Tim Newsome skrev:
Indeed we have a Thunderbird backend related bug here. To be more specific the mork implementation. It fails to load your .msf files containing information about what to index. This should however not affect how beagle index other things in other backends since all backends have separate indexes. Are you sure that they really are empty (check with beagle-info --index-info while running the daemon)? Might be another bug if they are. Perhaps you could include a complete log file in the bug you just opened?Joe Shaw wrote:Basically after I shut down beagle and start it up again, it appears to have lost all its indexes.This must be a bug in the Thunderbird backend. Can you file a bug about this at bugzilla.gnome.org and (if you can) attach the .msf fileI filed a bug. If it is a Thunderbird backend problem, then it's also causing problems with the rest of the index. I just reran the test, but let beagle index for an hour instead of 5 minutes. This way there are quite a few files that have "test" in them. But after shutting down, starting up, and rerunning the query, it still looks like the index is empty.
Don't know what we should do about the mork errors yet, though. I'm a bit skeptical if any effort should be put in "fixing" current code, since it already contains a bad design, lots of incorrect and bad code etc. A new and more thought out implementation is pretty much the only way to go at this point. This would most likely fix all open bugs in the bugzilla related to the thunderbird code as well (they are as far as I know all related to mork). Biggest problem is that I don't have the time to begin working on this, so I will need some help here. Maybe Kevin (or anybody else with some free time) wants to give this a try? I'll help out with as much assistance as I possibly can.
Thanks! Pierre