test(merger): add LS-A-19 regression for exact resource-import lookup by avrabe · Pull Request #163 · pulseengine/meld

avrabe · 2026-05-17T07:14:33Z

Summary

Closes the missing-regression gap for LS-A-19 that the LS-N verification gate (#161) surfaced on its first run. PR #156 fixed the imp.name.ends_with(rn) suffix-collision bug in Merger::add_unresolved_imports but landed without a dedicated regression test.

Refactors the dedup-skip path's inline .values().find(...) lookup into a private helper find_exact_resource_import_idx and adds three regression tests pinning the suffix-collision boundaries.

Why this matters

The bug: when two resources shared a suffix (float / bigfloat), the dedup-skip path's imp.name.ends_with("float") matched both [resource-rep]float AND [resource-rep]bigfloat. .find() returned whichever the HashMap iterated first (also LS-A-15 territory). The wrong import index got stored in resource_rep_by_component, routing subsequent float calls on the second component to the bigfloat import. Validator passes (both share (i32) -> i32); runtime sees bigfloat rep as float — type confusion with no trap (H-10).

Tests added

Test	What it pins
`ls_a_19_exact_match_picks_float_not_bigfloat`	Both names in tracking; exact match returns the right index regardless of HashMap iteration order
`ls_a_19_no_match_returns_none_even_with_suffix_collision`	Only `bigfloat` in tracking, ask for `float` → must return None (buggy `ends_with` would have returned bigfloat's index)
`ls_a_19_resource_new_lookup_is_also_exact`	Same case for the `[resource-new]` table

Gate impact

LS-N verification gate verdict moves from 15/19 verified (4 missing) → 16/19 verified (3 missing).

Remaining missing-bucket entries:

LS-CP-4 (DWARF passthrough) — likely subsumed by DWARF / source-map remapping for fused output (witness MC/DC integration) #130 Phase 2 (DWARF Phase 2: address remap (#130) #143)
LS-A-8 (HashMap rep_func selection)
LS-A-9 (async callback POLL fallthrough)

End-to-end test for mythos-auto.yml

This is also the first real PR exercising mythos-auto.yml (added in #162). It touches a Tier-5 file (merger.rs), so the auto-runner should fire end-to-end:

Detect Tier-5 changes job passes the actor gate (avrabe)
Matrix runs claude-code-action on merger.rs with scripts/mythos/discover.md
Aggregate job posts sticky  comment
If NO_FINDINGS, auto-applies mythos-pass-done label

Watching for: budget consumption per single-file scan, sticky-comment markdown rendering, label-application timing.

🤖 Generated with Claude Code

PR #156 fixed the `imp.name.ends_with(rn)` suffix-collision bug in `Merger::add_unresolved_imports` (the dedup-skip path propagating `resource_rep_by_component` / `resource_new_by_component` entries) but landed without a regression test. The LS-N verification gate (#161) surfaced this gap as missing coverage on the next-to-last sweep. Extracts the exact-match lookup into a private helper `find_exact_resource_import_idx` and adds three regression tests: - `ls_a_19_exact_match_picks_float_not_bigfloat` — both `float` and `bigfloat` in tracking; asking for `[resource-rep]float` must return float's index, not bigfloat's. The buggy `ends_with` form would match bigfloat under some iteration orders. - `ls_a_19_no_match_returns_none_even_with_suffix_collision` — only bigfloat in tracking, caller asks for plain float. Exact match must return None; the buggy `ends_with` form would return bigfloat's index. - `ls_a_19_resource_new_lookup_is_also_exact` — same suffix- collision case for the `[resource-new]` table. LS-N gate result moves from 15/19 verified (4 missing) to 16/19 verified (3 missing). Remaining missing-bucket entries are LS-CP-4 (likely subsumed by #130 Phase 2), LS-A-8, LS-A-9 — tracked separately as research items. This is also the first real PR exercising the `mythos-auto.yml` workflow added in #162: it touches a Tier-5 file (`merger.rs`) so the auto-runner will fire end-to-end on PR open. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

github-actions · 2026-05-17T07:15:44Z

Mythos delta-pass required

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

meld-core/src/merger.rs

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

Follow scripts/mythos/discover.md
— one fresh agent session per touched Tier-5 file.
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).
Attach a comment on this PR with either the findings
(formatted per discover.md's output schema) or
NO FINDINGS.
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 · 2026-05-17T14:23:14Z

