A failed resume incorrectly retries queued messages prior to reattachment

[paddybyers]

When handling a failed resume in accordance with RTN15c3 (https://github.com/ably/specification/blob/ee511f8a62e3e09cdd2ce50fadddb2d66f2ebaa3/textile/features.textile#RTN15c3), the ConnectionManager processes outstanding queued messages via `ConnectionManager.queuedMessages (

ably-java/lib/src/main/java/io/ably/lib/transport/ConnectionManager.java

Line 1215 in 07ad898

addPendingMessagesToQueuedMessages(true);

).

However, these messages instead need to be requeued at the Channel level, so that they are correctly processed relative to the re-attach that happens. (Currently, replayed presence messages are reattempted before the channel attach has completed.)

Normal Messages need to be removed from ConnectionManager.queuedMessages and ConnectionManager.pendingMessages and put back into ChannelBase.queuedMessages (

ably-java/lib/src/main/java/io/ably/lib/realtime/ChannelBase.java

Line 1270 in e57183f

queuedMessages = new ArrayList<QueuedMessage>();

)

Presence messages need to be removed from ConnectionManager.queuedMessages and ConnectionManager.pendingMessages and put back into Presence.pendingPresence (

ably-java/lib/src/main/java/io/ably/lib/realtime/Presence.java

Line 850 in e57183f

private final Map<String, QueuedPresence> pendingPresence = new HashMap<String, QueuedPresence>();

)

[sync-by-unito]

➤ Automation for Jira commented:

The link to the corresponding Jira issue is https://ably.atlassian.net/browse/SDK-3281