Hi all! Yesterday Kris pointed out that we are not following the SimplePush Spec[1] when it comes to the 'pushEndpoint' that is returned when registering a channel. The spec shows that an absolute URL is returned but we are returning a relative URL. For the AeroGear SimplePush Server this is being address by AGPUSH-261[2], and an instance that returns absolute URL is available at http://endpointurldns-dbevenius.rhcloud.com:8000/simplepush. This change will affect the JavaScript client, the UnifiedPush Server, the Admin UI and possibly QE integration tests. I won't go into all the changes for the components listed above as others can provide more accurate information about them if needed. Just wanted to give a heads up to all about this. /Dan [1] https://wiki.mozilla.org/WebAPI/SimplePush/Protocol#PushServer_-.3E_UserA... [2] https://issues.jboss.org/browse/AGPUSH-261 _______________________________________________ aerogear-dev mailing list aerogear-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/aerogear-dev

>Looks like we need/should get rid of the "pushNetworkURL" on the SP Variant class and store a (SP installation) specific (simple)PushEndpoint. +1 >I know that the SP spec names it pushEndpoint, but I feel it's better to name it "simplePushEndpoint", to explicitly make sure that this is only for Installations of a SimplePushVariant. +1 That makes sense naming it simplePushEndpoint I think. >I wonder if, instead of the ENTIRE URL, we only could just store the "hash" (here "d9b74644") and continue to use the pushNetworkURL on the SP Variant -1 for this with the reasons you outlined. I think it might be more flexible if the UnifiedPush Server does not have to deal with a 'pushNetworkURL' for SimplePush.

/Dan On 8 August 2013 10:51, Matthias Wessendorf <matzew(a)apache.org&gt; wrote: > Hi, > > here is an update for UnifiedPush (not yet working), that indicates a > strategy: > > > https://github.com/matzew/aerogear-unified-push-server/commit/dcf1c96d637... > > > Looks like we need/should get rid of the "pushnetwork URL" on the SP > Variant class and store a (SP installation) specific (simple)PushEndpoint. > I know that the SP spec names it pushEndpoint, but I feel it's better to > name it "simplePushEndpoint", to explicitly make sure that this is only for > Installations of a SimplePushVariant. > > > With that change we now store a lot of redundant metadata (for all the SP > installations), which is most of the URL. Looking at the endpoint JSON > object: > { > "messageType": "register", > "channelID": "d9b74644-4f97-46aa-b8fa-9393985cd6cd", > "status": 200, > "pushEndpoint": "http://pushserver.example.org/d9b74644" > } > > I wonder if, instead of the ENTIRE URL, we only could just store the > "hash" (here "d9b74644") and continue to use the pushNetworkURL on the SP > Variant (here: "http://pushserver.example.org"). > > HOWEVER, I think this does not work when the URL of the "request > pushendpoint" server (e.g. "pushserver-foo.something.org") is different > from the endpoint that receives the HTTP PUTs, from the > AppServer/UnifiedPushServer. > > SP connect: > > SPClient = AeroGear.SimplePushClient({ > simplePushServerURL: "http://foobar.server.com:7777/simplepush", > onConnect: spConnect, > onClose: spClose > }); > > > received endpoint payload: > > { > ... > "pushEndpoint" : "http://something.different.server.com/blahBlahBlub" > } > > > > > -Matthias > > > > > On Thu, Aug 8, 2013 at 9:53 AM, Daniel Bevenius < > daniel.bevenius(a)gmail.com&gt; wrote: > >> Hi all! >> >> Yesterday Kris pointed out that we are not following the SimplePush >> Spec[1] when it comes to the 'pushEndpoint' that is returned when >> registering a channel. The spec shows that an absolute URL is returned but >> we are returning a relative URL. >> >> For the AeroGear SimplePush Server this is being address >> by AGPUSH-261[2], and an instance that returns absolute URL is available at >> http://endpointurldns-dbevenius.rhcloud.com:8000/simplepush. >> >> This change will affect the JavaScript client, the UnifiedPush Server, >> the Admin UI and possibly QE integration tests. >> >> I won't go into all the changes for the components listed above as >> others can provide more accurate information about them if needed. Just >> wanted to give a heads up to all about this. >> >> /Dan >> >> [1] >> https://wiki.mozilla.org/WebAPI/SimplePush/Protocol#PushServer_-.3E_UserA... >> [2] https://issues.jboss.org/browse/AGPUSH-261 >> >> _______________________________________________ >> aerogear-dev mailing list >> aerogear-dev(a)lists.jboss.org >> https://lists.jboss.org/mailman/listinfo/aerogear-dev >> > > > > -- > Matthias Wessendorf > > blog: http://matthiaswessendorf.wordpress.com/ > sessions: http://www.slideshare.net/mwessendorf > twitter: http://twitter.com/mwessendorf > > _______________________________________________ > aerogear-dev mailing list > aerogear-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/aerogear-dev > _______________________________________________ aerogear-dev mailing list aerogear-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/aerogear-dev

