Oddbean new post about | logout
▲ ▼
Egge | 9 months ago (raw) | root | parent | reply | flag 
 Working on this now 🤙
I read through the spec and I think it’s truely a missed opportunity that a remote signer response does not reference the requests ID. Did you add that to your implementation by any chance?
▲ ▼
brugeman | 9 months ago (raw) | root | parent | reply | flag 
 The spec is too badly written. The reply has format of 
"content": "METHOD({id: <request-id>, result: <string>, error: <reason-string>})",
And the "result" field contains the method-specific payload.
The spec is being rewritten now with oauth-stuff added, hopefully it gets more readable.
▲ ▼
Egge | 9 months ago (raw) | root | parent | reply | flag +1 
 Yes I noticed. I was able to integrate it into my SDK and successfully requested my Cashu-address balance using nsec.app 👏 great work btw…

However request-id is the random string from my request, not an event id right?

I want to build a set of asynchronous functions that send a request and then resolve once I get the response, but from what I can tell there is no way to subscribe to ONLY the response, because the response does not include the requests e-tag, right?
▲ ▼
brugeman | 9 months ago (raw) | root | parent | reply | flag +1 
 Right, request-id is random string, not the request event id.
▲ ▼
Egge | 9 months ago (raw) | root | parent | reply | flag 
 Not including the requests e-tag in the response is a huge missed opportunity IMO
▲ ▼
brugeman | 9 months ago (raw) | root | parent | reply | flag 
 You can chime in to this huge rewrite if you're in for some arguing 😉 
https://github.com/nostr-protocol/nips/pull/1047
▲ ▼
brugeman | 9 months ago (raw) | root | parent | reply | flag 
 The only justification I have to omit e-tag is to decrease the metadata leakage (this is reply to this req), although it's a very weak argument.