',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique link ID.',
      ),
      'urlhash' => array(
        'type' => 'varchar',
        'length' => 32,
        'not null' => TRUE,
        'description' => 'The indexable md5 hash of the {linkchecker_links}.url.',
      ),
      'url' => array(
        'type' => 'text',
        'not null' => TRUE,
        'description' => 'The full qualified link.',
      ),
      'method' => array(
        'type' => 'varchar',
        'length' => 4,
        'default' => 'HEAD',
        'not null' => TRUE,
        'description' => 'The method for checking links (HEAD, GET, POST).',
      ),
      'code' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => -1,
        'description' => 'HTTP status code from link checking.',
      ),
      'error' => array(
        'type' => 'text',
        'not null' => FALSE,
        'description' => 'The error message received from the remote server while doing link checking.',
      ),
      'fail_count' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 0,
        'description' => 'Fail count of unsuccessful link checks. No flapping detection. (Successful = 0, Unsuccessful = fail_count+1).',
      ),
      'last_checked' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 0,
        'description' => 'Timestamp of the last link check.',
      ),
      'status' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 1,
        'description' => 'Boolean indicating if a link should be checked or not.',
      ),
    ),
    'primary key' => array('lid'),
    'unique keys' => array('urlhash' => array('urlhash')),
    'indexes' => array(
      'method' => array('method'),
      'code' => array('code'),
      'fail_count' => array('fail_count'),
      'last_checked' => array('last_checked'),
      'status' => array('status'),
    ),
  );

  $schema['linkchecker_boxes'] = array(
    'description' => 'Stores all link references for boxes.',
    'fields' => array(
      'bid'  => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {boxes}.bid.',
      ),
      'lid' => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {linkchecker_links}.lid.',
      ),
    ),
    'primary key' => array('bid', 'lid'),
    'indexes' => array('lid' => array('lid')),
  );

  $schema['linkchecker_nodes'] = array(
    'description' => 'Stores all link references for nodes.',
    'fields' => array(
      'nid'  => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {node}.nid.',
      ),
      'lid' => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {linkchecker_links}.lid.',
      ),
    ),
    'primary key' => array('nid', 'lid'),
    'indexes' => array('lid' => array('lid')),
  );

  $schema['linkchecker_comments'] = array(
    'description' => 'Stores all link references for comments.',
    'fields' => array(
      'cid'  => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {comments}.cid.',
      ),
      'lid' => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {linkchecker_links}.lid.',
      ),
    ),
    'primary key' => array('cid', 'lid'),
    'indexes' => array('lid' => array('lid')),
  );

  return $schema;
}

/**
 * Implementation of hook_requirements().
 */
function linkchecker_requirements($phase) {
  $requirements = array();
  $t = get_t();

  switch ($phase) {
    case 'install' :
      // MySQL version with subselect support is required.
      $version = db_version();
      if (in_array($GLOBALS['db_type'], array('mysql', 'mysqli')) && version_compare($version, LINKCHECKER_MINIMUM_MYSQL) < 0) {
        $requirements['linkchecker_minimum_mysql'] = array(
          'title' => $t('MySQL database'),
          'value' => $version,
          'severity' => REQUIREMENT_ERROR,
          'description' => $t('Your MySQL Server is too old. Link Checker requires at least MySQL %version.', array('%version' => LINKCHECKER_MINIMUM_MYSQL)),
        );
      }
      break;
  }

  return $requirements;
}

/**
 * Upgrade module to new D5 schema.
 */
function linkchecker_update_5200() {
  $ret = array();

  // Module functions are required. Make sure the module is loaded.
  drupal_load('module', 'linkchecker');

  // Remove obsolete tables no longer required.
  db_drop_table($ret, 'linkchecker_tasks');
  db_drop_table($ret, 'linkchecker_results');

