Note - the schema.xml has minor changes and a new protwords.txt file in included. All the configuration files on your sever should be updated and the Solr core reloaded or the servlet container restarted. You will need to re-index your content to take full advantage of the changes from #528086: html entities are not properly handled on search/display.
Note also that search result documents are new returned as stdClass objects, which could affect some custom code.
IMPORTANT: there are significant changes to the configuration files (especially schema.xml) in this release. You must install these new configuration files, and delete and re-index your data.
Apache Solr integration 7.x-1.0-beta4, 2011-04-04
------------------------------
#1116030 by cpliakas, pwolanin: Added titles to settings pages to add transparency as to which server's settings are being edited.
#1112022 by pwolanin, elliotttf: index node last_comment_timestamp too.
#1114798 by pwolanin: enhancements to config based on Solr 3.1 examples.
#1108618 by pwolanin: make numeric fields in 7.x use a sortable data type by default.
#1103602 by pwolanin: prevent PHP Notice when there is an invalid term reference field.
#1097988 by jpmckinney: avoid 'using default converter' warning on Solr startup.
#920482 by jpmckinney: $info_split['date'] template variable should contain changed, not created, date.
#379512 by pwolanin: separate schema field for indexing comments and "extra" information.
#871440 by jpmckinney: Solr taxonomy page displays search form and blocks when the user has no access.
#761990 by jhedstrom, pwolanin: switch to POST for long search ULRs.
#1112362 by pwolanin: cleanup following #1107502 to better use drupal_http_request.
#1107502 by pwolanin: Merge relevant parts of php client into DrupalApacheSolrService.
Apache Solr integration 7.x-1.0-BETA3, 2011-01-06
------------------------------
#1017836 by Janusman, pwolanin: fix empty search behavior.
#1017624 by pwolanin: Fix for "Did you mean" suggestion does not show.
#1007848 by pwolanin: Fixes for apachesolr_nodeaccess for Drupal 7.
#1017258 by pwolanin: Fix help text when there are no search results.
#1013122 by pwolanin: some JS cleanup to use jQuery proxy.
#1013136 by pwolanin: fix retain filters checkbox.
#1009398 by pwolanin: fix content bias forms.
#878996 by pwolanin, weri: don't break out of indexing loop if an exception is caught.
#880488 by pwolanin: security hardening: don't use unserialize on untrusted data.
#867448 by B-Prod | jpmckinney: Fixed Apache taxonomy search does not include the search.pages.inc file.
This is primariy a bug fix release, but minor new features have been added or backported from the 6.x-2.x version.
Note that the schema.xml and solrconfig.xml have changed, but should be backward-compatible with existing indexes. Update these xml files in your Solr conf directory and restart Solr.
#864160 by pwolanin: Allow the caller a last chance to modify the query and params.
#847040 by Agileware: Allow other modules to use facet only searching.
#536990 by pwolanin | jpmckinney, janusman: always index content as an anonymous user.
#751004 by jpmckinney, camidoo, rjbrown99, pwolanin: fix use of CCK API, handle NODE_BUILD_SEARCH_INDEX exclusion setting in the UI.
#373824 by pwolanin: show count of attached files to mimic upload module with node search.
#853564 by pwolanin: fix regression from #614644 - still need to see if there is POST data.
#592522 by pwolanin, quaoar: use multi-table sytax for MySQL to avoid killer sub-selects.
#614644 by pwolanin: Improved compatibility with other forms (based on Drupal 7 core issue patch).
This release contains several important bug fixes - in particular without the follow-up patch at #630798: Implement cache_apachesolr table, sites may experience significant performance issues due to an extra metadata request being made for every MLT or search request.
This release contains several important bug fixes - in particular without the follow-up patch at #630798: Implement cache_apachesolr table, sites may experience significant performance issues due to an extra metadata request being made for every MLT or search request.
This release contains several important bug fixes - in particular without the follow-up patch at #630798: Implement cache_apachesolr table, sites may experience significant performance issues due to an extra metadata request being made for every MLT or search request.
Also, the Drupal 7 language-neutral code was change from 'zxx' to 'und', and this module changed to follow suit - if you are using the language facet block with language-neutral content, you may want to reindex to get consistent codes in your index.
changelog entries:
#630798 by robertDouglass, pwolanin - critical fix for cache_get() calls.
#672518 by robertDouglass: Fixed Add new trie prefixes to helper function.
#668396 by pwolanin, closer to fix from #655006 for PHP notices.
#657648 by kcoop, Add Smaller Limit Options to Apache Solr Cron Indexing.
#664818 by pounard, robertDouglass, pwolanin fix incorrect watchdog() call.
#667650 by Dave Reid, Results of apachesolr_process_response() should return absolute URLs.
#667124 by pwolanin, Use current query not altered query for the breadcrumb.
