http://gwt-code-reviews.appspot.com/1107801/diff/3001/4005 File user/src/com/google/gwt/user/client/rpc/ServiceDefTarget.java (right):
http://gwt-code-reviews.appspot.com/1107801/diff/3001/4005#newcode43 user/src/com/google/gwt/user/client/rpc/ServiceDefTarget.java:43: RpcToken getRpcToken(); On 2010/11/22 03:52:05, meder wrote:
On 2010/11/20 04:33:44, xtof wrote: > How likely is it that developers will be providing their own
implementations
of > ServiceDefTarget? We'd be breaking them by adding methods here.
I've seen it being done, so this will most definitely break some code.
I'm
waiting for someone from GWT team to look at this and tell me if this
is fine or
if I should add another interface.
Do you have any reference to non-generated code that implements ServiceDefTarget? I would expect only mocks would do so. That said, it wouldn't be much of a problem to put the RpcToken APIs on another interface. http://gwt-code-reviews.appspot.com/1107801/diff/13001/14002 File user/src/com/google/gwt/user/client/rpc/RpcToken.java (right): http://gwt-code-reviews.appspot.com/1107801/diff/13001/14002#newcode28 user/src/com/google/gwt/user/client/rpc/RpcToken.java:28: public interface RpcToken extends Serializable { On 2010/11/20 04:33:44, xtof wrote:
As mentioned in the draft design, I'm not sure RpcToken is the best
name for
this. It's really just a value that's implicitly passed along with
each RPC.
Maybe, RpcRequestHeader? But really I'd defer to GWT team for advice
on naming
here...
I like RpcToken better than that. Basically, all we are saying is it is some token automatically passed with each request, and RpcToken seems descriptive and appropriately generic. http://gwt-code-reviews.appspot.com/1107801/show -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
