docs(invarianter): the five other types, as the outboxes actually lie [skip-docs]

Third writing of the same sentence, third time untrue. Measured in this working tree, by the arm
committed red before this one: FOUR of the five types outside the measured seven do not sit in an
outbox that also holds a coverage -- and for `plan-review` that says nothing at all, because the
type has 0 artefacts in the repo. `multibase` DOES sit there, in four outboxes
(`p17b-multibase/`, `p20-stress/`, `p21-stress/`, `p22-stress/`, all `lindaas`), and each of those
four holds TWO coverage files. That is why they fall outside the "exactly one coverage" rule the
union of seven is counted over -- and the same fact is the proof that seven is a FLOOR and not a
ceiling, which is the argument the false sentence was trying to make.

The comment above `_RUN_LEVEL_TYPES` carried the same claim in English and is corrected with it.
The arm now counts the outboxes and holds the ledger's row to the count, so a fourth writing is a
failing test rather than a review finding.

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

docs/invarianter.md

@@ -3146,9 +3146,16 @@
   igjen SJU artefakt-typer — ikke «en ekte kjøring», som lovet mer enn målingen (RETTET 19.09:
   `run.py` kaller TI av `outbox.py`s ti skrivere, så `exploration`, `prepass`, `multibase`,
   `plan-review` og `proposal-reviews` er fem typer til, skrevet når flaggene som lager dem er
-  gitt; ingen av dem ligger i dag i en utboks som også har coverage. Sju er gulvet målingen gir,
-  ikke et tak — og binderen kopierer på glob, ikke på denne lista, så en type utenfor den bæres
-  uansett). Talt over de fire arkiverte kjøringene sjekkpunktet
+  gitt. FIRE av de fem ligger ikke i en utboks som også har coverage — og for `plan-review` sier
+  det ingenting, for den typen finnes ikke som artefakt i repoet i det hele tatt (0 filer).
+  `multibase` GJØR det: fire utbokser (`p17b-multibase/`, `p20-stress/`, `p21-stress/`,
+  `p22-stress/`, alle `lindaas`) har både multibase og coverage, og hver av dem har to
+  coverage-filer. Det er derfor de faller utenfor «nøyaktig én coverage»-regelen unionen telles
+  over — og samtidig beviset på at sju er GULVET målingen gir, ikke et tak. Binderen kopierer
+  dessuten på glob, ikke på denne lista, så en type utenfor den bæres uansett). **Setningen er
+  skrevet om TRE ganger og var usann hver gang; den er nå pinnet** av en arm som teller utboksene
+  og krever at raden sier det tellingen sier
+  (`test_the_ledgers_claim_about_the_five_other_types_is_what_the_repo_measures`). Talt over de fire arkiverte kjøringene sjekkpunktet
   leste (`tunnel-hauglia-2027` -04/-06/-07/-08; hver `<run_id>-<rest>.json` typet som
   `proposal`/`outcome` når `<rest>` ender der, ellers `<rest>` selv): `-06`, `-07` og `-08` har
   alle sju (15 filer hver), `-04` har seks (12 filer, ingen `parse-failures` — den skrives bare når

tests/test_round_builder_loadbearing.py

@@ -177,8 +177,11 @@ _STAGE_SENTENCES: dict[str, tuple[str, str]] = {
 #:
 #: The wider claim would be FALSE. ``run.py`` calls TEN of ``outbox.py``'s ten writers —
 #: ``exploration``, ``prepass``, ``multibase``, ``plan-review`` and ``proposal-reviews`` are five
-#: more types, written when the flags that produce them are given, and none of them happens to
-#: sit in an outbox that also has a coverage today. Seven is what the runs measured left; it is
+#: more types, written when the flags that produce them are given. FOUR of those five do not sit
+#: in an outbox that also holds a coverage; ``multibase`` does, in four outboxes that hold TWO
+#: coverage files each, which is why they fall outside the «exactly one coverage» rule this union
+#: is counted over — and why seven is a FLOOR. Counted by the arm at the end of this file, which
+#: also holds the ledger's sentence to the count. Seven is what the runs measured left; it is
 #: not a ceiling. The builder copies on a glob and not on this list, so a type outside it is
 #: carried anyway — this table is what the FIXTURE must be able to witness, nothing more.
 #:
@@ -1289,9 +1292,14 @@ _TYPES_OUTSIDE_THE_SEVEN = (
     "proposal-reviews",
 )
 
-#: The false clause, verbatim, in the form it was written THREE times. Pinned as a needle so the
-#: fourth writing of it is a failing test rather than a review finding.
-_THE_FALSE_CLAUSE = "ingen av dem ligger i dag i en utboks som også har coverage"
+#: The false clause, verbatim, in BOTH languages it has been written in — Norwegian in the ledger
+#: row, English in this file's own comment above ``_RUN_LEVEL_TYPES``. Pinned as needles so a
+#: fourth writing is a failing arm rather than a review finding. Only the LEDGER is scanned for
+#: them: this file cannot scan itself for a string it must also contain to scan for it.
+_THE_FALSE_CLAUSES = (
+    "ingen av dem ligger i dag i en utboks som også har coverage",
+    "none of them happens to sit in an outbox that also has a coverage",
+)
 
 _NUMERAL = {0: "null", 1: "én", 2: "to", 3: "tre", 4: "fire", 5: "fem", 6: "seks"}
 
@@ -1343,10 +1351,10 @@ def test_the_ledgers_claim_about_the_five_other_types_is_what_the_repo_measures(
         if all(f"`{kind}`" in block for kind in _TYPES_OUTSIDE_THE_SEVEN)
     ]
     assert len(row) == 1, f"the seven-type row is not one block: {len(row)}"
-    sentence = row[0]
-    assert _THE_FALSE_CLAUSE not in sentence, "the false clause was written a fourth time"
-    assert _THE_FALSE_CLAUSE not in Path(__file__).read_text(encoding="utf-8").replace(
-        f'_THE_FALSE_CLAUSE = "{_THE_FALSE_CLAUSE}"', ""
-    ), "the false clause is back in this file's own comment"
+    # Markdown wraps, so the needles are read against the row as ONE line: where the sentence
+    # happens to break is not a claim about anything.
+    sentence = " ".join(row[0].split())
+    for clause in _THE_FALSE_CLAUSES:
+        assert clause not in sentence, f"the false clause was written a fourth time: {clause}"
     assert f"{_NUMERAL[len(multibase_outboxes)]} utbokser" in sentence, sentence[-400:]
     assert f"{_NUMERAL[2]} coverage-filer" in sentence, sentence[-400:]