docs(vcr): record landed optimized-path fixes + the optimizer_bridge control-flow frontier (#242)

[avrabe]

What

Behavior-frozen, doc-only update to the rivet-traced #242 roadmap (artifacts/verified-codegen-roadmap.yaml). VCR-ORACLE-001's running log was stale — it ended at "#490 blocks the non-leaf prologue lever / the lever stays unmerged until #490 lands", but both that fix and a companion control-flow fix have since landed. This records the resolution and the broader frontier the work surfaced, so the roadmap reflects reality and prioritizes the next gated step.

Appended to VCR-ORACLE-001

• Optimized path: small leaf functions clobber callee-saved r4-r8 without a saving prologue (latent ABI bug) #490 LANDED (PR fix(vcr-ra): optimized path preserves callee-saved r4-r8 — AAPCS prologue (#490, #242) #495): optimized path now wraps a body touching r4-r8 in push {r4-r8,lr}/pop {r4-r8,pc}, decided on the post-realloc body (no spurious push; >4-param fns decline to the direct selector so incoming stack-arg offsets are untouched), CI oracle callee_saved_490_differential.py. The non-leaf prologue lever it blocked is now unblocked.
• Optimized (non-relocatable) path miscompiles block/br_if: branch target lands mid-instruction #483 LANDED (PR fix(optimizer): optimized-path block/br_if branch lands mid-instruction (#483, #242) #497): forward block+br_if mid-instruction landing — End labelled with its own id not the closed block's, plus Strh/Strb/Ldrh/… sized 2 vs the 4-byte .w the high-base optimized path emits; CI oracle block_brif_483_differential.py.
• The multi-modal control-flow sweep established these are one cluster of optimizer_bridge control-flow/ABI defects — concrete measured evidence for VCR-SEL-001's "collapse the two/three-selector accretion" motivation — and the next gated frontier: Optimized path: br_if to a forward block end still unresolved for if/else and sequential-block shapes (#483 class, broader) #500 (br_if still unresolved for if/else + sequential blocks), Optimized path: spill frame not deallocated before return (missing 'add sp') on a function with control flow #499 (spill frame not deallocated under control flow; Optimized path: small leaf functions clobber callee-saved r4-r8 without a saving prologue (latent ABI bug) #490's pop epilogue made the SP leak fatal — v0.11.17 regression: guard-elision miscompiles fully-dissolved flight_algo (controller tail wrong; control_step OK; was correct in v0.11.16) #215-class), Optimized-path byte-size estimator is incomplete (Cmp/Cmn/Adds/Subs/Popcnt) — latent branch-displacement drift #498 (size estimator incomplete for Cmp/Cmn/Adds/Subs/Popcnt), Optimized path silently miscompiles register-exhausting i32 folds (no spill, no decline) #496 (register-exhausting folds silently miscompiled). All separate gated steps; none in shipped firmware (gale ships --relocatable/direct).

Verification

• rivet validate profile unchanged vs baseline (49 errors / 75 warnings / 0 broken cross-refs — all pre-existing xref/coverage gaps; NON_XREF = 0, the CI gate's pass condition); rivet coverage exit 0.
• No code changed — frozen byte gate / differentials unaffected.

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!