Beagle 0.3.0

* Many speed and memory performance improvements in both indexing and
  searching throughout the code.

* Support for date queries in the query language. For example,
  "date:2006" will return documents from 2006. "date:2005.04" will
  return documents from April 2005. Ranges are also supported. For
  example, "date:2006.05.01-2007" will return all documents
  between May 1, 2006 and December 31, 2006.

* Rewritten Thunderbird backend, which uses a Thunderbird extension to
  send data to be indexed to the Beagle daemon.

* All new Firefox and Epiphany extensions.

* Experimental network searches. You can search between multiple
  Beagle daemons, which can be automatically discovered through
  Avahi. Daemons can also be manually configured, useful for WAN
  searches.

* Experimental web interface to the Beagle daemon. An AJAX interface
  allows users to search for data through their web browser.

* Support for additional external metadata on previously-indexed
  documents. This allows new backends or external applications to
  attach metadata that can then be searched on.

* New Nautilus metadata backend, implemented on top of the new
  external metadata support. Indexes Nautilus emblems and notes, and
  adds search keywords for these.

* Snowball analyzers are now used to process text. This is the first
  step toward language specific analyzers.

* New Opera history backend.

* Individual backends can now be enabled and disabled using the
  beagle-settings tool.

* Ignored file patterns (like *.o, *.a, *~, etc.) are no longer
  hardcoded, and can be set through the configuration.

* Changes to our TextCache to use a sqlite database for small files
  and on-disk files for larger ones. This makes the disk usage of the
  TextCache much more efficient.

* Use Mono's updated Mono.Data.Sqlite code, rather than our own hacked
  up copy of the old Mono.Data.SqliteClient code. As a result of
  this, we now require sqlite3.

* A new configuration system, which allows system administrators to
  set defaults in /etc/beagle/config-files. This also makes it easier
  for backend and filter authors to add their own configuration
  settings.

* Use taglib-sharp (replacing unmaintained entagged-sharp) for
  extracting metadata from audio files.

* An option to turn off aggressive indexing of data while the
  screensaver is on.

* A new classification for files -- the "file type" -- which
  classifies similar types together. For instance, PNGs, JPEGs, and
  GIFs all have file type "image". These can be searched using the
  "filetype" keyword.

* Now store snippets for emails.

* New extended attribute format. Beagle now uses a single extended
  attribute, user.Beagle, to store all of its information. The old
  format is migrated to the new one as files are reindexed.

* Queries are run serially rather than in parallel, since most of the
  search overhead is in disk IO. Reduces memory usage considerably,
  and helps prevent disk thrashing on large searches.

* Print out warnings about possibly insufficient inotify watches, and
  don't periodically recrawl directories that don't have them. Fixes
  the 100% CPU indexing loop bug of death.

* Various tweaks to the scheduler, to make throttle large numbers of
  tasks better.

* Handle large batches of file deletions (ie, an "rm -rf" on a
  directory) and file additions (ie, untarring into a watched
  directory) in a much more efficient manner.

* Require evolution-sharp 0.12, which implies Evlution 2.8 or 2.10
  because of API changes to the library.

* Some experimental support for XMP sidecar files.

* A new QueryPart for searching explicitly for URIs.

* Support for excluding KMail folders.

* Support for indexing Evolution notes.

* Refactor the indexing of IM conversations.

* New TeX filter.

* Structure snippets rather than returning them as HTML strings.
  This gives application authors considerably more control over how
  to display them.

* Scan the first 4k of files to determine their MIME type in addition
  to looking at the file extension, and choose the better match.

* PDFs are now included when indexing system documentation.

* Improvements to indexing files inside archives. Now all extracted
  files are handled in the index helper process, removing a costly
  round trip.

* Redirect crash logs from Mono into the Beagle log files. They were
  previously lost and users had to reproduce problems with --fg.

* Add many more keywords to the query language; run beagle-query
  --keywords to see a full list.

* Lots and lots of indexing and searching bugfixes. Too many to
  count.