docs: 3.3.0 release notes for the merged fixes

[fgmacedo]

Why

The 3.3.0 notes on main had only the extension keys entry. Two fixes already merged for 3.3.0 change behavior a user may notice when upgrading, and later 3.3.0 PRs need a skeleton to add to.

Refs #692
Refs #713

What changes

docs/releases/3.3.0.md only:

• Skeleton matching 3.2.0 and 3.2.2: *Unreleased*, a comment placeholder for the maintainer's intro paragraph, then What's new, Behavior changes to check when upgrading, Deprecations (placeholder: deprecations are removed in 4.0, nothing deprecated yet) and Bug fixes.
• What's new: PR link on the extension keys entry, and a short entry on the UML test cases based on PSSM, linking to the contributing section (which links to tests/uml/pssm/README.md). No conformance claim, no spec text quoted.
• Bug fixes: history default entry (fix: enter history states by default without running the incoming effect twice #693, with a doctest showing first-visit entry and the incoming effect running once) and errors raised while an invocation processes done.invoke (fix: report errors raised while an invocation processes its done.invoke #698).
• Behavior changes: five bullets, two from fix: enter history states by default without running the incoming effect twice #693 (effect no longer doubled, empty history enters the compound's initial state) and three from fix: report errors raised while an invocation processes its done.invoke #698 (StateMachine errors reported via threading.excepthook / loop exception handler, TransitionNotAllowed without a done_invoke transition, on_finalize errors through the engine's error handling).

Evidence

• timeout 180 uv run pytest -n 4: 2973 passed, 26 skipped, 126 xfailed; the new doctest passes.
• uv run sphinx-build -W docs docs/_build/html: 48 warnings, all pre-existing in older release notes; none from 3.3.0.md.
• The doctest example run against 3.2.2 (d5282e0) gives an empty configuration after the first resume, as the text says.
• Factual review by a fresh subagent against gh pr diff 693 and gh pr diff 698, with scripts run on 3.2.2 and on main: two material errors found and fixed (the on_finalize bullet said a StateMachine error "still propagates", but when the invocation drives done.invoke it used to become error.execution; the first done.invoke bullet lacked the StateMachine scope). Three wording fixes applied (doubled effect only with a recorded value or default transition, run() scoping, lowered PSSM cases listed apart).

Not verified

• The intro paragraph is left for the maintainer.
• The ReadTheDocs build of this branch was not checked; only the local Sphinx build.

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.00%. Comparing base (75c9181) to head (a8c6af3).

Additional details and impacted files

@@            Coverage Diff            @@
##              main      #725   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files           53        53           
  Lines         5900      5900           
  Branches       939       939           
=========================================
  Hits          5900      5900           

Flag	Coverage Δ
unittests	100.00% <ø> (ø)

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.