================================================================================ $Id: CHANGELOG.txt,v 1.1.2.13 2008-09-03 01:38:55 markuspetrux Exp $ ================================================================================ -------------------------------------------------------------------------------- 2008-09-03, tag: 5.x-1.1 release - Fixed bug "Could not obtain last nid." caused by error building SQL condition for node type, when only one node type has been enabled. Thanks to valcker, issue: #302614 - Added flood control options with different thresholds for guests/reg. users. - Added option to disable logging of search keywords, or to log them only when keywords have been specified. Note that most requests could only be made by using tagadelic/faceted search related filters. - Added options to determine where the links of terms in the tag cloud should point to (taxonomy managed links or start search with predefined filters). - Changed functions arguments for sphinxsearch_tagadelic_build_data() and sphinxsearch_faceted_build_data(). These funcions are now easier to reuse, and eventually extend trying to not break backwards compatibility. -------------------------------------------------------------------------------- 2008-08-29, tag: 5.x-1.0 release - Removed configuration option for Sphinx document ids offset. I was confused by a big notice in the Sphinx manual stating that document ids should be unique, stop. The scope of such a restriction has been clarified by shodan, here: http://www.sphinxsearch.com/forum/view.html?id=2139#10402 :) It is required to run Drupal update.php and to rebuild all your indexes as soon as the module files have been updated! Sorry, for such a change at this point, though this is something that should be addressed as soon as possible. - Fix sphinxsearch_get_client() usage (missing ampersand operator, oops). - Added node type conversion when parsing grouping results by node type. - Added node tags in search results (theme). - Added support for author and content type tagadelic/faceted search. -------------------------------------------------------------------------------- 2008-08-27, tag: 5.x-1.0-RC1 - Added faceted search support. Faceted search block is rendered only under the context of search page. - Decode commas for query string stored along watchdog report for searches. - Added $debug argument to sphinxsearch_execute_query (developer use only). - Fix reentrancy issues in sphinxsearch_execute_query (reset filters, etc.). - Fix SetFilter() usage for content type and taxonomies. - Fix undefined notice in sphinxsearch_get_current_page(). - Changed name of tagadelic block variables (run Drupal update.php to apply). - Fix node text for xmlpipe document (title was indexed as part of the body). - Moving taxonomy related functions to a separate file. Reduce xmlpipe memory requirements when taxonomy module is not installed. - Added tagadelic alike pages. -------------------------------------------------------------------------------- 2008-08-26, tag: 5.x-1.0-BETA1 - Storing a number of common functions into a separate file. - Removed keys from search path. It all is based on query string. - Added option in module settings to select search page path. - Added options in module settings to customize excerpts for search results. - Added a check connection page to help us diagnose configuration issues. - Changed XMLPipe document, again. We now store the vids. All existing indexes get invalidated and need to be rebuilt. - Refactored almost everything related to the search form, so it now becomes a bit more modular. - Added basic tagadelic support. There is a single block with options to set number of tags, levels and sort order. -------------------------------------------------------------------------------- 2008-08-23, getting close to beta stage. - Removed changed attribute from XMLPipe indexes, added created attribute. - Added sort options (keywords relevance, node created, last updated time). - Changed internals of search page. Reusable components. This opens the door to create custom search pages and more. It might not be ready, but I started to see the light on how to build a framework from here to take advantage of Sphinx speed and features. - Started to add tagadelic alike blocks. Not finished yet. -------------------------------------------------------------------------------- 2008-08-22, not yet released. - Added support for indexing and filtering taxonomies. Note that this change invalidates indexes created previously. If you installed this module before, then you should check new module settings and rebuild all your indexes again. - preopen_indexes value in contrib/sphinx.conf is now set to 1, this is to minimize race conditions between XMLPipe generator and concurrent queries. - Added link to search syntax page from admin settings panel. - Hide content type filter in advanced search panel if only one content type has been enabled. Remove unnecessary clutter from user interface. - Fixed a few more typos and documentation issues here and there. - Removed options to select search matching modes. Not really necessary. Now, available modes are just All, Any and Phrase. Using Sphinx Extended2 mode. -------------------------------------------------------------------------------- 2008-08-21, not yet released. - Added "more information about search options" link to advanced search panel. - A few changes in module settings panel. -------------------------------------------------------------------------------- 2008-08-20, not yet released. - Added module settings to enable advanced matching modes. - Added selection of matching mode in advanced search fieldset. - Advanced search fieldset opened only by default when no results are obtained. - Revised help texts. -------------------------------------------------------------------------------- 2008-08-19, branched for Drupal 5. - Fixed typo in CSS. - Added method to check searchd daemon connection. - Enhanced user interface for searching. - Simplification of themable search results method. - Added missing option to setup XMLPipe to restart DB connection. - Added option to select size of excerpts in search results. -------------------------------------------------------------------------------- 2008-08-18, version not even alpha. - Rewritten XMLPipe generator. - Created sphinxsearch_scripts subdirectory to host the main entry point for the XMLPipe generator. This subdirectory allows to isolate PHP settings in .htaccess file that might be required to adjust memory_limit, etc. - Updated documentation in README.txt and contrib subdirectory. -------------------------------------------------------------------------------- 2008-08-15, version not even alpha. - Changed module name from sphinx to sphinxsearch. Please, see following issue for references on why there is another project called 'sphinx': http://drupal.org/node/296044 -------------------------------------------------------------------------------- 2008-07-30, version not even alpha. - Moved Sphinx client class script to lib subdirectory. - Added is_deleted attribute (boolean, default 0) to XMLPipe structure. - Added code in sphinx_queue_node_index() to update is_deleted attribute in realtime using Sphinx API UpdateAttributes(). - Fixed small bug in sphinx_get_node_text(). - Fixed search path used for watchdog in sphinx_search_page(). - Fixed query arguments returned by sphinx_search_form_submit(). - Added sample logrotate configuration script in contrib folder. - A few changes/improvements in contrib/sphinx script. -------------------------------------------------------------------------------- 2008-07-16, Started to work on this project. - It all started after reading a lot around Sphinx for Drupal (phpBB and other projects), and even looking and testing Solr and Xapian. Sphinx was looking nice and seemed to be more powerfull for our needs, and there was a PHP4 compatible API to work with, which is something we still need for a while. At the same time I had to resolve a few performance issues caused by our current search backends, based on MySQL queries. That was something that was in the TODO list for ages. However, Sphinx now started to look the solution, so I started to play with it. After a while, I realised it was necessary to find a method to bypass the need to build complex queries for Sphinx sources, it was going to be a lot easier to use XMLPipe. The problem then was more about performance, optimization, PHP memory_limits, max_execution_time, scalability, etc. but if it was possible to built a reasonable method to implement Sphinx for Drupal content, then it was going to be really great... Then I started to share my thoughts at g.d.o, then I shared a module with the d.o infra mailing list (eventhough it was early code, it covered XMLPipe and search user interface), then someone else looked interested in it, then... well, here we are. I need this module to solve a few of our problems now with D5, and in the near future with D6... it is nice and encouraging to do it in a way that this effort can be shared and reused by others.