[FIXED] JetStream: Fail Ack() (and the likes) for AckNone consumer

[kozlovic]

Resolves #1027

Signed-off-by: Ivan Kozlovic ivan@synadia.com

[kozlovic]

@mprimi I think that this would have let you know that you should not have been calling Ack[Sync]()for a AckNone ack policy (the OrderedConsumer is creating a consumer with AckNone).

[coveralls]

Coverage decreased (-0.1%) to 85.599% when pulling c040b4f on fix_1027 into cc189da on main.

[wallyqs]

LGTM

[derekcollison]

What is the problem we are trying to solve? The server can handle an ack for an AckNone consumer.

[kozlovic]

What is the problem we are trying to solve? The server can handle an ack for an AckNone consumer.

Two problems:

• Sending unnecessary ACKs to the server(s)
• Having users incorrectly Ack'ing and trying again and again after getting an "unexpected" error. The context was that Marco had a test using OrderedConsumer and was calling AckSync(), which was working fine until in some cases, when bringing the cluster down and back again, the AckSync() would start to return an error. He was told to ack again in case of error, well, in that case the ack would always fail because the ordered consumer had reset, so the message's ack subject on hand was no longer valid and not known by any of the server, so AckSync() would return an error such as "no responders available".

[derekcollison]

LGTM