test(installer): bind execution coverage callers
ci/woodpecker/pr/ci Pipeline failed
ci/woodpecker/pr/greenfield-install Pipeline was successful

This commit is contained in:
2026-08-06 03:48:03 -05:00
parent df705828a4
commit d66e91b1f2
7 changed files with 205 additions and 18 deletions
+5 -13
View File
@@ -132,16 +132,8 @@ steps:
- |
coverage_run="${CI_PIPELINE_NUMBER}-${CI_WORKFLOW_NUMBER}"
coverage_root=.mosaic-test-work/greenfield-execution-coverage
set +e
bash tools/verify-greenfield-execution-coverage.sh check cases \
tools/fixtures/greenfield-expected-red.tsv "$coverage_root" \
"$coverage_run"
cases_status=$?
bash tools/verify-greenfield-execution-coverage.sh check arms \
tools/fixtures/greenfield-expected-arms.txt "$coverage_root" \
"$coverage_run"
arms_status=$?
set -e
if [ "$cases_status" -ne 0 ] || [ "$arms_status" -ne 0 ]; then
exit 1
fi
bash tools/verify-greenfield-execution-coverage-gate.sh \
tools/verify-greenfield-execution-coverage.sh \
tools/fixtures/greenfield-expected-red.tsv \
tools/fixtures/greenfield-expected-arms.txt \
"$coverage_root" "$coverage_run"
+1 -1
View File
@@ -1393,7 +1393,7 @@ A from-zero install can report success while leaving the target host unusable be
12. The expected-RED comparator SHALL validate the complete manifest before selecting a case: exact case population, one exit and P0P9 disposition per case, pinned require/forbid classes, and no malformed, duplicate, or unknown rows.
13. The published installer contract SHALL reject failed fetches, HTTP-success empty bodies, and digest mismatch, then execute the exact digest-verified body. The remote CI arm SHALL enumerate every payload-acquisition path and report a bound/found denominator. It SHALL bind both the installer body and the downstream stack framework/source it consumes to the same immutable CI commit, while retaining `--next` as the lane selector. Source resolution/acquisition SHALL fail closed, and the arm SHALL verify the realised source commit and archive digest after installation. Any out-of-scope unpinned sibling SHALL be named and counted rather than silently included in a broader pinning claim. `ASSUMPTION:` the configured repository's authenticated exact-commit endpoint is trusted to map that commit ID to the returned archive bytes; independent signed provenance/authenticity against repository or TLS trust-root compromise remains excluded by canonical greenfield-install PRD v2 §3.
14. Phase diagnostics SHALL be redacted before terminal or durable-log output. A seeded positive-control canary SHALL remain absent from observed argv, output, command logs, npm configuration, generated files, and shell history.
15. The CI fixture SHALL publish pipeline-level execution coverage for both the unconditional case set derived from the expected-RED manifest and the explicitly declared required pipeline-arm set, including the immutable remote-installer contract. Per-case and per-arm success markers SHALL be run-scoped, stored beneath an existing checkout-archive exclusion, and written only after that arm's verifier passes. A final step SHALL depend on the complete fixture matrix, run after prior success or failure, emit both `cases_defined=N cases_executed=M` and `arms_defined=N arms_executed=M`, and fail unless both expected/executed name sets are exactly equal; missing, unexpected, stale, newly added unexecuted, or checkout-contaminating state SHALL fail closed.
15. The CI fixture SHALL publish pipeline-level execution coverage for both the unconditional case set derived from the expected-RED manifest and the explicitly declared required pipeline-arm set, including the immutable remote-installer contract. Per-case and per-arm success markers SHALL be run-scoped, stored beneath an existing checkout-archive exclusion, and written only after that arm's verifier passes. A final step SHALL depend on the complete fixture matrix, run after prior success or failure, emit both `cases_defined=N cases_executed=M` and `arms_defined=N arms_executed=M`, and invoke the same testable aggregation helper whose complete case/arm PASS/FAIL truth table is covered. It SHALL fail unless both expected/executed name sets are exactly equal; missing, unexpected, stale, newly added unexecuted, or checkout-contaminating state SHALL fail closed. The archive-purity control SHALL bind to the production checkout-archive selector so deleting the production exclusion makes the control RED.
### C1 acceptance criteria
@@ -0,0 +1,60 @@
# #1050 C1 fix round — Round 6 caller coupling and production archive binding
BASE HEAD
- df705828a439c6795cd47c938ed7f838b956ebf3
- Reviews 114 (rev-security-02) and 115 (be-coder-06) independently reproduced the same caller-level mutant: ignoring arms_status in the workflow left the focused helper suite green.
- Both reviewers confirmed the production aggregation was correct and requested test reachability, not a logic rewrite.
- rev-974 separately deleted only the production e2e-install-test.sh .mosaic-test-work tar exclusion; the copied-selector suite remained green.
RED
- After adding the four-cell caller truth-table expectations but before creating the workflow helper:
bash tools/verify-greenfield-execution-coverage.test.sh => rc=1
The expected helper did not exist, so the test could not reach a passing aggregation implementation.
