ci: Improve peer logging to help identify race #93
Conversation
Codecov Report
@@ Coverage Diff @@
## main #93 +/- ##
==========================================
+ Coverage 68.87% 71.87% +2.99%
==========================================
Files 85 89 +4
Lines 3396 3612 +216
Branches 55 55
==========================================
+ Hits 2339 2596 +257
+ Misses 860 793 -67
- Partials 197 223 +26
Continue to review full report at Codecov.
|
| c.opts.Logger.Debug(context.Background(), "ice gathering state updated", | ||
| slog.F("state", iceGatherState)) |
There was a problem hiding this comment.
Thanks for adding this additional logging!
| c.opts.Logger.Debug(context.Background(), "flushing remote candidate") | ||
| hash := sha256.Sum224([]byte(pendingCandidate.Candidate)) | ||
| c.opts.Logger.Debug(context.Background(), "flushing buffered remote candidate", | ||
| slog.F("hash", hash), |
There was a problem hiding this comment.
Neat idea to log this, so we can track if a particular ICE candidate made it across
There was a problem hiding this comment.
Yeah, I'm concerned that somehow the ICE candidate isn't getting across the pipe, so at least now we'll know whether it does or not!
There was a problem hiding this comment.
Speaking of that - we could also add logging when we set up the channels in the peer/conn_test.go, too!
case c := <-channel1.LocalCandidate():
_ = channel2.AddRemoteCandidate(c)Line 304 in f9e594f
And it might make sense to assert no error on that AddRemoteCandidate call? I don't think it should in the local network case (and if it does, might give us extra clues)
No description provided.