logged out clients will be anonymous viewers to the keyserver, so we'll need a username and password to attest that the user is who they say they are
Details
tested in following diff by calling the responder and then using the message from the keyserver to successfully authenticate with identity service
Diff Detail
- Repository
- rCOMM Comm
- Lint
Lint Not Applicable - Unit
Tests Not Applicable
Event Timeline
lib/actions/user-actions.js | ||
---|---|---|
215 ↗ | (On Diff #40812) | this was timing out pretty consistently on staging so i increased the timeout threshold |
keyserver/src/endpoints.js | ||
---|---|---|
463 ↗ | (On Diff #40812) | It looks like this might cause all callsites to this endpoint on existing clients to fail. Are any existing clients using this endpoint? |
keyserver/src/responders/user-responders.js | ||
964–967 ↗ | (On Diff #40812) | Reduce indentation on these three lines |
979 ↗ | (On Diff #40812) | If !userRow.hash, should we throw invalid_parameters instead, since this account wasn't required originally with the authoritative keyserver, and can't be reserved? |
lib/actions/user-actions.js | ||
215 ↗ | (On Diff #40812) | Should we increase it further? Isn't our default timeout longer than this? Should we just remove this override? |
225 ↗ | (On Diff #40812) | We could probably avoid spreading here if we wrap the declaration with Object.freeze, but in my comment above I suggest removing this override entirely |
keyserver/src/endpoints.js | ||
---|---|---|
463 ↗ | (On Diff #40812) | no there are no existing clients using this endpoint |
keyserver/src/responders/user-responders.js | ||
979 ↗ | (On Diff #40812) | that makes sense |
lib/actions/user-actions.js | ||
215 ↗ | (On Diff #40812) | yeah it can just be removed. default timeout is 10 seconds i believe |