  // Create new tables.
  $schema['linkchecker_links'] = array(
    'description' => 'Stores all links.',
    'fields' => array(
      'lid'  => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique link ID.',
      ),
      'token' => array(
        'type' => 'varchar',
        'length' => 32,
        'not null' => TRUE,
        'description' => 'The indexable md5 hash of the {linkchecker_links}.url.',
      ),
      'url' => array(
        'type' => 'text',
        'not null' => TRUE,
        'description' => 'The full qualified link URL.',
      ),
      'method' => array(
        'type' => 'varchar',
        'length' => 4,
        'default' => 'HEAD',
        'not null' => TRUE,
        'description' => 'The method for checking links (HEAD, GET, POST).',
      ),
      'code' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 0,
        'description' => 'HTTP status code from link checking.',
      ),
      'error' => array(
        'type' => 'text',
        'not null' => FALSE,
        'description' => 'The error message received from the remote server while doing link checking.',
      ),
      'fail_count' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 0,
        'description' => 'Fail count of unsuccessful link checks. No flapping detection. (Successful = 0, Unsuccessful = fail_count+1).',
      ),
      'last_checked' => array(
        'type' => 'int',
        'not null' => TRUE,
        'default' => 0,
        'description' => 'Timestamp of the last link check.',
      ),
    ),
    'primary key' => array('lid'),
    'unique keys' => array('token' => array('token')),
  );

  $schema['linkchecker_boxes'] = array(
    'description' => 'Stores all link references for boxes.',
    'fields' => array(
      'bid'  => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {boxes}.bid.',
      ),
      'lid' => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {linkchecker_links}.lid.',
      ),
    ),
    'primary key' => array('bid', 'lid'),
  );

  $schema['linkchecker_nodes'] = array(
    'description' => 'Stores all link references for nodes.',
    'fields' => array(
      'nid'  => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {node}.nid.',
      ),
      'lid' => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {linkchecker_links}.lid.',
      ),
    ),
    'primary key' => array('nid', 'lid'),
  );

  // Create schema.
  db_create_table($ret, 'linkchecker_links', $schema['linkchecker_links']);
  db_create_table($ret, 'linkchecker_boxes', $schema['linkchecker_boxes']);
  db_create_table($ret, 'linkchecker_nodes', $schema['linkchecker_nodes']);

  // Upgrade settings. Could be less code, but is easier to follow.
  $ignore_response_codes = preg_split('/(\r\n?|\n)/', variable_get('linkchecker_ignore_responses', "200\n304\n401\n403"));

  // Filter all invalid responds codes and outdated error messages out.
  $ignore_response_codes = array_filter($ignore_response_codes, '_linkchecker_isvalid_response_code');

  // Make sure we have status code 200 and 304 in the ignore list.
  $ignore_response_codes = array_merge(array('200', '304'), $ignore_response_codes);
  $ignore_response_codes = array_unique($ignore_response_codes);

  variable_set('linkchecker_ignore_response_codes', implode("\n", $ignore_response_codes));
  $ret[] = array('success' => TRUE, 'query' => 'Ignored response codes have been upgraded to '. implode(",", $ignore_response_codes));

  // Remove obsolete settings.
  variable_del('linkchecker_ignore_responses');
  variable_del('linkchecker_rebuild');
  variable_del('linkchecker_maxtime');
  variable_del('linkchecker_socket_timeout');
  variable_del('linkchecker_max_links_per_node');
  variable_del('linkchecker_remove_after');
  variable_del('linkchecker_give_up');

  return $ret;
}

/**
 * Upgrade module to new D6 schema.
 */
function linkchecker_update_6200() {
  $ret = array();

  db_change_field($ret, 'linkchecker_links', 'lid', 'lid', array('type' => 'serial', 'not null' => TRUE));

  // Upgrade line break converter from D5 filter/2 to D6 filter/1.
  $linkchecker_filter_blacklist = variable_get('linkchecker_filter_blacklist', array());
  if (!empty($linkchecker_filter_blacklist) && $linkchecker_filter_blacklist['filter/2']) {
    $linkchecker_filter_blacklist['filter/2'] = 0;
    $linkchecker_filter_blacklist['filter/1'] = 'filter/1';
    variable_set('linkchecker_filter_blacklist', $linkchecker_filter_blacklist);
    $ret[] = array('success' => TRUE, 'query' => 'Upgraded the default exclusion of line break converter filter from D5 filter/2 to D6 filter/1.');
  }

  return $ret;
}

/**
 * Install linkchecker_comments table.
 */
function linkchecker_update_6201() {
  $ret = array();