What's the timeframe of the change?

On Thu, 8 Aug 2013 09:53:37 +0200 Daniel Bevenius <daniel.bevenius(a)gmail.com&gt; wrote: > Hi all! > > Yesterday Kris pointed out that we are not following the SimplePush Spec[1] > when it comes to the 'pushEndpoint' that is returned when registering a > channel. The spec shows that an absolute URL is returned but we are > returning a relative URL. > > For the AeroGear SimplePush Server this is being address by AGPUSH-261[2], > and an instance that returns absolute URL is available at > http://endpointurldns-dbevenius.rhcloud.com:8000/simplepush. > > This change will affect the JavaScript client, the UnifiedPush Server, the > Admin UI and possibly QE integration tests. I don't think that impact on QE integration tests is that big, it should be a simple change. What's the timeframe of the change? > > I won't go into all the changes for the components listed above as others > can provide more accurate information about them if needed. Just wanted to > give a heads up to all about this. > > /Dan > > [1] > https://wiki.mozilla.org/WebAPI/SimplePush/Protocol#PushServer_-.3E_UserA... > [2] https://issues.jboss.org/browse/AGPUSH-261 _______________________________________________ aerogear-dev mailing list aerogear-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/aerogear-dev

On Aug 8, 2013, at 6:41 AM, Daniel Bevenius <daniel.bevenius(a)gmail.com&gt; wrote: > > What's the timeframe of the change? > I can't really speak for all parties here, but for SimplePush and UnifiedPush we should be able to have this ready by Monday afternoon ready to be included in the OpenShift cartridge. That is if Kris and Luke, or anyone else for that matter, does not object. No objection from me. I believe the only change needed on the client has been made and is already in master. > > > > On 8 August 2013 13:30, Karel Piwko <kpiwko(a)redhat.com&gt; wrote: > On Thu, 8 Aug 2013 09:53:37 +0200 > Daniel Bevenius <daniel.bevenius(a)gmail.com&gt; wrote: > > > Hi all! > > > > Yesterday Kris pointed out that we are not following the SimplePush Spec[1] > > when it comes to the 'pushEndpoint' that is returned when registering a > > channel. The spec shows that an absolute URL is returned but we are > > returning a relative URL. > > > > For the AeroGear SimplePush Server this is being address by AGPUSH-261[2], > > and an instance that returns absolute URL is available at > > http://endpointurldns-dbevenius.rhcloud.com:8000/simplepush. > > > > This change will affect the JavaScript client, the UnifiedPush Server, the > > Admin UI and possibly QE integration tests. > > I don't think that impact on QE integration tests is that big, it should be a > simple change. What's the timeframe of the change? > > > > > I won't go into all the changes for the components listed above as others > > can provide more accurate information about them if needed. Just wanted to > > give a heads up to all about this. > > > > /Dan > > > > [1] > > https://wiki.mozilla.org/WebAPI/SimplePush/Protocol#PushServer_-.3E_UserA... > > [2] https://issues.jboss.org/browse/AGPUSH-261 > > _______________________________________________ > aerogear-dev mailing list > aerogear-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/aerogear-dev > > _______________________________________________ > aerogear-dev mailing list > aerogear-dev(a)lists.jboss.org > https://lists.jboss.org/mailman/listinfo/aerogear-dev _______________________________________________ aerogear-dev mailing list aerogear-dev(a)lists.jboss.org https://lists.jboss.org/mailman/listinfo/aerogear-dev