Can someone point me to a place in the GWT code where I might be able to produce a patch or something? I realize it's probably not the GWTers highest priority, but I could use some guidance.
Unless of course the silence is an indication that what I'm proposing is against the "spirit" of GWT. I don't think that is the case, but I would love to know for sure what can be done and what can't. Thanks, Nathan On Jan 26, 9:30 am, Nathan Wells <nwwe...@gmail.com> wrote: > Sony, > > I disagree with taking an event-based approach to RPC, but this isn't > really the forum for that discussion > > John, > > My problem is notClassserializability, but rather the fact that the > GWT compiler rejects any code path that has > "SomeServerProcedure.class", since the compiler thinks it will need to > translate thatclassregardless of whether anyclassor instance > members are referenced. I'm suggesting that the compiler is a little > bit too eager. When I use annotations, this isn't a problem, since > they are simply dropped by the compiler and kept by the server (where > I need them). However, that leads to an unnatural wrapping situation, > where I basically have a markerclasson the client-side that > references a server procedure. > > Does that make sense? I'm not sure if I'm explaining this well, and I > can give an example, but it will take a little more time than I have > immediately available. I'll probably write it up tonight. > > On Jan 26, 8:31 am, Sony <xsonymat...@gmail.com> wrote: > > > > > Think about an Event based RPC mechanism as outlined in this > > Article..http://sonymathew.blogspot.com/2010/01/gwt-jee-blueprint.html > > > Essentially, RPC is then merely firing/listening to events. You > > merely extend RemtoteRequestEvent and RemtoteRequestEvent and make > > sure any member content you add is serializable. > > > Sony > > > On Jan 25, 10:33 am, John Tamplin <j...@google.com> wrote: > > > > On Mon, Jan 25, 2010 at 11:26 AM, Nathan Wells <nwwe...@gmail.com> wrote: > > > > So, is it possible to get this sort of functionality? It would enable > > > > an RPC mechanism that is a lot easier and more natural. I don't think > > > > there would be any performance issues, but that is obviously hard to > > > > say for sure at this point. > > > > The RPC mechanism has no way to serialize a JavaClassinstance. > > > Conceivably, you could write a CustomerFieldSerializer forClassand send > > > the name to the client, and thenClass.forName (etc) on the server, though > > > I > > > haven't tried it. > > > > -- > > > John A. Tamplin > > > Software Engineer (GWT), Google -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
