E0277 root A residual (1): fn-generic bound inference is Clone-only — audience_subset/audience_join call rc_set_union<T: Ord + Clone> without their own signature earning P: Ord

[briansrls]

Auto-opened by session-dashboard for session tidy-otter-493.
Pushing to session/tidy-otter-493 advances this PR.

Worker attestation

Before flipping this PR to ready for review, confirm each item:

• Title describes the change (not the session id or branch).
• PR body summarises what and why (replace the TODO below).
• Tests run: name the command (e.g. npm test, cargo test) and the result.
• If this closes a work item, the body contains a Closes #N directive.
• No commits on this branch are surprises (no fork/cherry-pick I did not make).
• No secrets / credentials / large binaries staged.

Summary

TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.

Test plan

• TODO: list the commands that ran (or "no tests changed; relied on CI") and the outcome.

[gunbai-bot]

Closing: this is an auto-opened ghost of superseded work, not a proposal.

The dashboard opened this against session/tidy-otter-493 because the branch still carries the commits from #8761 after that PR was closed. The body is the untouched template (TODO summary, unchecked attestation) and the diff is the same 475 additions — nobody authored this as a submission.

Why the underlying work was superseded (agreed with its author before either session archived): it infers a trait bound by recognizing set_union call sites by name — its own body said "scoped to the set_union builtin's known lowering only, not a general trait-bound inference system." That is a heuristic standing where a structural fact was available. Once AudienceSet<P> carries an item-level P: Ord, fn audience_subset<P>(left: AudienceSet<P>, …) earns the bound by naming an Ord-bounded declared type — no walker, no builtin name. Checked three ways independently before the call: exactly two generic set_union call sites exist in the corpus, both naming AudienceSet<P> at value-param position, so no specimen requires the walker.

The author reached that conclusion themselves and named their own mechanism as the failure DESIGN §4 describes.

The subsuming work is #8770, whose successor lane is node://adhoc-630ac7f9-6e6. Since #8699 landed on main — generalizing the type-param renderer to a per-parameter trait map — the replacement is concrete rather than theoretical: the general rule feeds bounds_by_param into a renderer that now exists.

The branch is deliberately retained, not deleted, for the evidence recorded on #8770. Note that the auto-opener will keep re-raising a PR while the branch has commits — if that becomes noise, the branch can go once that evidence is captured elsewhere.

— sent from bright-moth-92