LS-N verification gate

⚠️ 16/19 verified — 3 missing regression tests

	count
Passed (≥1 test, all green)	16
Failed (≥1 test failure)	0
Missing (no `ls__NN_` test found)	3

_{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-CP-4
LS-A-8
LS-A-9

_{Updated automatically by tools/post_verification_comment.py.

Source of truth: safety/stpa/loss-scenarios.yaml.}

avrabe · 2026-05-18T05:03:13Z

Admin-merging + plumbing follow-up

Two distinct issues blocked the PR's CI gates; only the first one is a real test failure that would block merge.

#139 capacity (real)

3 standard CI jobs still queued (Clippy + 2 fuzz smokes) after ~2 hours. After cancelling 5 hung cross-org jobs (loom × 3 + spar × 2, ages 9-22h, matching #139 §4 "hung jobs hold slots forever" pattern), capacity partially recovered but the queue cleared in zigzag.

mythos-auto.yml plumbing (will fix in #164)

The Mythos pass (meld-core/src/merger.rs) job failure is workflow plumbing, NOT a Mythos finding:

14:32:12 Error: Unable to locate executable file: unzip. Please verify either the file path exists...
14:32:12 /var/lib/runners/runner5/_work/_temp/.../d9f8a6f8.sh: line 3: bun: command not found
14:32:12 ##[error]Process completed with exit code 127.

oven-sh/setup-bun failed because the rust-cpu runner lacks unzip. The action's fallback to a system bun also failed because there is no system bun. claude-code-action's bun-based post-step entrypoints then exit 127.

The placeholder-FINDING fallback we wired in for exactly this scenario fired correctly, but the slug step output was empty (separate plumbing bug), producing mythos-out/.json as the artifact path.

Why admin-merge is OK here

The LS-A-19 regression is verified locally (cargo test ls_a_19_ → 3 pass; full suite 236/236)
The LS-N gate (verification-gate.yml) passes — that's the test-coverage gate this PR was written to satisfy
Format, Test, Coverage, Bench, fuzz_merger_idempotent, fuzz_resolver_terminates, LS-N gate: all 7 ran and passed
The 3 remaining stuck jobs are independent of this PR's contents
Same Smithy fleet reliability — rolling tracker (meld side) #139 admin-merge precedent as feat(ci): LS-N verification gate (spar-pattern port) #161 / feat(ci): Mythos delta-pass auto-runner (single-actor, OAuth-token) #162

Admin-merge counter for #139:

PR test(merger): add LS-A-19 regression for exact resource-import lookup #163 — admin-merged (this PR)

Follow-up PR #164 (next) will fix mythos-auto plumbing:

Pre-install unzip (and verify bun once setup-bun works)
Debug the empty-slug step (likely matrix-expression evaluation order)
Confirm OAuth-token wiring is the right input for non-interactive use

PR #163 was the first end-to-end test of mythos-auto.yml (added in #162). It surfaced three plumbing issues: 1. The action's `oven-sh/setup-bun` step requires `unzip`, which is not installed by default on the rust-cpu runners. Without it the action's bun-based post-step entrypoints exit 127, and the whole scan-step exits failure before emitting structured output. 2. The `Slugify file path for artifact name` step sat AFTER the discover step with no `if: always()`. When discover failed, the slug step was skipped, leaving `steps.slug.outputs.slug` empty. Downstream `if: always()` steps then wrote `mythos-out/.json` (no slug) and `upload-artifact` complained "No files were found with the provided path: mythos-out/.json". 3. The `Save structured output as artifact` step embedded `${{ steps.slug.outputs.slug }}` in the run-block via direct interpolation. Silently substituting an empty slug into a file path is a footgun even if the slug step had run — better to read slug from an env var and fail loudly on empty. Fixes: - Slugify step moves BEFORE the discover step, so it always runs (no `if: always()` needed because both detect+slug are the precondition for everything below). - New `Install unzip (required by setup-bun)` step, best-effort apt install mirroring the action's own subprocess-isolation install pattern. `continue-on-error: true` so non-Debian runners don't break the workflow. - `Save structured output as artifact` reads slug from env (`SLUG`) rather than `${{ }}` interpolation; explicitly errors out if SLUG is empty rather than silently writing to a malformed path. - `upload-artifact` step gains an extra `steps.slug.outputs.slug != ''` guard so it never tries to upload with an empty name. The placeholder-FINDING fallback (the part that surfaced these issues by writing "discover step failed before emitting structured output" into the aggregate comment) is intentional and stays — it guarantees the gate blocks on workflow failure rather than silently passing. Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

