Add scripted return status support for TestNode mocks

[pepeRossRobotics]

This adds support for computing a TestNode return status from a script, so mock behavior can depend on blackboard state without adding extra nodes to the tree.

It also updates the substitution tests and tutorial examples to cover both the observable and strict-failure mock flows.

Addresses #995

[facontidavide]

Thanks for this contribution — the feature itself is welcome and addresses #995 well, the test coverage is good, and I verified locally that everything builds and passes.

However, there are a few things that need to be addressed before this can be merged:

1. We can not accept API changes: return_status must remain a plain NodeStatus.

Changing TestNodeConfig::return_status from NodeStatus to std::optional<NodeStatus> is a source-compatibility break for any downstream code that reads the field directly (e.g. NodeStatus s = cfg.return_status;, toStr(cfg.return_status), or a switch on it).

The optional isn't actually needed for the feature: since return_status_script takes precedence over the fixed status anyway (as your own documentation states), you can keep

NodeStatus return_status = NodeStatus::SUCCESS;

unchanged. Concretely:

• drop the .reset() calls from the examples and tests;
• in loadSubstitutionRuleFromJSON, use a local "was return_status present in the JSON" flag for the "must contain return_status or return_status_script" validation, instead of resetting the member.

That makes the whole PR purely additive.

2. Misleading error message for out-of-range script results.

validateTestNodeStatus() hardcodes "resolved to invalid status IDLE". A script that evaluates to e.g. 7 reports "IDLE", which would send users debugging in the wrong direction. Please include the actual offending value in the message (the raw integer is fine, since toStr can't render invalid enum values).

3. Unrelated change in tests/gtest_wakeup.cpp.

The system_clock → steady_clock switch and the relaxed assertion bound are reasonable on their own, but they are unrelated to this feature and not mentioned in the PR description. Please move them to a separate PR (or at minimum call them out explicitly in the description).

4. Please leave the existing tutorial t15_nodes_mocking.cpp untouched.

Tutorial 15 corresponds to a published page on behaviortree.dev; restructuring its tree (the added Fallback, the mock flag) makes the code diverge from the website text. Keep the original example as it is and demonstrate the scripted return status only in your new t15_nodes_mocking_strict_failure.cpp (or a dedicated new example).

[pepeRossRobotics]

Comments addressed, also opened #1164