Reinstate pool tests

[Ben-PH]

• document all added functions
• try in sandbox /simulation/labnet
• unit tests on the added/changed features
  • make tests compile
  • make tests pass
• add logs allowing easy debugging in case the changes caused problems
• if the API has changed, update the API specification

[Ben-PH]

still need to refactor operation_pool_tests.rs::test_pool to use pool_test wrapper

EDIT: might not end up using it...

[Ben-PH]

67d2921 adds some dbg!s, and this is the typical output:

(will get through a few threads from time to time...)

     Running unittests src/lib.rs (target/debug/deps/massa_pool_worker-c92c862378c0b4ac)

running 1 test
[massa-pool-worker/src/tests/operation_pool_tests.rs:129] thread = 0
[massa-pool-worker/src/operation_pool.rs:224] &slot = Slot {
    period: 31,
    thread: 0,
}
[massa-execution-exports/src/test_exports/mock.rs:217] "never seen this not err..." = "never seen this not err..."
[massa-execution-exports/src/test_exports/mock.rs:218] response_rx.recv_timeout(Duration::from_millis(100)) = Err(
    Timeout,
)
thread 'tests::operation_pool_tests::test_pool' panicked at 'called `Result::unwrap()` on an `Err` value: Timeout', massa-execution-exports/src/test_exports/mock.rs:218:68

Still need to work out the source of the problem...

[Ben-PH]

At this point, two of the methods are unignored, but the third one is not clear for me yet. @AurelienFT what do you think of merging these two so far, and do the rest in a follow up PR. It at least makes these two available in the meantime.

[Ben-PH]

Turns out needed to forward on the messages, and provide a balance that allowed the operations to occur.

Still has an issue with storage cleanup.

[Ben-PH]

test is flakey: run this inside massa-pool-worker to (eventually) get a fail;

while cargo test test_pool; do :; done

[AurelienFT]

I didn't follow well everything that has been going on but to give you maybe some context that might help you. We always run the test at root with feature testing (to have exports only when test cfg (no other way with cargo yet)).

[Ben-PH]

updated script:

COUNTER = 0;
while cargo test test_pool -F testing --release ; do
    COUNTER=$[$COUNTER +1]
    :
done

echo $COUNTER;

@Eitu33 had a look at it with me, and it seems that this comes down to concurrent nature of the way the tests are setup?

The means to fix it is similar to the other two tests: Use the mock-receiver to intercept requests for data, and reply with hardcoded/intercepted data.

With that in mind, I'll do the final tidy up then mark for review.

[Ben-PH]

Was it the case in the past that you could create multiple instances of storage without panic!ing? That could explain the need for this change.