  // Create new tables.
  $schema['linkchecker_comments'] = array(
    'description' => 'Stores all link references for comments.',
    'fields' => array(
      'cid'  => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {comments}.cid.',
      ),
      'lid' => array(
        'type' => 'int',
        'not null' => TRUE,
        'description' => 'Primary Key: Unique {linkchecker_links}.lid.',
      ),
    ),
    'primary key' => array('cid', 'lid'),
  );

  // Table may be created in 5201 by 5.x-2.x.
  if (!db_table_exists('linkchecker_comments')) {
    db_create_table($ret, 'linkchecker_comments', $schema['linkchecker_comments']);
  }

  return $ret;
}

/**
 * Remove unnecessary elements from custom submit buttons.
 */
function linkchecker_update_6202() {
  $ret = array();

  variable_del('linkchecker_analyze');
  variable_del('linkchecker_clear_analyze');

  return $ret;
}

/**
 * Make fields unsigned.
 */
function linkchecker_update_6204() {
  $ret = array();

  // Make linkchecker_links.lid unsigned.
  db_change_field($ret, 'linkchecker_links', 'lid', 'lid', array('type' => 'serial', 'unsigned' => TRUE, 'not null' => TRUE));
  if ($GLOBALS['db_type'] == 'pgsql') {
    db_add_primary_key($ret, 'linkchecker_links', array('lid'));
  }

  // Make linkchecker_boxes.bid and linkchecker_boxes.lid unsigned.
  db_drop_primary_key($ret, 'linkchecker_boxes');
  db_change_field($ret, 'linkchecker_boxes', 'bid', 'bid', array('type' => 'int', 'unsigned' => TRUE, 'not null' => TRUE));
  db_change_field($ret, 'linkchecker_boxes', 'lid', 'lid', array('type' => 'int', 'unsigned' => TRUE, 'not null' => TRUE));
  db_add_primary_key($ret, 'linkchecker_boxes', array('bid', 'lid'));

  // Make linkchecker_comments.cid and linkchecker_comments.lid unsigned.
  db_drop_primary_key($ret, 'linkchecker_comments');
  db_change_field($ret, 'linkchecker_comments', 'cid', 'cid', array('type' => 'int', 'unsigned' => TRUE, 'not null' => TRUE));
  db_change_field($ret, 'linkchecker_comments', 'lid', 'lid', array('type' => 'int', 'unsigned' => TRUE, 'not null' => TRUE));
  db_add_primary_key($ret, 'linkchecker_comments', array('cid', 'lid'));

  // Make linkchecker_nodes.nid and linkchecker_nodes.lid unsigned.
  db_drop_primary_key($ret, 'linkchecker_nodes');
  db_change_field($ret, 'linkchecker_nodes', 'nid', 'nid', array('type' => 'int', 'unsigned' => TRUE, 'not null' => TRUE));
  db_change_field($ret, 'linkchecker_nodes', 'lid', 'lid', array('type' => 'int', 'unsigned' => TRUE, 'not null' => TRUE));
  db_add_primary_key($ret, 'linkchecker_nodes', array('nid', 'lid'));

  return $ret;
}

/**
 * Remove any references to nodes that are not published and also remove their comments references.
 */
function linkchecker_update_6205() {
  $ret = array();

  $ret[] = update_sql("DELETE FROM {linkchecker_nodes} WHERE nid IN (SELECT nid FROM {node} WHERE status = 0)");
  $ret[] = update_sql("DELETE FROM {linkchecker_comments} WHERE cid IN (SELECT c.cid FROM {node} n INNER JOIN {comments} c ON c.nid = n.nid WHERE n.status = 0)");

  return $ret;
}

/**
 * Add status column to links table.
 */
function linkchecker_update_6206() {
  $ret = array();

