§317 ended with "That is a five-line script, not an audit -- and it
would have found all four." No such script existed when I wrote that.

Built and measured:

  naive diff of constructed variants against each list
      -> 10 lists of 10 flagged. Almost every list is a legitimate
         subset, so it finds everything, which is finding nothing.

  the same grouped into families (control flow / binding / exit),
  flagging a list that covers PART of a family
      -> 3 of 10, and on the commit before #2875 it points at the exact
         `has_body` line. But all three of its hits on a clean tree are
         correct code: `StmtLocal | StmtAssign` collects NAMED bindings
         and `StmtExpr` has no name; the loop arm at 10979 excludes
         `StmtIf` on purpose.

  two of the four are not NodeKind lists at all: `compound_binop` maps
  operator strings and `expr_is_bool` is a match whose missing arm
  nobody enumerated.

So: 1 of 4, three false positives. `tri kinds drift` also finds 1 of 4
and costs zero false positives, which is why that is the one that
shipped. §317 now carries these numbers instead of the prediction.

319 is the general form. A sentence in the conditional tense, inside a
document whose purpose is to hold measurements, takes authority from the
numbers around it -- and unlike a wrong claim about existing code, it
cannot be checked until somebody builds the thing. Either build it and
write the number, or write "untested" beside it.

Refs #2876

The Rust header exists to name what this backend drops:

    // NOT LOWERED BY THIS BACKEND: 7 test(s), 2 invariant(s).
    // This backend emits declarations only.

Its own comment says why: "Emitting library code without tests is a
defensible policy; emitting it silently is the defect."

It said nothing about the third category. 817 functions across the
corpus are emitted as `unimplemented!()`, and a file holding two dozen
of them read as complete -- the banner accounted for the tests and the
invariants and left the stubs out of its own accounting.

The banner is written before the body, so the count needs a pre-pass.
The `has_body` predicate was inline in `gen_fn`; it is now a function
both call, because two copies of that list is exactly how `StmtAssign`
went missing from one of them (#2875).

Cross-checked over all 559 specs that generate: banner total 817,
`unimplemented!()` emitted 817, zero files where the two disagree.

FROZEN_HASH resealed. Suite unchanged.

Refs #2875