InSite Search Release Notes

Build Compatibility

  • 2.1.x is compatible with CMS 9 releases up to SR5.
  • 2.2.x through 2.8.x is compatible with CMS 9 SR6 through CMS 9 SR8.
  • 2.9.x - 2.13.x is compatible CMS 9 SR6+ and CMS 10.0 - 10.3

    Note: New Asset indexing enhancements in 2.13+ are only supported in CMS 10.3 or above. Refer to the notes on Legacy Asset Indexing if you are indexing content from CMS 10.2 and below.

Current Version: InSite Search 2.13.44 - December 31, 2019

  • Asset indexing is now optimized so that changes to metadata or categorization to assets do not require full asset re-indexing.

    Note: This optimization applies to CMS 10.3+ published content only. For customers publishing from CMS 10.2 or below, refer to the note in Legacy Asset Indexing Mode.

  • Optimizations to asset binary file indexing, along with additional logging output during binary file indexing.
  • Renamed Assets from CMS 10 would previously not trigger a re-index of the assets. Now, when an asset is renamed, the indexer creates a new record for the newly-renamed asset. In order to remove the previous asset entry, a full publish will need to occur so that the indexer is aware of the binary file no longer being present on disk.
  • Added two new attribute settings for QueryFields.xml entries: 1) partialfieldnamematch, which allows for partial field name support, and 2) wildcard. See QueryFields for details.
  • Keymatch now supports a fifth data column, Abstracts, which can be used to provide abstract text to be returned with a keymatch search result.
  • The sample SiteCrawlerSource.config file contained a misspelling in the useCanonicalUrlInHtml setting.
  • Exact phrase search is now supported by encapsulating multiple terms in double quotes. This feature neither supports spell checking nor does it accept additional terms that do not match the exact phrase passed into the query.

    Note: This may cause performance issues if your query includes fields with large bodies of text (e.g., fulltext fields) and is not recommended for all implementations.

  • Added the urlExclusionPrefixes configuration for HTML Site Crawler to exclude indexing of pages that contain the provided path.
  • Asset changes that were not in a Documents folder path would not trigger re-indexing on changes.
  • Asset metadata changes now trigger re-indexing of data.
  • Asset binary file data would still be indexed even when the asset type had been flagged as non-indexabale at the CMS schema level.
  • Synonym entries now support multi-word values.
  • H1-H6 fields can be used as matching frag fields in the HiliteFields.xml configuration.
  • The file used for side-by-side re-indexing was not properly denormalizing the path values to the current read/write index location during test scenarios.
  • Terms with underscores will be tokenized as a single term instead of separate terms (e.g., Spider_Man).
  • Asset ID is now stored as a field value in the document index.
  • Asset binary files published from a CMS 10.3 environment are now indexing based on timestamp change from the Assets.xml file in the published data set. A legacy asset indexing mode setting has been added to support content from CMS 10.2 and below.
  • Asset descriptive metadata is returned in search results.
  • Asset metadata list items are indexed in a single field. Each list item value is delimited with an ellipses within this single field.
  • Fixed an issue with search implementations in a DSS where permission to the index folder location would be locked, resulting in an unresponsive DSS.
  • For implementations that are indexing two different publishing target data sets, the asset metadata would not be indexed for the second source in the configuration list.
  • Searches against the API using q as a parameter would return broken results. The q parameter is not a valid and will be ignored.
  • Added the setting ContentTermSizeLimit (default value: 50) for spellchecker source configuration file in order to set a max limit of characters per field to index. This prevents large indexing times for large sources of content.
  • Content would not be indexed on pages that contained List, Link, or Linkset elements and were set to expand on publish. To support this change in index behavior, the IGX_Presentations and LingualMaps elements are ignored during indexing.
  • Categories Stats count of associated indexed items would not match the count of search results filtered on the same category ID.
  • External asset metadata from CMS 10.3 published content would not be picked up by the indexer.
  • The search query analyzer would not pass in hyphenated terms to a search query; instead, it split on the hyphenated term.

For previous release notes, see InSite Search - Previous Releases.