I'm not sure Chromium has any fts1 databases, I think the original patch was applied there for completeness.
The change from fts2 to fts3 has been made in the history system, but it only applies to new data, and hasn't yet rolled out to stable. So we wouldn't be able to even start to cease using it for awhile. Furthermore, for performance reasons we segment and finalize the history data by month, so older data will continue to use fts2. We could add new code to upgrade that older data, though there would be some risk of that causing new problems (some of the data has been idle for a long time, so no doubt we'll flush out undetected corruptions, etc), and the straight-forward migration strategy would require fts2 to stay linked until we had some confidence that most of the users had launched their browsers for long enough that the migration had completed. So it wouldn't really be the preferred solution, given that the current code is fairly well-tested (for our case). I suppose we could pull in the time window by using internals knowledge to upgrade fts2 tables manually. Either we could steal the document table and load it into an fts3 table, or we could edit the sqlite_master table directly to convert it to an fts3 table. Hmm. Or it's possible we could have fts3.c register as handling fts2 tables, which I think would work alright in the read-only case, though it's mis-leading at best. It is also possible that there are web-developer-controlled tables in the wild using fts2 (I mean like WHATWG Web SQL Database). I lobbied to include fts3 for that because I didn't think we should encourage fts2, but I don't know whether I caught it soon enough. I also don't know the status of this WRT Gears, though for purposes of Chromium upstreaming things I don't think that matters. -scott On Wed, Aug 18, 2010 at 1:49 PM, Richard Hipp <d...@sqlite.org> wrote: > I'm thinking that you shouldn't be using FTS1 and FTS2 in the first place. > They are untested and unsupported. We'll get around to patching them, if > you insist, but right now we are busy trying to 3.7.1 out the door. > > On Wed, Aug 18, 2010 at 4:41 PM, Paweł Hajdan, Jr. > <phajdan...@chromium.org>wrote: > >> On Tue, Aug 10, 2010 at 13:16, Paweł Hajdan, Jr. <phajdan...@chromium.org >> >wrote: >> >> > Now, how about fts1 and fts2? The original chromium patch is at >> > http://codereview.chromium.org/174387 . Could you take a look and >> suggest >> > a way to upstream those fixes to SQLite? >> > >> >> Ping about the above. Or have the fixes already been made and we just need >> to upgrade to new sqlite? >> _______________________________________________ >> sqlite-users mailing list >> sqlite-users@sqlite.org >> http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users >> > > > > -- > D. Richard Hipp > d...@sqlite.org > _______________________________________________ > sqlite-users mailing list > sqlite-users@sqlite.org > http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users > _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
