JooqRecordServiceParamPipelineTest asserts on generated helper body strings

JooqRecordServiceParamPipelineTest matches literal fragments of generated method bodies through two private helpers, methodBody(spec, name) and singularHelperBodies(spec), both of which render a MethodSpec to text and hand it to assertThat(...).contains(...). The helper-routing arms are where it concentrates: contendedSingularShapes_emitTwoDistinctHelpers_andEachFetcherRoutesToItsOwn, contendedListShapes_emitTwoDistinctListHelpers_andEachFetcherRoutesToItsOwn and identicalShapesAcrossDifferentInputTypes_collapseToOneBareHelper each assert on createFilmRecordN(env.getArgument( appearing (or not appearing) in a fetcher body, and one arm counts how many rendered helper bodies contain the string RELEASE_YEAR. docs/architecture/principles/development-principles.adoc bans code-string assertions on generated method bodies "at every tier", and docs/architecture/how-to/testing.adoc repeats the ban for the pipeline tier this file sits in.

Filed at the member-axis sibling’s Done gate, where the same shape was copied into InputBeanGroupingPipelineTest and then removed on review. That removal is worth reading first because it shows what the replacements look like in this neighbourhood: a descent’s declaration order became a compile error at a graphitron-sakila-example fixture, an unchecked cast turned out to be already enforced by that module’s -Xlint:all -Werror compile of the emitted tree, and the remaining pins moved onto the resolved carriers instead of the rendered text.

What makes this file harder than that one, and the reason it is filed rather than fixed in passing: the thing it wants to pin is genuinely a call-site routing fact ("this fetcher calls that helper"), which has no carrier of its own today. The rendered body is the only place the routing is currently visible, so the item has to decide where routing becomes observable before it can delete the scans. Candidates worth weighing: a structural assertion over the MethodSpec’s emitted call sites if the javapoet model exposes enough to name them, a carrier on the fetcher-generation side that records helper-name-per-arg so the pipeline tier can assert on it, or an execution-tier round-trip through two contended shapes where a mis-routed helper produces a visibly wrong record. The `RELEASE_YEAR count is the easiest half: which columns a helper writes is already a property of the binding shape, so it can be asserted before emit.

Adjacent, same rule, different files: R669 (RootLauncherRendererTest), R554 (the TypeSpecAssertions body-scan helpers), R522 (whether a "seam pin" is a legitimate exception to the ban at all). R522 in particular should probably be settled first, since a seam-pin carve-out would cover part of this file’s routing assertions and change how much of it needs to move.