Conversation
Conflicts: tests/msc2716_test.go
| }) | ||
|
|
||
| t.Run("Historical messages are visible when joining on federated server - auto-generated base insertion event", func(t *testing.T) { | ||
| t.Skip("Skipping until federation is implemented") |
There was a problem hiding this comment.
Removing as matrix-org/synapse#10245 is now merged
Synapse support was added in matrix-org/synapse#10498
| }) | ||
|
|
||
| t.Run("When messages have already been scrolled back through, new historical messages are visible in next scroll back on federated server", func(t *testing.T) { | ||
| t.Skip("Skipping until federation is implemented") |
There was a problem hiding this comment.
Removing as matrix-org/synapse#10498 is now merged
Synapse change added in matrix-org/synapse#10552
| }) | ||
| }) | ||
|
|
||
| t.Run("Batch send endpoint only returns state events that we passed in via state_events_at_start", func(t *testing.T) { |
There was a problem hiding this comment.
Test added to verify matrix-org/synapse#10552
tests/msc2716_test.go
Outdated
| // We only expect 1 state event to be returned because we only passed in 1 | ||
| // event into `?state_events_at_start` | ||
| if len(stateEventIDs) != 1 { | ||
| t.Fatalf("Expected only 1 state event to be returned but received %d: %s", len(stateEventIDs), stateEventIDs) |
There was a problem hiding this comment.
| t.Fatalf("Expected only 1 state event to be returned but received %d: %s", len(stateEventIDs), stateEventIDs) | |
| t.Fatalf("Expected only 1 state event to be returned but received %d: %v", len(stateEventIDs), stateEventIDs) |
There was a problem hiding this comment.
For my own reference, what's the benefit of using %v vs %s here?
They both appear like:
Expected only 1 state event to be returned but received 1: [$EVEa7PzafkdmDIu4kEYwlqJFSSKV_pj5h5xYfRspFs4]
| }, | ||
| } | ||
| // We can't use as.SendEventSynced(...) because application services can't use the /sync API | ||
| markerSendRes := as.MustDoFunc(t, "PUT", []string{"_matrix", "client", "r0", "rooms", roomID, "send", markerEvent.Type, "txn-m123"}, client.WithJSONBody(t, markerEvent.Content)) |
There was a problem hiding this comment.
Is it a problem that the txn id never changes?
There was a problem hiding this comment.
Since the homeserver is new every run, I think it's fine. I did notice that txn-m123 was re-used once, so I updated to something different.
Might be nice to add a helper function to client to get a random unique one easily (using txnId in there). I added an internal one in #192
|
Thanks for the review @kegsay and catching the Python patterns that slipped in 🐆 |
MSC2716 federation changes
batchSendHistoricalMessagesto be able to compose state and message eventsstate_eventsthat we passed in via?state_events_at_start, Only return state events that the AS passed in viastate_events_at_start(MSC2716) synapse#10552Dev notes
Todo
state_events_beforeincludes invitations and join eventsRedacting an insertion, chunk, marker so remove it from our new tracking tablesnext_chunk_idwill reject. ((room_id, next_chunk_id)should be unique))