  // Add the column only if not exists. Column may been added in D5 (5207).
  if (!db_column_exists('linkchecker_links', 'status')) {
    db_add_field($ret, 'linkchecker_links', 'status', array('type' => 'int', 'not null' => TRUE, 'default' => 1));
  }

  return $ret;
}

/**
 * Remove obsolete variable.
 */
function linkchecker_update_6207() {
  $ret = array();

  variable_del('linkchecker_cleanup_links_interval');

  return $ret;
}

/**
 * Change default value for linkchecker_links.code to -1.
 *
 * fsockopen may return 0 as an indication that the error occurred before
 * the connect() call. This is most likely due to a problem initializing the
 * socket.
 */
function linkchecker_update_6208() {
  $ret = array();

  db_change_field($ret, 'linkchecker_links', 'code', 'code', array('type' => 'int', 'not null' => TRUE, 'default' => -1));

  return $ret;
}

/**
 * Update the permissions table, to reflect changes to hook_perm.
 */
function linkchecker_update_6209() {
  $ret = array();

  $res = db_query('SELECT rid, perm FROM {permission}');
  $perms = array();
  while ($p = db_fetch_object($res)) {
    $perm = $p->perm;
    $perm = preg_replace('/access linkchecker/', 'access broken links report', $perm);
    $perms[$p->rid] = $perm;
  }

  foreach ($perms as $rid => $renamed_permission) {
    db_query("UPDATE {permission} SET perm = '%s' WHERE rid = %d", $renamed_permission, $rid);
    $ret[] = array('success' => TRUE, 'query' => 'UPDATE {permission} SET perm = ' . check_plain($renamed_permission) . ' WHERE rid = '. $rid);
  }

  return $ret;
}

/**
 * Add status code 302 to the default list of ignored response codes.
 */
function linkchecker_update_6210() {
  $ret = array();

  $ignore_response_codes = preg_split('/(\r\n?|\n)/', variable_get('linkchecker_ignore_response_codes', "200\n302\n304\n401\n403"));
  if (!in_array('302', $ignore_response_codes)) {
    $ignore_response_codes[] = '302';
    sort($ignore_response_codes);
    variable_set('linkchecker_ignore_response_codes', implode("\n", $ignore_response_codes));
    $ret[] = array('success' => TRUE, 'query' => 'Added the status code 302 to the list of ignored response codes.');
  }
  else {
    $ret[] = array('success' => TRUE, 'query' => 'No action taken. The status code 302 was already on the list of ignored response codes.');
  }

  return $ret;
}

/**
 * Remove obsolete variable.
 */
function linkchecker_update_6211() {
  $ret = array();

  variable_del('linkchecker_pathfilter_support');
  $ret[] = array('success' => TRUE, 'query' => 'Removed obsolete linkchecker_pathfilter_support variable.');

  return $ret;
}

/**
 * Change unique key and column name 'token' into 'urlhash'.
 */
function linkchecker_update_6212() {
  $ret = array();

  db_drop_unique_key($ret, 'linkchecker_links', 'token');
  db_change_field($ret, 'linkchecker_links', 'token', 'urlhash', array('type' => 'varchar', 'length' => 32, 'not null' => TRUE));
  db_add_unique_key($ret, 'linkchecker_links', 'urlhash', array('urlhash'));

  return $ret;
}

/**
 * Add RFC documenation domains back if they have been removed.
 */
function linkchecker_update_6213() {
  $ret = array();

  $linkchecker_disable_link_check_for_urls = array_filter(preg_split('/(\r\n?|\n)/', variable_get('linkchecker_disable_link_check_for_urls', LINKCHECKER_RESERVED_DOCUMENTATION_DOMAINS)));
  variable_set('linkchecker_disable_link_check_for_urls', implode("\n", array_unique(array_merge(explode("\n", LINKCHECKER_RESERVED_DOCUMENTATION_DOMAINS), $linkchecker_disable_link_check_for_urls))));
  $ret[] = array('success' => TRUE, 'query' => 'Added RFC documenation domains back if they have been removed.');

  return $ret;
}

/**
* #1321378: Improve performance of queries.
*/
function linkchecker_update_6214() {
  $ret = array();

  db_add_index($ret, 'linkchecker_boxes', 'lid', array('lid'));
  db_add_index($ret, 'linkchecker_comments', 'lid', array('lid'));
  db_add_index($ret, 'linkchecker_nodes', 'lid', array('lid'));

  return $ret;
}

/**
* Add status code 206 to the default list of ignored response codes.
*/
function linkchecker_update_6215() {
  $ret = array();

