Ask the maiden take's numeral candidates the one-case fact (#731)

[derek73]

Closes #731.

M2's maiden take builds a clause-free view of the name from the clause's trailing candidates and lets tail_reading decide what the run takes. Its candidate test (_pieces.trailing_candidates) asked the numeral fork's shape test with no one-case fact, meaning to admit more than the fork as a safe superset. But a candidate is also cut out of the view, which changes the word in front of the numeral there:

input	before	after
john smith nee x x	suffix x, maiden x, reported	maiden x x, as JOHN SMITH NEE X X
john smith nee j x	suffix x, maiden j, reported	maiden j x, as JOHN SMITH NEE J X

Change. trailing_candidates takes a one-case fact, and the take passes the whole name's (one_case and is_one_case(<every token>)). The word in front of a candidate is always a clause word, which rules/T4 reads by the whole name's case. The first version passed the own-words fact, and the design-docs review found it moved MARY SMITH NEE j x (mixed case) to maiden j x; ee0962e fixes that.

Versions. 2.2.0 through 2.4.2 gave the old lowercase reading; 2.0.0 and 2.1.0 gave the one this restores. The #723 release bullet links #731.

Tests.

• The one-case invariant's grid gains {a} {b} nee {n} {n} and {a} {b} nee j {n}: ad486f6 disagreed on 36 of 4,752 pairs, this branch on none (new control: 36).
• A new invariant, test_a_clause_letter_beside_capitals_reads_as_beside_title_case, pairs own words in capitals with Title case, the clause letters lowercase in both: 0 field differences, 72 with the first version's fact (its control).
• Two case rows. No corpus name moved; the snapshot gains only the new T4 example.

A gap left open, on master too. The view itself still reads by the own words' case, so a lowercase clause numeral behind own words in capitals counts as a capital would. With two own words only the report differs (JOHN SMITH nee x x reports suffix-or-name, John Smith nee x x does not; 108 pairs). With one own word the fields differ too (SMITH NEE j x maiden j x, Smith Nee j x suffix x). Passing the whole name's fact to the view closes both, but it also moves S6's credential lean, which P3 asks of the own words: test_a_maiden_clause_does_not_change_how_a_trailing_word_reads fails on 270 of 20,412 pairs (J. nee Smith BA would read the credential where J. BA reads a name). Closing it needs the peel to take the numeral's case fact apart from the lean's. That is declined here with this evidence and recorded in decisions/T4; the new invariant's grid states it holds two own words only.

Docs. rules/T4 (statement clause, example, implemented:/interacts:), rules/M2 names S10 among its trailing rules, decisions/T4 and decisions/M2 entries (the candidate test's one stated exception to its superset contract), and the trailing_candidates and is_trailing_numeral_suffix docstrings.

Reviews. The design-docs review found the mixed-case regression, the superset wording, the version dates and M2's missing S10. The code review confirmed the fix (8,100 lowercase/capital moves on a ~1M-parse grid, no mixed-case moves, superset check holding with lowercase input) and found the one-word-head field gap and two stale docstrings. All are addressed in ee0962e and 5f989e2.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.04%. Comparing base (ad486f6) to head (5f989e2).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #732   +/-   ##
=======================================
  Coverage   99.04%   99.04%           
=======================================
  Files          46       46           
  Lines        4588     4589    +1     
=======================================
+ Hits         4544     4545    +1     
  Misses         44       44           

☔ 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.