On Tuesday, October 22, 2013 6:13:49 PM UTC+2, John A. Tamplin wrote: > > On Tue, Oct 22, 2013 at 12:11 PM, Thomas Broyer > <[email protected]<javascript:> > > wrote: > >> As far as modularization is concerned, we'd like to have dependencies >> always in the same direction: rebind->client->shared->server (or >> rebind->client->server->shared, depending on modules) so we can make a >> "client" artifact depending on a "server" artifact, or a "client" and >> "server" artifacts both depending on a "shared" artifact (but with no >> dependency between "client" and "server"). In some cases, we'll split >> packages into distinct artifacts (there are cycles at the package level, >> but not when looking only at the classes); e.g. c.g.g.user.client (Window, >> Timer, etc.) c.g.g.resources and c.g.g.junit to have a package not >> depending on I18N and other things, and another one (or several) with >> additional dependencies. >> > > Why would it be acceptable to have shared code depend on server code? >
I probably meant "rebind" rather than "shared"; but I seem to remember some shared code making direct calls to server/vm, with a super-source version. In this case, the shared and server packages would live in a "server" artifact, and the client package and super-source version would live in a "client" artifact, with no "shared" artifact. There's also the case of classes referenced in annotations (e.g. @ProxyFor and @Service in RequestFactory) > -- http://groups.google.com/group/Google-Web-Toolkit-Contributors --- You received this message because you are subscribed to the Google Groups "GWT Contributors" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
