On 07/04/2008, Jamie McCracken <[EMAIL PROTECTED]> wrote: > > > On Mon, 2008-04-07 at 16:29 +0300, Ivan Frade wrote: > > Hi all, > > > > We can discuss the roadmap for the next versions. > > > > Jamie: do you have any concrete plans? > > > > Here is my proposal: > > > > The final goal (whatever version number it is) must be a BIG release > > for application developers. In that release tracker will > > > > - Offer a complete configuration of service types from .service files: > > including the "Dirs" and "Extensions" properties and what we could > > need. > > > > - Include a nice documentation describing the .service format, the > > extractors architecture and how tracker index the stuff. (Using tomboy > > as example?) > > > > - Additional doc: how to add new languages, RDF query language. > > > > - If it is possible we should complete and document the "VFS" API. > > (Using "Daze" as example?) > > > > Then we can announce the big new version and try to convince > > applications developers to integrate their applications, because at that > > point it would be possible without recompiling tracker. > > > > Maybe it sounds very ambitious... At least is a nice long-term goal, > > isn't it? > > > > To achieve this we must: > > > > * finish the refactoring (long work, but neccesary) > > * maybe add unit tests (patch in bugzilla) > > * complete service type definition in .service files > > * check for new .service files when tracker starts. > > * split the current trackerd in two processes > > * implement the user DB support and complete VFS api > > * the eternal pending task of XESAM integration > > > > > > We can organize these tasks in different releases. What do you think? > > More tasks to the list? Maybe the refactor is enough to the 0.7.0 > > version... and we can develop the other features in 0.7.x releases. > > > > This is only my proposal to start the discussion. I expect a lot of > > comments :) > > > > 0.7 will pretty much be refactoring + xesam > > also possible is your patches for service file and other minor > improvements > > I will have all outstanding patches reviewed today > > > The end of the 0.7.x series will include everything needed for devs to > integrate firmly > > I also want to finalise index structure for 0.7 so no more reindexing > (that means adding things like a 8-bit language ID to the index and > language detection to the indexing)
If anybody starts hacking on the Xesam API then I should point you to xesam-tools[1]. It contains a small unit test suite for Xesam API providers in test/xesam-search-api-test.py, it is currently up to scratch with Xesam RC1, and probably also RC2 (haven't checked the details, but I will make a release shortly that will be guaranteed RC2 compat.). Both Strigi and Beagle pass it... If you take a hint :-) xesam-glib[2] should see a 0.2 release within one to two weeks. It does in fact contain some nifty classes and interfaces for implementing Xesam servers, but I believe Tracker is not meant to depend on dbus-glib, which xesam-glib does (although it would not be rocket science to remove the dbus-glib dep it is not a task I have planned to do in the short term future). Cheers, Mikkel [1]: http://xesam.org/people/kamstrup/xesam-tools [2]: http://xesam.org/people/kamstrup/xesam-glib
_______________________________________________ tracker-list mailing list tracker-list@gnome.org http://mail.gnome.org/mailman/listinfo/tracker-list
