[Tracker] huge database

[Garrett Mitchener <garrett mitchener gmail com>]

Hi, I just upgraded to Fedora 14, and since beagle is not under active development, I decided to switch to tracker.  However, I'm having a terrible time with the speed and size of the database.  I do a lot of research, writing, and programming, and I keep a lot of my files under an svn-controlled directory tree.  When I add that in as a directory to be indexed in the 'Locations' tab under 'Recursively indexed', tracker churns for hours on end, and creates a huge database:

$ du -h ~/.cache/tracker/; du -h ~/.local/share/tracker/data

6.6G /home/garrett/.cache/tracker/

16G /home/garrett/.local/share/tracker/data

And searching for a key word is generally quite slow.  In the options dialog box, I gave tracker a list of things not to index, such as auxiliary TeX files (*.bbl, *.blg, ...) and mathematica files (*.nb) and not to index version control directories like .svn or _darcs.  I told it to skip directories named 'Data' and 'Samples' where I keep large text files full of numerical data.

At least I think that's what I did.  I entered those in the 'Ignored Content' tab under 'Directories', 'Glob patterns to ignore'.

I can't figure out what's taking up so much time and space, and I don't know where to begin.  Is there a way to get some kind of list of what files it's searching and how much space they're taking up in the database?

By the way, there's only 32GB of data in that entire directory subtree, including stuff I told it to ignore.  When I leave that subtree out of the index, the database is a few hundred MB.  With it in, it grows to the above sizes, ~22GB.

Help!

-gm