OG PR is #600, recreated to ensure correct emails are being used for commits.

Received the following diagnostics while designing an application with postgres-nio 1.27.0...1.29.0:

00:56:35.221 [NotificationsListener.swift:517] - Notifications: Failed to LISTEN {error: "PSQLError(code: server, serverInfo: [sqlState: 57P01, file: be-secure.c, line: 209, message: terminating connection due to unexpected postmaster exit, routine: secure_read, localizedSeverity: FATAL, severity: FATAL])", topic: hawq_39}
00:56:35.222 [PostgresChannelHandler.swift:104] - Channel error caught. {psql_error: "I/O on closed channel", psql_connection_id: "1"}
00:56:35.222 [NotificationsListener.swift:286] - Notifications: Restarting subscription {topic: hawq_205}
00:56:35.222 [NotificationsListener.swift:148] - Notifications: Starting listening {psql.conn.id: "0", psql.conn.isClosed: true, topic: hawq_157}

SWIFT TASK CONTINUATION MISUSE: listen(_:) leaked its continuation without resuming it. This may cause tasks waiting on it to remain suspended forever.

And then the application deadlocks because listen() doesn't ever return.

Reproduction:

1. The code looks like that:

            try await self.postgresClient.withConnection { connection in
                148: log.debug("Notifications: Starting listening", metadata: [       
                    "topic": "\(topic)",                                         
                    "psql.conn.id": "\(connection.id)",                          
                    "psql.conn.isClosed": "\(connection.isClosed)"               
                ])                                                               
                sequence = try await connection.listen(topic.stringValue)        
                154: log.debug("Notifications: Now listening", metadata: [            
                    "topic": "\(topic)",                                          
                    "psql.conn.id": "\(connection.id)",                          
                    "psql.conn.isClosed": "\(connection.isClosed)"               
                ])
   

2. brew services restart postgresql

3. The error log similar to the log above gets generated, and the system goes into a deadlock.

Notice the log from line 148 is present, the connection appears to be CLOSED, and then the line 154 is unreachable. This might mean that listen() state machine is not correct on closed channels and result in deadlock.