Th 6.2 branch has been driving development for some months. There are several new features, such as comment search, OR facets, and the conditional display of facet blocks, that are unique to this branch. Some other features have already been backported to the 6.1 branch.
The plan for development between alpha1 and alpha2 is to continue adding features, date facets from CCK fields is first on the list, and to start significantly refactoring key APIs.
This branch is not recommended for production environments as long as it is in alpha.
The alpha1 is schema-compatible with the 6.1 branch and Acquia Search.
You must update to r22 of the client PHP module in order to use this version of the module. Do this by deleting the SolrPhpClient directory that is currently in the apachesolr module, and using svn to get the right one to replace it:
#502976: Other GET parameters ignored by Apache Solr Facet Blocks by pwolanin, Scott Reynolds, robertDouglass facet links and form submissions respect non ApacheSolr $_GET parameters. Note that this changes the interface API: get_url_querystring is now get_url_queryvalues and returns and array instead of a string.
Note the apachesolr_mlt module is removed by this release and combined with the framework module. Be sure that you remove all files from contrib/apachesolr_mlt when upgrading to avoid possible PHP errors due to duplicate functions.
Important note: due to #343252: make node access code multi-site aware, those using the Apache Solr Nodeaccess module will need to re-index their content for access control on searches to work correctly.
#305370 by pwolanin, Handle failed delete requests so unpublished/deleted content doesn't stay in the index.
#407570 by pwolanin and moshe weitzman, _alter for sort links, hide for < 2 results.
#392978 by pwolanin and ncameron, workaround for those using php 5.1, update README.
#402984 by JacobSingh and pwolanin, put MLT menu under the general ApacheSolr settings.
#401442 by Janusman and pwolanin, no sort block when 0 results.
#405732 by JacobSingh, pwolanin: Update to new SolrPhpClient (r6) and make ping() use drupal_http_request.
Important: as of this release you must get and install the PHP client library from its external project page at: http://code.google.com/p/solr-php-client/ See the README.txt for more details.
CHANGELOG since beta2:
#378222 by janusman and pwolanin, add boost settings for "sticky" and "promote".
#378566 by pwolanin, nodeaccess not correctly marking single nodes for re-indexing.
#378270 by pwolanin, suppress MLT admin link when there is no block content.
#378196 by pwolanin, remove PHP client from CVS per Drupal.org policy.
#231200 by janusman and pwolanin, turn on mapping of accented to non-accented ISO characters for indexing and search.
#377494 by pwolanin, Update text type in schema to new example.
#376270 by pwolanin, also add option to bias on recent comments/changes.
#337879 by pwolanin and blackdog, Store relative not absolute paths.
#376255 by pwolanin, Index more node fields, use boolean fields.
The Drupal 5 and Drupal 6 branches have fallen out of sync. The DRUPAL-5 branch doesn't have as many features as the DRUPAL-6 branch, but is likely more stable at this moment in time precisely because of all the new features being added to DRUPAL-6--1.
D5 and D6 now synced and aligned on ALPHA-3 because new features are still expected.
- Only one form setting per block for facet limits.
- Validation on block form now works.
- f.fieldname.facet.limit syntax.
- facet.field dynamically generated from block list.
- using individual dynamic fields for each vocabulary for accurate facet counts.
- fixed bug in author facets that excluded Anonymous from facet list.
Note: You will have to reconfigure your facet blocks after this update.
D5 and D6 now synced and aligned on ALPHA-3 because new features are still expected.
- Only one form setting per block for facet limits.
- Validation on block form now works.
- f.fieldname.facet.limit syntax.
- facet.field dynamically generated from block list.
- using individual dynamic fields for each vocabulary for accurate facet counts.
- fixed bug in author facets that excluded Anonymous from facet list.
Note: You will have to reconfigure your facet blocks after this update.
#230402 repairing CCK integration and changing strategy a bit. Now only checkboxes, selects and radio buttons generate facet blocks automatically. Other CCK fields can have facets and facet blocks by implementing hook_apachesolr_cck_field_mappings
#248552 add explicit check for required field nid
#230377 add results sorting to ApacheSolr. There is now a block that allows you to sort by Title, Type, Author, and Date. The solrsort GET parameter can additionally be used to sort by comment count, uid and nid, as well as any CCK text fields that are indexed. This change requires you to update the schema.xml file in your Solr installation, restart Solr, and re-index your site.
#248882 Fixed problem with indexing when no solr instance available
#248881 Adding hook_requirements so that Solr server status can be monitored more easily by administrator
#247358 created a function, apachesolr_has_searched, that acts a semaphore to only do Solr related stuff if there has been a search (like generate blocks). Fixed form_alter (hacked it, really) so that it doesn't instantiate a Solr object every time the search form is built (like on q=search/node). See the issue for discussion of deeper problem.
