Oh yes... apologize... I did not see any bad side-effect after removing the scheme from defaultShindigTestHost. Only endPoints as you suggested.
doug On 6/20/11 5:55 PM, "Li Xu" <[email protected]> wrote: > Hi, Doug > Thanks for the comments. > There were two properties.... I have seen the exception only when I remove > scheme from this property from container.js > "osapi" : { > // The endpoints to query for available JSONRPC/REST services > "endPoints" : [ "//%host%${CONTEXT_ROOT}/rpc" ] > }, > I agree with you that the java code should be fixed to get scheme from > request...will look further. > > For the other property: > "defaultShindigTestHost":"//%authority%", > > I haven't seen any side effects yet... I'd like to suggest to update that > property first if there's no known problem. > > thanks! > li > > > > > > From: > daviesd <[email protected]> > To: > <[email protected]> > Date: > 06/20/2011 05:35 PM > Subject: > Re: remove scheme from defaultShindigTestHost in container.js > > > > If I remove the scheme then the server-side fails during listMethods (it > uses the same js value that is used client side). > > org.apache.shindig.gadgets.render.DefaultServiceFetcher retrieveServices > SEVERE: Failed to fetch services methods from endpoint > //myshindigserver:8443/shindig/rpc. Error Missing schema for request: > //myshindigserver:8443/shindig/rpc?method=system.listMethods > > I think the java code would need to be changed to tack on the current > scheme > of the servlet request. > > doug > > > On 6/20/11 4:26 PM, "Li Xu" <[email protected]> wrote: > >> Hello, all >> >> Can we remove scheme from "defaultShindigTestHost" in container.js? >> eg >> "defaultShindigTestHost":"//%authority%", >> insteadof >> "defaultShindigTestHost":"http://%authority%";, >> Right now it's not convenient. Each time we need to modify container.js > if >> scheme needs to be updated. >> Another approach I could think of is to replace the scheme based on the >> value from incoming request.... >> >> Could anyone please suggest if there's any flaw to remove scheme here? I >> have run a few tests and didn't catch anything bad... >> >> thanks! >> li >> > > > >
