On Thu, 2009-03-05 at 17:07 +0000, Martyn Russell wrote: > Hi all, > > Jamie, this week we realised that QDBM is really taking a lot of time > and is now the most prominent thing to fix to get some speed increase in > all areas. Couple this with the fact that we will be trying to move to > vstore shortly and FTS right after that, I wanted to have a peek and see > what is needed to get FTS going. > > First I thought I would check the code out you started and compare it to > SQLite's FTS3 version to see what changed. I noticed they have made > quite some changes, but mostly code white space updates. > > I have included in this email a patch with those (mostly white space > changes) which we could apply to our version so you can see the real > differences which need more care and attention. Those real differences > are attached in the two extra patches which diff the hash file and the > fts file. > > Are you happy for me to commit the update patch?
AFAICT it looks ok so please commit > > Does it look like it is worth updating our FTS support from the upstream > changes in SQLite? Not especially - see: http://www.sqlite.org/cvstrac/rlog?f=sqlite/ext/fts3/fts3.c I believe version in our source is from 1.21 so not a lot has changed there jamie > _______________________________________________ > tracker-list mailing list > tracker-list@gnome.org > http://mail.gnome.org/mailman/listinfo/tracker-list _______________________________________________ tracker-list mailing list tracker-list@gnome.org http://mail.gnome.org/mailman/listinfo/tracker-list