RootLauncherRendererTest asserts on generated body strings

RootLauncherRendererTest asserts on the rendered text of generated method bodies. It calls render(row).code().toString() through a body(...) helper and then matches literal Java and jOOQ source fragments against it, roughly 33 contains / doesNotContain calls across its arms (.contains("rows = filmByIdInputRows(env, filmTable)"), .contains(".values(rows).as(\"filmByIdInput\", \"idx\", \"film_id\")"), and so on). docs/architecture/principles/development-principles.adoc bans exactly this: "Code-string assertions on generated method bodies are banned at every tier: they test implementation, not behaviour, and break on every refactor", with the compile and execution tiers named as the replacement. docs/architecture/how-to/testing.adoc repeats the ban for the neighbouring sub-families. The ban is review-enforced rather than build-enforced, and this file is where it has drifted furthest.

The reason to file it rather than fix it opportunistically is that the file is not wrong about what it wants to pin, only about how. A renderer is a total function over a command’s sealed arms and per-arm coverage of the emit is genuinely valuable; testing.adoc’s renderer-arm-test paragraph says so. The question the item has to answer is what the structural form of each assertion should be once the body string is off the table: `MethodSpec shape (name, return type, parameters) is already asserted separately in the same file, the SQL that actually reaches PostgreSQL is pinned by RootLauncherSqlBaselineTest, and emitted runtime helpers can be invoked reflectively the way ScatterLookupByIdxTest does. A plausible outcome is that most of the 33 assertions are already covered at a sanctioned tier and simply delete, with a residue that needs a baseline or a reflective-invocation test built for it. The audit is the work; the deletions are cheap.

Provenance: found at the third In Review gate on the lookup positional-contract item, which added three such assertions for its scatter arm. That item removes its own three; this one covers the pre-existing remainder, which is out of its scope.