  $ignore_response_codes = preg_split('/(\r\n?|\n)/', variable_get('linkchecker_ignore_response_codes', "200\n206\n302\n304\n401\n403"));
  if (!in_array('206', $ignore_response_codes)) {
    $ignore_response_codes[] = '206';
    sort($ignore_response_codes);
    variable_set('linkchecker_ignore_response_codes', implode("\n", $ignore_response_codes));
    $ret[] = array('success' => TRUE, 'query' => 'Added the status code 206 to the list of ignored response codes.');
  }
  else {
    $ret[] = array('success' => TRUE, 'query' => 'No action taken. The status code 206 was already on the list of ignored response codes.');
  }

  return $ret;
}

/**
* #965720: Add indexes to improve performance of views queries.
*/
function linkchecker_update_6216() {
  $ret = array();

  db_add_index($ret, 'linkchecker_links', 'method', array('method'));
  db_add_index($ret, 'linkchecker_links', 'code', array('code'));
  db_add_index($ret, 'linkchecker_links', 'fail_count', array('fail_count'));
  db_add_index($ret, 'linkchecker_links', 'last_checked', array('last_checked'));
  db_add_index($ret, 'linkchecker_links', 'status', array('status'));

  return $ret;
}

/**
 * #1450672: Set user 1 as default user to impersonate content updates.
 */
function linkchecker_update_6217() {
  $ret = array();

  $linkchecker_default_impersonate_user = user_load(1);
  variable_set('linkchecker_impersonate_user', $linkchecker_default_impersonate_user->name);
  $ret[] = array('success' => TRUE, 'query' => 'Set user 1 as default user to impersonate content updates.');

  return $ret;
}

/**
 * Removed obsolete linkchecker_check_links_max variable.
 */
function linkchecker_update_6218() {
  $ret = array();

  variable_del('linkchecker_check_links_max');
  $ret[] = array('success' => TRUE, 'query' => 'Removed obsolete linkchecker_check_links_max variable.');

  return $ret;
}

/**
 * Upgrade outdated HTTP user agents.
 */
function linkchecker_update_6219() {
  $ret = array();

  $linkchecker_check_useragent = variable_get('linkchecker_check_useragent', 'Drupal (+http://drupal.org/)');

  $useragent_upgrade = array(
    'Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1;)' => 'Mozilla/5.0 (compatible; MSIE 9.0; Windows NT 6.1; WOW64; Trident/5.0)',
    'Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 6.0;)' => 'Mozilla/5.0 (compatible; MSIE 10.0; Windows NT 6.2; WOW64; Trident/6.0)',
    'Mozilla/5.0 (Windows; U; Windows NT 5.1; de; rv:1.9.0.5) Gecko/2008120122 Firefox/3.0.5' => 'Mozilla/5.0 (Windows NT 6.1; WOW64; rv:17.0) Gecko/20100101 Firefox/17.0',
    'Mozilla/5.0 (Windows; U; Windows NT 6.0; de; rv:1.9.0.5) Gecko/2008120122 Firefox/3.0.5' => 'Mozilla/5.0 (Windows NT 6.2; WOW64; rv:17.0) Gecko/20100101 Firefox/17.0',
  );

  if (isset($useragent_upgrade[$linkchecker_check_useragent])) {
    variable_set('linkchecker_check_useragent', $useragent_upgrade[$linkchecker_check_useragent]);
    $ret[] = array('success' => TRUE, 'query' => 'Upgraded outdated HTTP user agent.');
  }
  else {
    $ret[] = array('success' => TRUE, 'query' => 'User agent already up to date.');
  }

  return $ret;
}