On Wednesday 02 December 2015 07:54:39 ASHOKBABU CHANNA wrote: > Hi Thiago, > > > >If you use dashes, however, it's valid: > >coap+bt://00-11-22-33-44-55:777 > > Our approach was to use coap://00:11:22:33:44:55 with transport > having IP/BT/BLE information not a raw mac address.
That's not a valid URI. You cannot use ':' in the hostname without []. And you can only use [] for IPv6 and IPvFuture. > This is same as our gerrit changes. Which is why I had it reverted and started this discussion: it was requiring and parsing an invalid URI. > >In any case, please see the discussion going on inside the SWG about the > >use of fully qualified URIs. The proposed solution is to encode the device > >ID in the hostname, in which case you can't put the MAC address there. The > >URI would be of the form: > > > >oic://bf3718fa-3ba4-4898-b9e9-cb88a681cc4a/ > > > This will be valid for all transport not particularly for BT/BLE address. We > will adopt this approach once it is finalized. I hope SWG is also > discussing about direct unicast request where device ID is unknown. Explain the use-case: how do you know the device address, without a discovery? And besides, even in that case, why does it need to be a URI? Why can't it be an OCDevAddr? -- Thiago Macieira - thiago.macieira (AT) intel.com Software Architect - Intel Open Source Technology Center
