Add Hypothesis property-based tests for substitution logic

[adamtheturtle]

Summary

• Add hypothesis dependency for property-based testing
• Add 8 property-based tests covering _apply_substitutions and flag behavior:
  • Empty substitutions is identity
  • All placeholder occurrences are replaced
  • Text without placeholders is unchanged
  • No leftover placeholders after substitution
  • MyST {{ }} delimiters work correctly
  • Multiple delimiter pairs work in one call
  • :nosubstitutions: prevents replacement regardless of config
  • :substitutions: forces replacement regardless of config

Test plan

• All 47 tests pass (8 new + 39 existing)
• 100% code coverage maintained
• All pre-commit hooks pass (mypy, pyright, ruff, interrogate, etc.)

Closes #1423

🤖 Generated with Claude Code

---

Note

Low Risk
Low risk: only adds a dev-only testing dependency and new property-based tests; no runtime code or substitution logic is modified.

Overview
Adds hypothesis to the dev dependency set and introduces a new tests/test_hypothesis.py suite of property-based tests.

The new tests stress _apply_substitutions across randomized inputs (identity cases, full replacement, MyST {{ }} support, and mixed delimiter pairs) and verify :nosubstitutions:/:substitutions: flags override substitutions_default_enabled when building a minimal Sphinx project.

^{Written by Cursor Bugbot for commit 5821c2c. This will update automatically on new commits. Configure here.}

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

[cursor]

Missing deadline=None causes flaky Sphinx build tests

Medium Severity

The @settings for both Sphinx build tests suppresses HealthCheck.function_scoped_fixture but doesn't set deadline=None. Hypothesis defaults to a 200ms deadline per example. Each example creates a temp directory, writes files, initializes a SphinxTestApp, runs a full build, and reads HTML output — very likely exceeding 200ms, especially in resource-constrained CI environments. This can cause intermittent DeadlineExceeded failures.

Additional Locations (1)

• tests/test_hypothesis.py#L237-L240