#481: preserve true bool PMP defaults

[GHX5T-SOL]

Summary:

• Preserve boolean PMP defaults that are already coerced before final type conversion.
• Add a regression with a PMP XML bool default of true and no factbase override.

Closes #481

Validation:

• Pre-fix probe with in-memory PMP XML default true returned value=false.
• Post-fix probe with the same fixture returned value=true.
• bundle exec ruby -Itest -e "ARGV << "--no-cov"; require "./test/fbe/test_pmp"; ARGV.delete("--no-cov")" -> 9 tests, 17 assertions, 0 failures, 0 errors, 0 skips.
• bundle exec rake test -> 284 tests, 804 assertions, 0 failures, 0 errors, 9 skips.
• bundle exec rubocop -> 65 files inspected, no offenses detected.
• git diff --cached --check -> passed.
• gitleaks protect --staged --no-banner --redact --verbose -> no leaks found.

No secrets were added.

[akmhatey-ai]

Checked the current head b85f0855194d6fad4c307eccf48d41e6963f71b1 against #481.

The one-line change is scoped to the failing path: an XML PMP bool default is already coerced to Ruby true, and the final bool conversion now accepts it via result.to_s == 'true'. Existing string overrides still behave as before ('true' stays true, 'false' stays false), and the new regression test covers the no-factbase-override/default-true case from the issue.

Validation I ran locally:

• bundle install with isolated BUNDLE_PATH passed.
• Focused PMP tests passed: 9 tests, 17 assertions, 0 failures/errors/skips.
• bundle exec rubocop lib/fbe/pmp.rb test/fbe/test_pmp.rb --format simple passed.
• Full bundle exec rubocop --format simple passed: 65 files, no offenses.
• git diff --check origin/master...HEAD passed.
• Diff-only gitleaks stdin --no-banner --redact --exit-code 1 passed with no findings.

One local limitation: full bundle exec rake test on my Windows checkout hit unrelated environment errors around SQLite temp DB cleanup permissions and /bin/bash not existing. The touched TestPmp coverage passes locally, and the hosted Ubuntu/macOS rake checks are green, so I do not see that as caused by this PR.

[0crat]

@akmhatey-ai Hey! Nice work on that review 👍 You earned +10 points this time: started with the standard +18 base points, but had to deduct 8 since there weren't any comments posted during the review. Next time, try leaving a few comments to avoid that deduction and boost your score even higher! Your running total is now +119 - keep it up and don't forget to check your Zerocracy account too.

[yegor256]

Please fix the merge conflicts so this pull request can be merged.

[morphqdd]

Correct fix for #481 — result == 'true' returns false when result is already the boolean true (true == 'true' is false), so a bool PMP with a true default was silently flipped to false. result.to_s == 'true' handles both the string-from-XML and the already-boolean cases. The test with a <default>true</default> bool covers it. LGTM.

[morphqdd]

Good — .to_s normalizes both paths: 'true' (parsed from XML) and a boolean true (when the default is already typed) both yield truthy, while false/'false'/anything else stays false. Handles the exact case #481 describes.

[0crat]

@morphqdd Thanks for the review! You've earned +13 points for this: +18 as a basis; -5 for very few (1) comments. Your running score is +778; don't forget to check your Zerocracy account too).