Funnelback 9.1.0

Release notes for Funnelback 9.1.0

Released: 5th August 2010

New Features

• Integration with an enterprise repository connector system, supporting SharePoint and Lotus Domino gathering with high performance document level security.
• Improved query report presentation, including a dashboard and interactive charts.
• Incremental updating of query reports and ability to view reports while updating.
• Higher performance document filtering, multi-threaded filtering and in-crawl filtering.

Upgrade Issues

• The "SEARCH_SERVICE" environment variable is no longer supported, having been replaced by the bureau_mode_enabled setting in global.cfg. In environments where SEARCH_SERVICE was set to BUREAU, bureau_mode_enabled should be set to true.
• The change in default filtering frameworks noted above results in much improved filtering performance, however please be aware that:
  • Cached copies of Microsoft office documents will no longer reflect the original document formatting.
  • Microsoft Word 95 and Microsoft Word 6 documents are no longer supported.

Selected improvements and bugfixes

• Feature renaming as follows :
  • Fluster becomes Contextual Navigation.
  • Query Spike Detection becomes Pattern Analyser.
  • Featured Pages becomes Best Bets.
  • Query Expansion becomes Synonyms.
• Ability to specify both inclusive and exclusive date constraints.
• Activation of faceted and contextual navigation by default on filecopy collections.
• Relocation of Jetty working directory to avoid conflict with tmpwatcher and similar.
• Assorted improvements to document title fixing to reduce undesirable title changes.
• Improved query reports performance on 64bit windows platforms.
• Form tag reordering to allow s:cgi access from within form plugins.
• Security plugin mechanism for padre, allowing custom early binding document level security.
• Case insensitivity option for best bets and synonyms.
• Better query handling of German umlauts and Maori macrons.
• Improved result summaries for Chinese, Japanese, Korean and Thai.
• Fixed leaking of file-handles when updating large pattern analyser data sets.
• Improved support for HTML base href elements within forms.
• Support for HP TRIM license numbers.
• Improved reporting PDF layout.
• Fixed handling of facets containing parentheses.
• Fixed handling of BASE HREFs in indexed documents.
• Fixed issue with stemming when processing document at a time queries.
• Improved customisation options for contextual navigation.
• Corrected handling of collections with missing metamap.cfg and xml.cfg.
• Better error reporting where primary and secondary indexes differ in format.
• Indexing performance and scalability improvements.
• Dynamic resizing and increased scalability of indexer memory structures.
• Support for complex queries in spelling suggestions and contextual navigation.
• Improved stemming algorithm catching more cases where stemming should apply.
• Corrected reporting counts for geolocated data and queries using include or exclude scopes.