Re: [SyncEvolution] Concurrent sync sessions update - Almost done

On Thu, 2012-03-08 at 18:58 +0100, Chris Kühl wrote: > On Thu, Mar 8, 2012 at 6:23 PM, Patrick Ohly <patrick.ohly(a)intel.com&gt; wrote: > > On Thu, 2012-03-08 at 17:18 +0100, Chris Kühl wrote: > > 6c9a05a9db72f001d9834d2d24ac589f48fc5798 > > > >    dbus-server: Run sync sessions in separate processes > > > > ... > > > >    Sessions are separated into SessionResource and Session classes. A > >    SessionResource instance resides in the server process and serves as a > >    proxy to the Session instance which is in the child process. > > > > This naming seems rather arbitrary to me. Why call it "Resource" and not > > something like "Stub" or "Proxy"? > > > > Yeah, I'm not 100% happy with the naming either. They are subclasses > of Resource so it was the obvious choice. Renaming is not a problem > but I'd rather get finished with the more substantive changes needed > to complete this before doing that. Everything that minimizes the number of changes that I need to look at helps.

> > Or, perhaps even better, don't rename it at all on the server side. Then > > a whole range of diffs goes away: > > > > It's just that the Session is actually not in the Server anymore so it > seems a tad misleading to call it that in the Server. It's still the implementation of the D-Bus Session API in the server, isn't it? So for the sake of minimizing code churn, Session (in the server) and SessionImpl (in the client) might work.