Pin the maiden rule to its vocabulary, and open the 2.1 ledger

AGENTS.md

@@ -88,9 +88,14 @@ uv run sphinx-build -b html docs dist/docs
 #    bare compare.py measures against two minors back while reporting the
 #    previous one, and _allowlist_for hard-errors on the missing file.
 #    tests/v2/test_regex_sync.py sweeps every expected_since_*.toml (#333), so
-#    the new ledger's hand copies of _SCRIPT_RANGES and the honorific
-#    vocabulary are checked from the day the file lands -- but it enrolls
-#    itself in neither roster, and both failures are loud, not silent:
+#    the new ledger's hand copies of _SCRIPT_RANGES and of the honorific and
+#    Latin vocabularies are checked from the day the file lands. All three
+#    rosters find copies by their SYNTAX -- a span class or an alternation --
+#    so a copy spelled some other way is not undeclared but unseen. What
+#    covers those is _CORPUS_CLAIMS, which records what every rule claims
+#    -- its regex's corpus reach, its roles, and which names -- and so
+#    needs no notion of how a copy is spelled. Enrol the new ledger in
+#    _CORPUS_CLAIMS always, and in the others where they apply:
 #      - _SPAN_BEARING_RULES: add the filename, mapped to the set of issue
 #        tags whose rules carry a script-span class (empty set if none).
 #      - _HONORIFIC_SOURCES: if the ledger has a CJK honorific rule, add a
@@ -99,6 +104,14 @@ uv run sphinx-build -b html docs dist/docs
 #        that issue. A retroactive ledger can repeat an older one's rule
 #        verbatim (fix(#271/#272/#298) is in both today), and every rule
 #        must match exactly one key.
+#      - _CORPUS_CLAIMS: REQUIRED, not conditional -- a ledger with no
+#        entry hard-fails. Add the filename mapped to {} while the ledger
+#        is empty, then one _Claim per rule as rules land. The test prints
+#        the values to record.
+#      - _LATIN_ALTERNATION_SOURCES: same, for a rule copying a Latin
+#        vocabulary (maiden markers, ambiguous acronyms). An alternation
+#        matching no key fails as undeclared -- add it, or record it in
+#        _NOT_A_VOCABULARY_COPY if it copies nothing.
 ```
 
 Enable debug logging to see the parser's internal decisions:

tests/v2/test_differential.py

@@ -298,15 +298,56 @@ def test_validate_rules_rejects_a_rule_that_would_silently_widen(
         compare.validate_rules([rule], "expected_since_2.0.0.toml")
 
 
+def test_default_baseline_has_a_ledger_and_nothing_else_in_it() -> None:
+    """Two facts about the OPEN cycle's ledger, which the carve-out
+    below leans on and neither of us should have to derive.
+
+    It must exist: _allowlist_for treats a missing ledger as a hard
+    error, so a DEFAULT_BASELINE with no file makes a bare compare.py
+    run abort -- and it would also make the carve-out inert, since it
+    would name a file nothing iterates over.
+
+    And it must define nothing at the top level except `change`. This is
+    the one ledger allowed to be empty, so a mistyped table header --
+    `[[changes]]`, `[[rules]]` -- reads as a legitimately empty open
+    cycle everywhere instead of as a broken file: every sweep gets zero
+    rules and passes, while the author believes they shipped a rule.
+    The other ledgers are protected by having to be non-empty; this one
+    needs saying out loud.
+    """
+    import tomllib
+    open_cycle = _TOOLS / f"expected_since_{compare.DEFAULT_BASELINE}.toml"
+    assert open_cycle.exists(), (
+        f"DEFAULT_BASELINE is {compare.DEFAULT_BASELINE!r} but "
+        f"{open_cycle.name} does not exist; a bare compare.py run would "
+        f"hard-error, and the empty-ledger carve-out would be inert")
+    keys = set(tomllib.loads(open_cycle.read_text(encoding="utf-8")))
+    assert keys <= {"change"}, (
+        f"{open_cycle.name} defines {sorted(keys - {'change'})} at the top "
+        f"level. Only `change` is read, so anything else is a typo that "
+        f"would read as an empty ledger rather than as a broken one")
+
+
 def test_validate_rules_accepts_the_shipped_ledgers() -> None:
-    """The guards above must not be so strict they reject real rules."""
+    """The guards above must not be so strict they reject real rules.
+
+    Every ledger but one must carry rules. The exception is the OPEN
+    cycle's -- the one DEFAULT_BASELINE names -- which is created the day
+    its baseline is released and is legitimately empty until that
+    cycle's first behavior change lands. An older ledger is history, and
+    history is not empty, so emptying one is still a mistake this
+    catches.
+    """
     import tomllib
+    open_cycle = f"expected_since_{compare.DEFAULT_BASELINE}.toml"
     ledgers = sorted(_TOOLS.glob("expected_since_*.toml"))
     assert ledgers, "no ledgers found; this test would pass vacuously"
     for ledger in ledgers:
         rules = tomllib.loads(
             ledger.read_text(encoding="utf-8")).get("change", [])
-        assert rules, f"{ledger.name} has no [[change]] rules"
+        assert rules or ledger.name == open_cycle, (
+            f"{ledger.name} has no [[change]] rules, and it is not the "
+            f"open cycle's ledger ({open_cycle})")
         compare.validate_rules(rules, ledger.name)