On Mon, 2006-01-09 at 07:16 -0400, Patrick O'Callaghan wrote: [snip] > Unless I misunderstand what you're saying, you're moving it from > somewhere remote to somewhere local, which is a big difference. Yup. Locally. All 'journal'led (like you are in the offline mode and replayed when back online)
> But you can never give an upper bound to how long a remote sync will > take. Servers go down, networks are disconnected, power failures happen. > The essence of a distributed application (of which remote mail is a > fairly simple example) is that you can survive these events without loss > of data (which is not the same as perfect consistency between client and > server, but that was never guaranteed anyway). Once you can do that, you > can impose an upper bound on stopping Evo, say 30 seconds. That would be a good idea.(viz. A 30 sec upper bound.) With the Network manager support that has gone into HEAD (2.5.x) - a lot of issues are handled by evolution. Like when the network is down, Evolution switches to offline, and journals everything and replays it when Evo is back online. -partha _______________________________________________ Evolution-list mailing list [email protected] http://mail.gnome.org/mailman/listinfo/evolution-list
