/[drupal]/contributions/modules/apachesolr/apachesolr.index.inc
ViewVC logotype

Diff of /contributions/modules/apachesolr/apachesolr.index.inc

Parent Directory Parent Directory | Revision Log Revision Log | View Revision Graph Revision Graph | View Patch Patch

revision 1.1.2.13 by pwolanin, Wed Oct 14 19:08:19 2009 UTC revision 1.1.2.14 by pwolanin, Sun Nov 15 14:52:28 2009 UTC
# Line 1  Line 1 
1  <?php  <?php
2  // $Id: apachesolr.index.inc,v 1.1.2.12 2009/10/14 13:28:40 pwolanin Exp $  // $Id: apachesolr.index.inc,v 1.1.2.13 2009/10/14 19:08:19 pwolanin Exp $
3    
4  /**  /**
5   * @file   * @file
# Line 16  function apachesolr_add_node_document(&$ Line 16  function apachesolr_add_node_document(&$
16  }  }
17    
18  /**  /**
  * Strip control characters that cause Jetty/Solr to fail.  
  */  
 function apachesolr_strip_ctl_chars($text) {  
   // See:  http://w3.org/International/questions/qa-forms-utf-8.html  
   // Printable utf-8 does not include any of these chars below x7F  
   return preg_replace('@[\x00-\x08\x0B\x0C\x0E-\x1F]@', ' ', $text);  
 }  
   
 /**  
19   * Strip html tags and also control characters that cause Jetty/Solr to fail.   * Strip html tags and also control characters that cause Jetty/Solr to fail.
20   */   */
21  function apachesolr_clean_text($text) {  function apachesolr_clean_text($text) {
22    // Add spaces before stripping tags to avoid running words together.    // Add spaces before stripping tags to avoid running words together.
23    $text = filter_xss(str_replace(array('<', '>'), array(' <', '> '), $text), array());    $text = filter_xss(str_replace(array('<', '>'), array(' <', '> '), $text), array());
24    // Decode entities and then make safe any < or > characters.    // Decode entities and then make safe any < or > characters.
25    $text = htmlspecialchars(html_entity_decode($text, ENT_NOQUOTES, 'UTF-8'), ENT_NOQUOTES, 'UTF-8');    return htmlspecialchars(html_entity_decode($text, ENT_NOQUOTES, 'UTF-8'), ENT_NOQUOTES, 'UTF-8');
   // We must strip low bytes second in case there was an encoded  
   // low-byte character.  
   return apachesolr_strip_ctl_chars($text);  
26  }  }
27    
28  /**  /**
# Line 79  function apachesolr_node_to_document($ni Line 67  function apachesolr_node_to_document($ni
67        }        }
68      }      }
69      $text .= "\n\n" . implode(' ', $extra);      $text .= "\n\n" . implode(' ', $extra);
     $text = apachesolr_strip_ctl_chars($text);  
70    
71      $document = new Apache_Solr_Document();      $document = new Apache_Solr_Document();
72      $document->id = apachesolr_document_id($node->nid);      $document->id = apachesolr_document_id($node->nid);
# Line 99  function apachesolr_node_to_document($ni Line 86  function apachesolr_node_to_document($ni
86      }      }
87      $document->body = apachesolr_clean_text($text);      $document->body = apachesolr_clean_text($text);
88      $document->type  = $node->type;      $document->type  = $node->type;
89      $document->type_name = apachesolr_strip_ctl_chars(node_get_types('name', $node));      $document->type_name = node_get_types('name', $node);
90      $document->created = apachesolr_date_iso($node->created);      $document->created = apachesolr_date_iso($node->created);
91      $document->changed = apachesolr_date_iso($node->changed);      $document->changed = apachesolr_date_iso($node->changed);
92      $last_change = (isset($node->last_comment_timestamp) && $node->last_comment_timestamp > $node->changed) ? $node->last_comment_timestamp : $node->changed;      $last_change = (isset($node->last_comment_timestamp) && $node->last_comment_timestamp > $node->changed) ? $node->last_comment_timestamp : $node->changed;
93      $document->last_comment_or_change = apachesolr_date_iso($last_change);      $document->last_comment_or_change = apachesolr_date_iso($last_change);
94      $document->comment_count = isset($node->comment_count) ? $node->comment_count : 0;      $document->comment_count = isset($node->comment_count) ? $node->comment_count : 0;
95      $document->name = apachesolr_strip_ctl_chars($node->name);      $document->name = $node->name;
96    
97      $path = 'node/' . $node->nid;      $path = 'node/' . $node->nid;
98      $document->url = url($path, array('absolute' => TRUE));      $document->url = url($path, array('absolute' => TRUE));
# Line 118  function apachesolr_node_to_document($ni Line 105  function apachesolr_node_to_document($ni
105        $language = empty($node->language) ? '' : $node->language;        $language = empty($node->language) ? '' : $node->language;
106        $output = drupal_get_path_alias($path, $language);        $output = drupal_get_path_alias($path, $language);
107        if ($output && $output != $path) {        if ($output && $output != $path) {
108          $document->path_alias = apachesolr_strip_ctl_chars($output);          $document->path_alias = $output;
109        }        }
110      }      }
111    
# Line 314  function apachesolr_nodeapi_mass_update( Line 301  function apachesolr_nodeapi_mass_update(
301    $time = time();    $time = time();
302    try {    try {
303      $solr = apachesolr_get_solr();      $solr = apachesolr_get_solr();
304      $solr->deleteMultipleById($ids);      $solr->deleteByMultipleIds($ids);
305      apachesolr_index_updated($time);      apachesolr_index_updated($time);
306      foreach ($nodes as $node) {      foreach ($nodes as $node) {
307        // There was no exception, so update the table.        // There was no exception, so update the table.
# Line 340  function apachesolr_nodeapi_mass_delete( Line 327  function apachesolr_nodeapi_mass_delete(
327    }    }
328    try {    try {
329      $solr = apachesolr_get_solr();      $solr = apachesolr_get_solr();
330      $solr->deleteMultipleById($ids);      $solr->deleteByMultipleIds($ids);
331      apachesolr_index_updated($time);      apachesolr_index_updated($time);
332      // There was no exception, so update the table.      // There was no exception, so update the table.
333      db_query("DELETE FROM {apachesolr_search_node} WHERE nid  IN (" . db_placeholders($nids) . ")", $nids);      db_query("DELETE FROM {apachesolr_search_node} WHERE nid  IN (" . db_placeholders($nids) . ")", $nids);
Legend:
Removed from v.1.1.2.13  
changed lines
  Added in v.1.1.2.14
  ViewVC Help
Powered by ViewVC 1.1.3