Whole-tree compile speed: rc_map_insert quadratic + M1 resolve memo

[gunbai-bot]

Auto-opened by session-dashboard for session bold-crane-271.
Pushing to session/bold-crane-271-compile-speed 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Closing as superseded — the mergeable slice of this work lives in #6996.

Disposition:

• Fix B (reconcile_with_typed_cache clone elimination) — isolated, reviewed, and ready in Fix variant_surfaces double-clone in reconcile_with_typed_cache #6996.
• Fix A (rc_map_insert try_unwrap fast path) — local whole-tree benchmarks showed no wall-clock win vs main (A+B 193.5s vs baseline 188.5s); with im::HashMap + Rc::make_mut the extra path is likely inert. Per §6, not landing churn without displaced cost.
• M1 resolve memo — already on main; no PR needed.
• Benchmark harness (measure_cold_compile, frozen unit order) — WIP scaffolding only; not ready for review (incomplete roster provenance, no attributed A/B receipt). Can reopen as a separate probe PR if the lane still wants a frozen protocol.

— sent from bold-crane-271