On Tue, Jan 17, 2006 at 07:13:53PM +0100, Dieter Maurer wrote: > Paul Winkler wrote at 2006-1-16 14:32 -0500: > >On Mon, Jan 16, 2006 at 07:48:45PM +0100, Dieter Maurer wrote: > >> Synchronization modifies "bobobase_modification_time" on the target. > > > >Yep. > > > >> A safe method to check whether you have a current state is to > >> register the "bobobase_modification_time" (on source and target) > >> that resulted from the synchronization. > >> A resynchronization is necessary when the "bobobase_modification_time" > >> moved away from the registered "sync_bobobase_time"s. > > > >Ahhh, now I see. Thank you, this is probably cheaper than > >most of my ideas. > >I could store that info in a BTree of some flavor. > > An alternative would be a persistent subobject of the synchronized > object (such that its modification does not modify the > "bobobase_modification_time" of the primary object).
That's not good: unless i find a suitable core zope class to use, if the user ever uninstalls ZSyncer, all of his objects are broken! in general, i don't like to pile more data onto the poor user's unsuspecting objects. I have done it, i just don't like it :-) > I just notice that synchronization of structured (e.g. "ObjectManager") > objects might be a problem. As subobjects might also be synchronizable, > they, too, should get the correct times. Ugh, yes, that's a problem. -- Paul Winkler http://www.slinkp.com _______________________________________________ Zope maillist - Zope@zope.org http://mail.zope.org/mailman/listinfo/zope ** No cross posts or HTML encoding! ** (Related lists - http://mail.zope.org/mailman/listinfo/zope-announce http://mail.zope.org/mailman/listinfo/zope-dev )