NEW in 0.9.4 - 2010-05-11
=========================

Improvements / New:

  General:
  * Require poppler 0.12.2 instead of poppler-glib 0.4.5
  * Removed many unused variables from coverity reports
  * Various other fixes picked up from coverity reports
  * RSS miner, pkg-config should look for rest-0.6, not rest
 
  Functional Tests:
  * Added force-sqlite-misused-batch.sh and force-sqlite-misused.sh for NB#166684
  * Added for journal replay
  * Added for backup-restore
 
  Ontology:
  * Fixed typo for nfo:softwareCmdLine comment
  * Added tracker:notify for nco:EmailAddress
  * Added tracker:notify for nco:PhoneNumber
  * Added tracker:notify for nco:PostalAddress
  * Added tracker:notify for nfo:Document
  * Added tracker:notify for nfo:SoftwareApplication
  * Added tracker:indexed for nfo:belongsToContainer

  Data Generators:
  * Fixed %u use since it is deprecated in favour of %d in Python
 
  libtracker-common:
  * Fixed compilation error in validating ints for tracker-keyfile-object
 
  libtracker-db:
  * Enable read_uncommitted to fix locking issues
  * Properly shut down thread data on thread destruction, fixes sqlite3_close() log errors
  * Don't fsync/close already open databases if g_open() succeeds
  * Avoid strstr in uri_is_{parent|descendant} functions
  * Performance improvement for tracker:uri-is-parent function
 
  libtracker-extract:
  * Don't modify setlocale() return value, as it's statically stored
  * Protect against invalid values in tracker_date_guess()
 
  libtracker-miner:
  * Added missing .deps file for Vala bindings
  * Fixed memory leak in VAPI file
  * Fixed includes for libtracker-client in VAPI file
  * Updated nfo:fileLastModified query to use nfo:belongsToContainer, much faster
 
  libtracker-client:
  * Fixed typo in documentation for tracker_resources_sparql_query()
 
  tracker-store:
  * Fixed possible memory corruption
  * Fixed crash when client disappears
  * Fixed unref of TrackerDBCursor when NULL
  * Handle commit transaction error when importing turtle files and rollback

  tracker-extract:
  * Ported PDF extractor to not use poppler-glib, speed improvement
  * Don't run past genre array in mp3 extractor
  * Use nfo:HtmlDocument instead of nfo:Document in html extractor
  * Fixed compilation warning for msoffice extractor, use G_GSIZE_FORMAT not %d

  tracker-writeback:
  * Added MP4 and 3GPP video files xmp module 
 
  tracker-search:
  * Added --emails and list subjects/dates
  * Added --contacts and list names/addresses
  * Updated --detailed so we only report URNs if this is supplied
 
  tracker-tag:
  * Fixed possible use of uninitialised memory
 
  tracker-info:
  * Show results in shortened form, added --full-namespaces for old behaviour
 
  tracker-search-tool:
  * Fixed segmentation fault when there are no results
  * Added "Folders" category
  * Renamed "Office Documents" category to "Documents"
 
Bugs:
 
  * GB#616548 - DIrectory lists and de-duplication in tracker-preferences
  * GB#617082 - RFE Useability - Learning How to Start the Miners was a Little Unintuitive
  * NB#166767 - Every word in a long tag is indexed separately
  * NB#166684 - Database table is locked issue, SQLITE_MISUSED is reported SQLite error here
  * NB#167616 - Tracker's ISO 8601 timestamp implementation is not correct
  * NB#166717 - Only first DC:Subject and IPTC:Keywords tags are extracted

Translations:
 
  * Updated cs: Petr Kovar
  * Updated de: Mario Blättermann
 
Notes:
 
  Due to the HTML extractor changing the document type from
  nfo:Document to nfo:HtmlDocument, we would advise anyone with these
  files indexed to use:
 
    tracker-control --reindex-mime-type text/html
 
  This should make sure all indexed files are correctly adjusted.