ANNOUNCE: Beagle 0.2.14


I'm pleased to announce the release of Beagle 0.2.14.  This version
has many bug fixes, and some great new features as well.  Memory usage
is also way down in this version.  We strongly recommend all users
upgrade to this release.  Read on for more info.


To download the 0.2.14 tarball or learn more, visit the Beagle wiki at:

The latest gossip is available at:

We still talk about Beagle on the dashboard-hackers mailing list:

In poker, the "dead man's hand" is two-pair hand of aces and eights:

Beagle is a desktop-independent service for indexing and searching
your data.

The Beagle daemon transparently monitors your data and updates the
index to reflect any changes.  On an inotify-enabled system, these
updates happen more-or-less in real time.  So for example,
* Files are immediately indexed when they are created, are re-indexed
  when they are modified, and are dropped from the index upon deletion.
* E-mails are indexed upon arrival.
* IM conversations are indexed as you chat, a line at a time.

Beagle supports many different file formats including OpenOffice
documents, Microsoft Word documents, PDFs, HTML files, images, audio
and video files, archive files and their contents, source code, plain
text files, and many more.

Beagle can extract information from your file system; Evolution,
Thunderbird, and KMail emails and their attachements; Evolution,
Thunderbird, and KAddressbook addressbooks; Evolution calendars; Gaim
and Kopete instant messenger conversations; feeds from several RSS
aggregators; Tomboy, KNotes, and Labyrinth notes; Konqueror browsing
history; system documentation; and more.  Beagle also indexes tags on
your photos from F-Spot and Digikam.

Beagle also provides Firefox and Epiphany extensions that index web
pages in real-time as the user visits them.

Beagle uses the Lucene indexing system from the Apache project and the
prodigious Doug Cutting, ported to .NET by George Aroush.

Beagle includes an optional GNOME-based graphical tool for searching
the index that the daemon creates.  This application doesn't query the
index directly; it passes the search terms to the daemon and the
daemon sends any matches back.  The user interface then renders the
results and allows you to perform useful actions on the matching

Indexing your data requires a fair amount of computing power, but the
Beagle daemon tries to be as unobtrusive as possible.  It contains a
scheduler that works to prioritize tasks and control CPU usage, based
on whether or not you are actively using your workstation.


Beagle requires:
* Mono or better, along with the full Mono stack
* glib-sharp 2.4.0 or better (for the daemon and tools)
* gtk-sharp 2.4.0 or better (for the UI and some backends)
* GMime 2.2.0
* Libexif 0.5.7 or better
* shared-mime-info

For the best possible Beagle experience, you should also have:
* Mono 1.2.2 or better -- 1.2 and 1.2.1 have bugs which affect
  Beagle and should be avoided.
* Evolution-sharp 0.11.1 for Evolution 2.4 or 2.6, or 0.12.0 for
  Evolution 2.8
* libgsf 1.14.1 and gsf-sharp 0.8.1 from
* Galago 0.5.x
* Either wv 1.2.2, from or a *patched*
  wv 1.0.3.  The patch is available from
* An inotify-enabled kernel.  Inotify is in the mainline Linux
  kernel as of 2.6.13.

And other optional dependencies:


* Added the infrastructure necessary to handle indexing of archive
  files.  (Debajyoti Bera, Daniel Drake)
* Fix many issues with dates because .NET 1.1 was interpreting the
  dates as local time, not UTC.  (Bera)
* Indexables are no longer marked as indexed until all of its children
  are indexed first.  (Bera)
* Add infrastructure to signal clients when we're doing the initial
  index.  (Joe Shaw)
* Changes to how child indexables are dealt with in the index.  (Bera)
* Work around a bug in Mono 1.2.1, so that Beagle doesn't incorrectly
  think your home directory is on a remote filesystem like NFS.  (Joe)
* Use the XDG autostart specification to autostart the daemon, and
  deprecate the old --autostarted flag.  (Joe)
* Fix a URI comparison problem that would cause files to be repeatedly
  reindexed.  (Bera)
* Fix a bug in the tokenizer so that "001234" is tokenized as "1234".
* Tokenize longer numbers, like international phone numbers.  (Bera)
* After we store child indexable streams in the helper, close them
  rather than letting the GC do it for us.  (Joe)
* Don't set the "source" on indexables if they are already set.  This
  allows indexing service clients to set their own "source" names.
* Added a handler for SIGUSR2 to the helper process, which will print
  out what file it is currently filtering.  Should help debugging 100%
  CPU issues greatly.  (Joe)
* Greatly improve logging in the daemon and helper.  (Joe)

* Support child indexables in the file system backend.  (Bera)
* Report progress percentages in the Evolution mail backend.  (Joe)

* Added a new archive filter, to handle zip, tar, gzip and bzip2
  archives.  (Joe, Veerapuram Varadhan, Bera)
* Limit the archive filter to only index 30 files inside archives for
  now, to avoid excessive disk usage.  (Joe)
* Added Scribus filter.  (Alexander Macdonald)
* Fix a crash and clean up the code in the SVG filter.  (Alexander)
* Fix a potential crash in the HTML filter.  (Bera)
* Use the current encoding when decoding URLs in the HTML filter.
* Add text/troff to the list of supported MIME types in the man page
  filter.  (Joe)
* Turn on snippeting in the man page filter.  (Joe)

* Fix a logic bug in libbeagle that would cause responses to get lost
  if different complex messages were sent to the daemon.  (Joe)
* Print out the response in libbeagle apps if ENABLE_XML_DUMP is
  defined.  (Joe)
* Set the correct time zone info in BeagleTimestamp.  (Bera)
* Add example code for using the indexing service APIs.  (Bera)

