Container sessions updates

[carpawell]

There were some changes in the node (nspcc-dev/neofs-node#2727) that changed the behavior (it was already mentioned in #718). However, there also was some bug about token's wildcard field that was fixed in nspcc-dev/neofs-node#2741.

Current Behavior

#726 changed the error message to expect to the "wrong container" version. After the fix, it should not be possible to get such a message if wildcard field is provided (the field means you may not specify a container at all, so "wrong container" does not relate to the situation).

Expected Behavior

If tests do correct things, no error is expected. I have not seen all the tests but every error case should have some readable message about the error. No "wrong container" for the wildcard tokens caused by the bug fixed in nspcc-dev/neofs-node#2741 and also no "timeout" error as it was before since the SN now validates requests before it sends them to the IR. Also, there is exactly one straight happy path (correct container, correct token signer, and so on) and i think any different token usages should be also tested. If anything that can be tested is already tested, it is great then. But if there are some untested cases, i would like them to be tested.

Possible Solution

1. Run some tests with 0.40.1+ version, see the message. If it is correct (it should describe the "wrong" things the tests sent), adapt the test's expected message. If it is incorrect, create an issue about it and we will fix it.
2. See if anything else can be tested but has not been tested yet.

Steps to Reproduce

0.40.1. I believe that is all you need.

Context

If any message is not fully clear, we may adapt it, just create an issue. Also, be ready that some wrong cases may not be reproducible cause tests use neofs-cli that may fix some token issues inside.

Regression

#726. Not sure it had to be merged. It does not adapt the tests, it just fixed (i mean the bug in the node is now "tested and expected") the bug we found.