test(fact): LS-A-9 regression for callback POLL dispatch

[avrabe]

Summary

Closes the missing-regression gap for LS-A-9 that the LS-N verification gate has been surfacing. Prior fix at meld-core/src/adapter/fact.rs:3992-3998 added the WAIT || POLL OR-pattern to the callback dispatch but landed without a dedicated regression test.

Why this matters

The async-callback adapter dispatches [waitable-set-poll] when the guest callback returns WAIT (2) or POLL (3). A previous version matched only WAIT, silently letting POLL fall through to the YIELD path — which sent (EVENT_NONE, 0, 0) to [callback] and dropped host-ready events. The fused module diverged from the composed module on P3 async export call paths (H-3, UCA-F-3), and there's no host trap to catch it.

Test approach

ls_a_9_callback_adapter_dispatches_both_wait_and_poll:

1. Builds a minimal MergedModule with the prerequisites the callback emitter requires (lift func, [callback]<export> companion, [waitable-set-poll] import).
2. Drives generate_async_callback_adapter end-to-end.
3. Scans the emitted function body for the canonical byte sequence:

   0x41 0x02 0x46 0x20    i32.const WAIT(2) / i32.eq / local.get
   0x41 0x03 0x46 0x72    i32.const POLL(3) / i32.eq / i32.or
   

   in order. A regression that drops the POLL arm or reorders the comparisons fails this test.

Pin-by-substring is robust against unrelated body changes (locals layout, surrounding control flow).

Gate verdict

State	Before	After
Verified	17/19	18/19
Missing	2 (LS-A-8, LS-A-9)	1 (LS-A-8)
Failed	0	0

Remaining: LS-A-8 (HashMap rep_func selection for list<record<borrow>>) — surveyed in task #52 at ~2-4h scope.

End-to-end mythos-auto smoke test

This is the first PR to touch a Tier-5 source file (adapter/fact.rs) since the auto-runner's plumbing fixes (#164). Watching for:

• Detect job lists meld-core/src/adapter/fact.rs ✓
• Scan matrix actually fires (the path that was broken in test(merger): add LS-A-19 regression for exact resource-import lookup #163's run)
• claude-code-action emits structured verdict: NO_FINDINGS | FINDING JSON
• Sticky comment posted with per-file verdict
• mythos-pass-done label auto-applied on NO_FINDINGS

Local tests: cargo test ls_a_9_ → 1 pass; full meld-core --lib → 237/237.

🤖 Generated with Claude Code

[github-actions]

Mythos delta-pass required

This PR modifies one or more Tier-5 source files (per
scripts/mythos/rank.md):

meld-core/src/adapter/fact.rs

Before merge, run the Mythos discover protocol on the
modified Tier-5 files:

1. Follow scripts/mythos/discover.md
   — one fresh agent session per touched Tier-5 file.
2. For each finding, the agent must produce both a Kani
   harness and a failing PoC test (per the protocol's
   "if you cannot produce both, do not report" rule).
3. Attach a comment on this PR with either the findings
   (formatted per discover.md's output schema) or
   NO FINDINGS.
4. Add the mythos-pass-done label to this PR.

Why this gate exists: LS-A-10
(CABI alignment padding in async-lift retptr writeback) was
found by the v0.8.0 pre-release Mythos pass — but it had
lived in the callback emitter since #128, across six
releases. A PR-time gate would have caught it at review
time instead of at the release boundary.

The gate check on this PR will pass once the label is
applied.

[github-actions]

LS-N verification gate

⚠️ 18/19 verified — 1 missing regression tests

	count
Passed (≥1 test, all green)	18
Failed (≥1 test failure)	0
Missing (no ls_*_NN_* test found)	1

_{Approved loss-scenarios.yaml entries are expected to have a
regression test named ls_<letter>_<num>_* (e.g. LS-A-11 →
ls_a_11_*). The gate runs each prefix via cargo test --lib --no-fail-fast and aggregates pass/fail/missing.}

Failed LS entries

(none)

Missing regression tests

• LS-A-8

_{Updated automatically by tools/post_verification_comment.py.
Source of truth: safety/stpa/loss-scenarios.yaml.}

[avrabe]

Admin-merge per #139 (filed 2 smithy issues)

9 of 13 checks green including everything substantive (Test, Clippy, Coverage, Bench, LS-N gate, 2 fuzz smokes, Detect Tier-5 changes). The 4 remaining failures are all #139 infra, not regressions in this PR's content:

• Fuzz smoke (fuzz_parse_component) + Fuzz smoke (fuzz_resolver_terminates) — same "sanitizer is incompatible with statically linked libc" config drift documented in Smithy fleet reliability — rolling tracker (meld side) #139 §3. Reruns queued ~25min on the 7/7-busy fleet without picking up.
• Mythos pass (meld-core/src/adapter/fact.rs) — unzip missing on rust-cpu runners. oven-sh/setup-bun can't extract the bun zip, claude-code-action's bun-based entrypoints exit 127, placeholder-FINDING fallback fires.
• Aggregate findings + label + Mythos delta-pass gate — cascading from the auto-runner failure (no NO_FINDINGS verdict → no label → label-only gate fails).

Two new tracking issues filed for smithy:

• smithy: unzip missing on rust-cpu runners blocks claude-code-action #167 — unzip missing on rust-cpu runners blocks claude-code-action
• smithy: sanitizer/libc config drift recurrence on fuzz smoke jobs #168 — sanitizer/libc config drift recurrence on fuzz smoke jobs

LS-A-9 regression test itself is sound (1 pass, 237/237 in meld-core lib). The auto-runner was supposed to validate end-to-end here; with the unzip block, that validation is deferred until #167 ships.

Admin-merge counter for #139 since last reset:

• PR test(fact): LS-A-9 regression for callback POLL dispatch #166 — admin-merged (this PR, +1)