* Only show the timestamp in beagle-extract-content if it's valid.
* Fix a problem in which child indexables from archives weren't being
  cleaned up in beagle-extract-content.  (Bera)
* Only show the number of total hits if we're in verbose mode with
  beagle-query, so that scripts can still easily deal only with URIs.
* Add support to beagle-settings to handle autostarting the daemon and
  UI with the XDG autostart spec.  (Joe)
* Show the percentages of progress in beagle-index-info for the
  backends that support it.  (Joe)
* Remove dead webservices code from beagle-settings tool.  (Joe)
* Fix a problem in which passing in --disable-directories to
  beagle-crawl-system would override --disable-filtering.  (Joe, 
  Pat Double)

* Show an informational box when the daemon is in the process of
  indexing the user's data.  (Lukas Lipka, Joe)
* New tile to show files inside archives.  (Joe)
* Added a status bar, and display in it the total number of matching
  documents and the number that are currently displayed.  (Joe)
* Fix some rendering ugliness in the details pane if you resized it.
* Add support for xdg-open if present.  (Joe)
* Fix the build so that Thunderbird was correctly opened from tiles.
  (Kevin Kubasik)
* Fix a crash in the image tile if the file didn't have an extension.
* Fix web tiles opening the handler for the MIME type, rather than the
  handler for the URL.  (Kevin, Joe)
* Fix a crash on right-click when the "Open With..." menu would have
  been blank.  (Joe)
* Use the XDG autostart spec to autostart beagle-search, and deprecate
  the old --autostarted method.  (Joe)
* Fix a potential crash in the image tile in the unlikely event that
  we'd attempt to composite the F-Spot logo on top of a standard MIME
  icon.  (Joe)
* Don't scale up RSS feed icons.  Liferea stores icons mostly at 16x16
  or 32x32, and those look terrible.  (Joe)

Memory optimizations:
* Fix a big leak in which we were leaking Lucene IndexReader
  instances.  This substantially reduces memory usage.  (Joe)
* Plug a leak in which QueryWorker instances were never removed from a
  static hash table.  (Joe)
* Don't store full Hit objects in the QueryWorker's URI hash table,
  since they're never used and hold references on a lot of objects.
* Fix a leak of URIs in LuceneFileQueryable, which is the base class
  for most of the smaller backends.  (Joe)
* Fix a leak in which the list of all the Evolution accounts were
  stored for each folder.  (Joe)
* Avoid some unnecessary boxing in Lucene, reducing memory usage.
* For all our non-tokenized fields, turn off storing norms in the
  Lucene index.  This will save disk and memory usage for those
  fields.  (Joe)
* Make the code in the IndexerRequest class a little better about
  object allocations.  (Joe)
* Fix an issue where characters were boxed every time we processed a
  URI.  This heavily reduces memory allocations.  (Joe)
* Use UTC when calculating load averages, to avoid uncessary
  allocations related to timezones.  (Joe)
* If --heap-buddy or --heap-shot are passed to the daemon, assume
  --debug-memory.  (Joe)
* If running with --heap-shot and the RSS increases by more than 5% or
  5 megs, or the Mono heap size increases by more than 10%,
  automatically send SIGPROF to the daemon to generate a memory
  snapshot.  (Joe)

* Updated Arabic translation.  (Djihed Afifi)
* Updated Catalan translation.  (Jordi Mas)
* Updated Czech translation.  (Jakub Friedl)
* Updated Galician translation.  (Ignacio Casal Quinteiro)
* Updated German translation.  (Hendrik Brandt)
* Updated Hungarian translation.  (Gabor Kelemen)
* Updated Japanese translation.  (Takeshi Aihana)
* Updated Norwegian bokm�translation.  (Kjartan Maraas)
* Updated Spanish translation. (Francisco Javier F. Serrador)
* Updated Swedish translation. (Daniel Nylander)

Everything else:
* Sync our xdgmime with upstream.  (Bera)
* Only check for wv1 if we have gsf-sharp.  We need both for MS Word
  filtering.  (Joe, Bera)
* Only build the po directory if we're building the GUI.  (Joe)
* Update the build system to use gnome-common's if
  it's available.  If not, use our own included copy.  (Joe)
* Require automake 1.8 to build.  (Joe, Pat, Max Weihle)
* Build against Mono's included SharpZipLib 0.84, instead of the older
  0.60 version.  (Joe)
* Fix an issue with the cron scripts not being included in the
  tarball.  (Joe)
* Only check for gtk-doc if libbeagle is enabled.  (Joe)


Thanks to new Mono profiling tools, memory usage is quite a bit lower
than in recent versions, but we still use more than we'd like.  We
continue to work on this.  However, in particular:

  * The Evolution mail backend can use a large amount of memory on
    large mailboxes.  This will be fixed in the next release.

  * The Thunderbird backends can take very large amounts of memory
    if you have large Mork files.  This issue is being addressed.

Certain extremely large documents can temporarily degrade your
system's performance while they are being indexed.

There are some race conditions that can occur with certain combinations of
file system operations.  In very rare cases it might be necessary to stop
and restart the daemon.

Certain files can crash the underlying libraries Beagle uses to
extract metadata.  This has been observed in MS Word and JPG files.
If you encounter such a crash, please report it to the upstream
developer of those libraries (wv1 and libexif for the above, respectively).

At this point in development, we cannot commit to stable APIs or file formats.
You will almost certainly need to delete your indexes and start again at some
point in the future.

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]