core: apply 2-state transition pattern to TransferProcessManager

[ndr-brt]

What this PR changes/adds

Migrate all the state transition to the 2-state pattern as described in #825 for TransferProcessManager
(PRs for ContractNegotiationManagers are about to come)

Why it does that

Use the 3-state pattern could lead to problems, described in #825

Further notes

now all the states (except ERROR, CANCELLED, ENDED and STREAMING) have their process* method that permit them to be transitioned to another state.

the state REQUESTED_ACK has been removed, a new REQUESTING state has been introducted.
the state DEPROVISIONING_REQ has been removed.

currently, the deprovisioning method on the listeners is called by the processDeprovisioning, but it feels like it would be better to have it called by the DeprovisionRequestHandler

I had to implement a lease mechanism into InMemoryTransferProcessStore

I had to raise the provisionWait on AbstractQueueProvisioner to make DemoPushStreamTransferTest run on the CI, it was working locally but not on Actions. It was pretty flaky also before this PR. (e.g: https://github.com/eclipse-dataspaceconnector/DataSpaceConnector/runs/5478021065 on main branch)

Linked Issue(s)

#825 (not ready to be closed yet)

Checklist

• added appropriate tests?
• performed checkstyle check locally?
• added/updated copyright headers?
• documented public classes/methods?
• added/updated relevant documentation?
• added relevant details to the changelog? (skip with label no-changelog)
• linked GitHub project? (see the section on the right-hand side -->)

[ndr-brt]

@algattik the minikube test is failing, but it works locally, do you know what the cause could be?

[ndr-brt]

I think understood the problem, the FileTransferLocalSimulation does not wait for the file to be transferred, just for the transfer process to be in completed state (in the sample 04 there's no status checker so the manager puts the process in COMPLETED directly without checking if the file is transferred).
I added a StatusChecker to verify if the transfer is completed for real

[codecov-commenter]

Codecov Report

Merging #831 (42479aa) into main (b5a5d34) will increase coverage by 0.56%.
The diff coverage is 71.54%.

@@             Coverage Diff              @@
##               main     #831      +/-   ##
============================================
+ Coverage     60.07%   60.64%   +0.56%     
- Complexity     2262     2279      +17     
============================================
  Files           607      607              
  Lines         13071    13078       +7     
  Branches        912      912              
============================================
+ Hits           7853     7931      +78     
+ Misses         4792     4709      -83     
- Partials        426      438      +12     

Impacted Files	Coverage Δ
.../spi/transfer/observe/TransferProcessListener.java	0.00% <0.00%> (ø)
...ansfer/demo/protocols/stream/DemoTopicManager.java	71.11% <50.00%> (-2.23%)	⬇️
...fer/store/memory/InMemoryTransferProcessStore.java	96.66% <66.66%> (-1.55%)	⬇️
...sfer/core/transfer/TransferProcessManagerImpl.java	75.84% <71.42%> (+31.69%)	⬆️
...tor/spi/types/domain/transfer/TransferProcess.java	73.72% <77.77%> (+1.06%)	⬆️
...re/command/handlers/DeprovisionRequestHandler.java	100.00% <100.00%> (ø)
...mo/protocols/common/AbstractQueuedProvisioner.java	83.33% <100.00%> (ø)
...i/types/domain/transfer/TransferProcessStates.java	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 7fa65d3...42479aa. Read the comment docs.

[paullatzelsperger]

From your PR description:

currently, the provisioning method on the listeners is called by the processProvisioning, but it feels like it would be better to have it called by the DeprovisionRequestHandler

Is that a typo: provisioning should probably be deprovisioning, right? Assuming that's the case that would mean that the TransferProcessObservable "leaks" all the way through the ProvisionManager into the DeprovisionRequestHandler, which sounds awkward

[ndr-brt]

@paullatzelsperger yes, it was a typo, I fixed it and also the other remarks, thanks

[ndr-brt]

I had to implement a lease mechanism into the InMemoryTransferProcessStore, now everything work as expected, ready to review/merge

[ndr-brt]

ready to review/merge

[ndr-brt]

@paullatzelsperger @algattik is ok for you?

[ndr-brt]

@jimmarino can this be merged?