Similar to #314, we might need a "repository error" event.
In the real world, things might go wrong in the repository. Clients should at least have the chance to inform the user about it.
The repository error event would carry a human-readable message, and an error code.
The error code would be an undefined string. Rationale: There are too many things that can go wrong for us to specify, but we also don't want to resort to parsing the human-readable message if our client has some specific means to handle specific kinds of issues.
This could be the same as the no-op event from #314.
Pro:
Less events
Con:
So far, we tend to have lots of very fine-grained events to not muddle with semantics. In this sense, No-op and error reporting are different.
Similar to #314, we might need a "repository error" event.
In the real world, things might go wrong in the repository. Clients should at least have the chance to inform the user about it.
The repository error event would carry a human-readable message, and an error code. The error code would be an undefined string. Rationale: There are too many things that can go wrong for us to specify, but we also don't want to resort to parsing the human-readable message if our client has some specific means to handle specific kinds of issues.
This could be the same as the no-op event from #314.
Pro:
Con: