When receiving messages in a chat room, Spark checks if the sender has voice, before displaying the message. This has at least two problems:
If the room is not moderated, having voice is of no interest here.
The responsibility of distributing messages should not be with the client, but with the server.
Effectively, this prevents users from seeing their own text in unmoderated rooms in which they have no voice.
Spark should simply not perform this check.