release: v1.1.0 — ægraph production substrate + first mechanized roundtrip proof

[avrabe]

Summary

Minor-version release consolidating the v1.1.0 sprint. The v1.0.4 ægraph substrate becomes a default-on pipeline pass with cost-driven extraction and a widened rule set, and the parser/encoder roundtrip proof (#48) gains a real Rocq scaffold.

Byte-neutral on the current corpus — this is an infrastructure and correctness release, not a size-win release.

Shipped

• Track A (feat(proofs): close #48 — parser/encoder roundtrip identity (v1.1.0 Track A, Path A) #135) — Path A for Prove parser/encoder round-trip identity #48: Rocq parser/encoder roundtrip proof scaffold. proofs/ Admitted. count 4 → 2.
• Track B (feat(opt): cost-driven ægraph extraction (v1.1.0 Track B) #134) — cost-driven ægraph extraction. Re-applied in this PR: feat(opt): cost-driven ægraph extraction (v1.1.0 Track B) #134 merged but its egraph.rs/lib.rs diff was silently clobbered when feat(opt): ægraph i64 ops + 8 new identity rules + commutativity helpers (v1.1.0 Track C) #137's rebase resolved conflicts by whole-file copy from a pre-feat(opt): cost-driven ægraph extraction (v1.1.0 Track B) #134 branch. This PR restores the byte-cost model + UF-root extraction.
• Track C (feat(opt): ægraph i64 ops + 8 new identity rules + commutativity helpers (v1.1.0 Track C) #137) — ægraph i64 ops (11 new variants) + 8 new identity rules + commutativity normalization helpers.
• Track F — ægraph pass is default-on, revert-safe by construction (egraph_optimize_body splices extraction back only when strictly shorter). New docs/measurements/v1.1.0-corpus-baseline.md. measure_corpus.sh pct_delta no longer fabricates a -100% "win" on error/timeout rows.

Deferred to v1.1.1

• Track D — Track-3 housekeeping (touches every fused-optimizer call site).
• Track E — real meld-fused fixture (blocked on a meld-binary permission wall); shipped as a documented placeholder.

Lint debt paydown

Repo-wide cargo fmt + 12 clippy warnings fixed (collapsible_if, needless_range_loop → slice idioms, empty_line_after_doc_comment, doc-overindent, too_many_arguments allow on the CLI entrypoint) so the fmt + clippy pre-commit gates pass.

Test plan

• cargo fmt --all -- --check clean
• cargo clippy --all-targets --all-features -- -D warnings clean
• 378 loom-core lib tests pass
• 25 egraph tests pass (Track B/C coverage)
• CI full integration suite
• 4 pre-existing Z3-inline tests (test_inline_*) — hang locally under heavy machine load; verify in CI