clients will call this RPC with a user ID and get in response a map of session initialization info (public keys, social proof, etc.) it needs, keyed by the device's ed25119 signing public key.
adding @ashoat to OK the API design
Details
Details
called this RPC from the keyserver and got back the test data i created in DDB. also tested the failure case with an invalid user ID and got back an exception in JS.
Diff Detail
Diff Detail
- Repository
- rCOMM Comm
- Branch
- identity
- Lint
No Lint Coverage - Unit
No Test Coverage
Event Timeline
| services/identity/src/service.rs | ||
|---|---|---|
| 286–289 | the DB client gives us back a HashMap<String, HashMap<String, String>>. We need to wrap the inner HashMap in a SessionInitializationInfo struct to satisfy the protobuf definition | |
| shared/protos/identity.proto | ||
| 88–90 | protobuf doesn't allow defining nested maps inline, but it's happy if we wrap the inner map in a message | |
| shared/protos/identity.proto | ||
|---|---|---|
| 89 | Can you add an inline comment explaining what the keys and values are? I think for the four keys we're starting with, @atul and @marcin have settled on a nested JSON format (see D6822). Perhaps we could have three keys here:
| |
| 196 | Can you add an inline comment explaining what it's keyed on? | |
| shared/protos/identity.proto | ||
|---|---|---|
| 90 ↗ | (On Diff #23208) | Might be worth explaining that this is where the signing ed25519 key is to be found, and that the signature part refers to primary (not notification) |
| shared/protos/identity.proto | ||
|---|---|---|
| 89 | Really not a fan of having a having a heterogenous collection of values in a hashmap. Would much rather see: message SessionInitializationInfo {
string payloadJSON = 1;
string payloadEd25519Signature = 2;
optional string socialProof = 3;
}It helps with compile time checking, bindings, and reasoning about the shape of your data. | |
Comment Actions
Yeah but it makes it much harder whenever we want to change it. Protobufs / gRPC don't handle versioning of schemas very well. @jon, do you have a good story on how we'll handle it if we need to add a new entry?
In general we need to land this ASAP, so would prefer to prioritize finishing up the monthly goal over further discussion here. Maybe we can create a stub task and discuss there once everything is landed?