With Nearby Connections, each device has an endpointId, something similar to zkHk.
Getting the endpointId of others is trivial since it is returned by the API when scanning or connecting to other devices.
I must miss something, but I cannot find a way to get my own endpointId (apart implementing a mechanism where a connected peer echoes my id). It can be useful for some protocols where I want to follow what is sent to who.
The only thing I found is getLocalEndpointName but it returns my name, not my id. Even though it seems the C++ version of Nearby have it!
Do you have some ideas for Java/Kotlin? I specifically seek to get the endpointId, and not use alternatives like using a kind of GUID in the localendpoint name as a replacement.
Edit: Some example of usage
1) For instance, it can be interesting to implement some network mesh protocols. Several devices are interconnected making a global network, and each device add its endpointId in the incoming payload before sending it again, so others can check if they should send the payload to a device that already has it.
2) I may also want to specifically send a packet from device A to C through B acting as a relay, and add some "from: A" and "to: C" field in the payload so the network would know how to route the data and avoid some retransmission cycles. It is simpler to do that with endpointId since each device has a list of endpointId to which it is connected.
3) It can also be interesting for debug purpose. If I do some tests with a phone connected to several others (e.g. star network), it is easier to know from which phone a new piece of data is coming, all the more if I want to use name for another purpose.
Note: all of that could be done differently (e.g. use some unique identifier for the "name" of the devices and check that instead of the endpointId) but it seems a little cumbersome. All the more since endpointId guarantee a kind of unicity, whereas I must enforce it for the name. Moreover there isn't lots of information I can have on another device before exchanging data (only endpointId and name), so I feel I remove my last metadata slot if I use name as a substitute for endpointId.
As of today, you can't get your own endpoint id. We didn't see a reason you'd need it. Can you give a more detailed example of an algorithm where you need to know your own id?