P3 cross-component stream-pair detection foundation + a fully operational Mythos delta-pass auto-runner. 12 commits since v0.8.1. Headline changes: - Cross-component stream<T> pairing detection (#141, ADR-3). The StreamPairGraph foundation for the in-module stream adapter: meld now inventories at resolve time which fused components form producer -> consumer stream pairings. The ring-buffer / copy-chain emitter is a runtime-verified follow-up (ADR-3 Path N). - Mythos delta-pass auto-runner (#162, #164, #170, #173, #175). The AI-driven discover protocol now runs automatically on every Tier-5 PR by the maintainer, via claude-code-action on a Max-plan OAuth token. Five plumbing fixes brought it to a working end-to-end state: scan -> NO_FINDINGS verdict -> sticky comment -> mythos-pass-done label. - LS-N verification gate (#161, #165). Every approved loss-scenario in safety/stpa/loss-scenarios.yaml is now enforced to have a matching ls_<letter>_<num>_* regression test; 19/19 verified. - DWARF / witness-mapping discovery (#131) — Phase 1 of the #130 epic; pins today's lossy passthrough as the green-to-red oracle for the Phase 2 remap work. - Regression coverage for LS-A-8/9/19 and LS-CP-4 (#163/165/166/169) — closed every missing-test entry the LS-N gate surfaced. - CI footprint reduction (#171) — bench/fuzz/ci skip on docs- and safety-only PRs; meld is a leaner consumer of the shared fleet. - fuzz.yml musl-target drop (#170, closes #168) — fixes the recurring "sanitizer incompatible with statically linked libc" fuzz failures. Co-authored-by: Claude Opus 4.7 <noreply@anthropic.com>

avrabe merged commit 84a1ed9 into main May 18, 2026
10 of 14 checks passed

avrabe deleted the fix/ls-a-19-exact-resource-import-lookup branch May 18, 2026 05:03

avrabe mentioned this pull request May 18, 2026

fix(ci): mythos-auto plumbing — slug ordering, unzip install #164

Merged

3 tasks

This was referenced May 19, 2026

test(fact): LS-A-9 regression for callback POLL dispatch #166

Merged

smithy: unzip missing on rust-cpu runners blocks claude-code-action #167

Closed

chore: release v0.9.0 #176

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

test(merger): add LS-A-19 regression for exact resource-import lookup#163

test(merger): add LS-A-19 regression for exact resource-import lookup#163
avrabe merged 1 commit into
mainfrom
fix/ls-a-19-exact-resource-import-lookup

avrabe commented May 17, 2026

Uh oh!

github-actions Bot commented May 17, 2026

Uh oh!

github-actions Bot commented May 17, 2026

Uh oh!

avrabe commented May 18, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant

Conversation

avrabe commented May 17, 2026

Summary

Why this matters

Tests added

Gate impact

End-to-end test for mythos-auto.yml

Uh oh!

github-actions Bot commented May 17, 2026

Mythos delta-pass required

Uh oh!

github-actions Bot commented May 17, 2026

LS-N verification gate

Uh oh!

avrabe commented May 18, 2026

Admin-merging + plumbing follow-up

#139 capacity (real)

mythos-auto.yml plumbing (will fix in #164)

Why admin-merge is OK here

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant