Hi all, Carlos, Good work! it looks like the improvement doesnt justify the complexity of the implementation. I think it is ok to use the standard dbus to communicate our processes, so no need to continue with the patch. I expected much more of this "direct communication" stuff in dbus.
Ivan On Fri, Feb 13, 2009 at 7:04 PM, Carlos Garnacho <[email protected]> wrote: > Hi! > > I've been investigating the usage of private peer-to-peer connections > between trackerd and tracker-indexer, and between tracker-indexer and > tracker-extract. > > Things turn out to be quite more complicated, since we have to make > tracker handle the other peer disconnecting, (re)spawning it, etc... > > >From the latency point of view, IMHO things don't improve sufficiently > to compensate the extra complexity. > > Wrt making tracker-indexer and tracker-extract "private" DBus-wise, I'd > better recommend making these drop petition that don't come from > org.freedesktop.Tracker, or just disable introspection and rename the > services as org.freedesktop.Tracker.Private.[Indexer|Extract], as the > miscellaneous gvfs services do. > > Any opinions? I'm attaching my half-cooked patch in case anybody finds a > reason to continue with this. > > Regards, > Carlos > > _______________________________________________ > tracker-list mailing list > [email protected] > http://mail.gnome.org/mailman/listinfo/tracker-list > >
_______________________________________________ tracker-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/tracker-list
