Correct two comments this pass got wrong itself

Both describe the branch's own mechanism, and both were checked
against origin/master afterwards rather than before.

The maiden clause said 씨 "stayed in maiden 박씨 on the unscoped
segments[0]". It did stay there, but not for that reason: before #312
the family-comma gate sat ABOVE the peel and the stage returned
without running it at all.

The comma-and-dot test comment said the 间隔号 gates the surname split
and the family comma "gates it too". Both stand over the split, but
the family comma returns first, so the dot's gate is never reached on
this input -- which is the only reason the test is a composition and
not two facts.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

Author: derek73

nameparser/_pipeline/_script_segment.py

@@ -349,8 +349,9 @@ def _peel_honorific_tail(state: ParseState) -> ParseState:
     regardless of which of her names it was glued to, and a
     name-final honorific is exactly what this peels. Since #312 that
     reach extends to FAMILY_COMMA along with the rest of the crossing
-    -- "김, 민준 née 박씨" now routes 씨 to suffix where on the
-    unscoped segments[0] it stayed in maiden "박씨".
+    -- "김, 민준 née 박씨" now routes 씨 to suffix, where before #312
+    the stage returned at the family-comma gate above the peel and
+    left it in maiden "박씨".
 
     The other direction is a LIMIT, stated rather than fixed: a maiden
     clause pushes the site off the person's own name, so "김민준씨 née

tests/v2/pipeline/test_script_segment.py

@@ -622,13 +622,13 @@ def test_an_empty_first_run_still_reaches_the_second() -> None:
 
 
 def test_the_peel_crosses_a_comma_and_a_dot_at_once() -> None:
-    # The two gates #312 made siblings, composed: the 间隔号 gates the
-    # surname split (威廉 stays whole -- 威 is a listed surname here,
-    # so without the dot it would divide) and the family comma gates
-    # it too, while the peel crosses both and reaches 太郎さん on the
-    # far side of the comma. Neither gate alone pins this: keying the
-    # cross-comma reach on interpunct-free input passes every other
-    # case in this file.
+    # The two gates #312 made siblings, composed. Both stand over the
+    # surname split here -- 威 is a listed surname in this lexicon, so
+    # bare 威廉 would divide 威 + 廉, and either gate alone stops that
+    # (the family comma reaches its return first) -- while the peel
+    # crosses both and reaches 太郎さん on the far side of the comma.
+    # Neither gate alone pins this: keying the cross-comma reach on
+    # interpunct-free input passes every other case in this file.
     lex = _LEX_TAILS.add(surnames={"威"})
     state = segment(tokenize(ParseState(
         original="威廉·莎士比亚, 太郎さん", lexicon=lex, policy=_HAN)))