- Prior independent controls:
ignore-arms workflow mutant => rc=0 before this round (reviews 114/115)
delete-production-archive-exclusion mutant => rc=0 before this round (rev-974 exact-head sandbox)
IMPLEMENTATION
- Extracted the existing set +e / two checks / status capture / final OR into tools/verify-greenfield-execution-coverage-gate.sh without changing its semantics.
- .woodpecker/greenfield-install.yml now invokes that helper directly; its inline duplicate was removed.
- The focused suite invokes the same helper used by the workflow across all four cells:
cases PASS / arms FAIL => RED
cases FAIL / arms PASS => RED
both FAIL => both diagnostic outputs and RED
both PASS => GREEN
- The archive-purity test now first locates the production checkout-archive command in tools/e2e-install-test.sh and requires its .mosaic-test-work exclusion, then proves the exclusion's archive effect.
- GREENFIELD_COVERAGE_GATE and GREENFIELD_E2E_INSTALLER are test-only subject injection seams used to prove the requested mutants; production paths remain the defaults.
MUTATION CONTROLS
- ignore-arms aggregation helper => focused suite rc=1
- ignore-cases aggregation helper => focused suite rc=1
- always-RED aggregation helper => focused suite rc=1
- delete only production .mosaic-test-work archive exclusion => focused suite rc=1
- restored production subjects => greenfield execution coverage tests passed
- final mutant workspace: /tmp/c1-round6-final-mutants.u0ZaL6
- logs: /tmp/c1-round6-final-gate-ignore-arms.log, /tmp/c1-round6-final-gate-ignore-cases.log, /tmp/c1-round6-final-gate-always-red.log, /tmp/c1-round6-final-archive-mutant.log
FENCED FILES
- tools/e2e-install-test.sh HEAD/worktree SHA-256: a93113565aa69f2c6f3d792b78251021bb3bbe3f7813d5fed547ab0099fa3b98 (byte-identical)
- tools/fixtures/greenfield-expected-red.tsv HEAD/worktree SHA-256: baae06c664d0ab7ffe4c32ec106bfcb6f9dd74b46461d3c4440b73dc582af1a2 (byte-identical)
- No per-case disposition or #869 wiring changed.
LOCAL BASELINES
- /home free before build: 11G
- bash -n focused scripts: PASS
- ShellCheck focused scripts: PASS
- bash tools/verify-greenfield-execution-coverage.test.sh: PASS
- pnpm test:installer: PASS
- pnpm typecheck: PASS (45/45)
- pnpm lint: PASS (25/25)
- pnpm format:check: PASS
- woodpecker-cli lint --strict .woodpecker/greenfield-install.yml: PASS
- git diff --check: PASS
INDEPENDENT REVIEW
- Codex code review: APPROVE, confidence 0.94, seven files reviewed, zero findings. It confirmed the workflow invokes the exact four-cell-tested helper, both diagnostics remain observable, and archive purity binds to the production selector.
- Codex security review: risk NONE, confidence 0.97, seven files reviewed, zero findings. It confirmed fail-closed behavior, quoted inputs, repository-controlled production arguments, and no new secret/injection/access-control exposure.
- The code-review sandbox could not execute the dynamic suite because its filesystem is read-only; the local measured baselines and mutation controls above provide dynamic evidence.
PENDING
- commit, queue guard, one lease-pinned push, provider attribution read-back
@@ -2,7 +2,7 @@
## Required artifacts
- [x] `docs/PRD.md` exists; #1050 C1 requirement 13 binds the complete in-scope remote stack source chain, and requirement 15 binds pipeline-level exact case+arm execution coverage without checkout-state contamination.
- [x] `docs/PRD.md` exists; #1050 C1 requirement 13 binds the complete in-scope remote stack source chain, and requirement 15 binds pipeline-level exact case+arm execution coverage, executable aggregation, and production-bound archive purity.
- [x] User guide: not applicable; no user-facing production installer behavior changed.
- [x] Admin guide: not applicable; no operator procedure or deployment behavior changed.
- [x] Developer guide: existing `docs/guides/installer-state-machine.md` defines the fail-closed installer model; the C1-specific remote-arm contract and evidence live in the PRD, scratchpad, and scoped verification report.
@@ -22,4 +22,5 @@
- [x] Round-3 Codex code review APPROVE (confidence 0.92, zero findings) and security re-review risk NONE (confidence 0.96, zero findings). The initial HIGH trust-root finding remains documented as the canonical signed-provenance deferral, not hidden.
- [x] Round-4 review cycle closed: two blocking Codex findings (missing dependency ordering; production manifest comments misparsed) were accepted and fixed. Final code review APPROVE (confidence 0.94, zero findings) and security risk NONE (confidence 0.96, zero findings).
- [x] Round-5 exact arm coverage and archive-purity controls reviewed: Codex code APPROVE (confidence 0.93, zero findings) and security risk NONE (confidence 0.96, zero findings).
- [x] Round-6 executable aggregation and production archive-selector binding reviewed: Codex code APPROVE (confidence 0.94, zero findings) and security risk NONE (confidence 0.97, zero findings).
- [x] Canonical evidence remains in-repo. No external publishing action was requested or performed.
@@ -160,3 +160,11 @@ Denominator corrections were sent to and accepted by the lane lead before implem
- State moved beneath `.mosaic-test-work/greenfield-execution-coverage`, which the existing checkout selector excludes. The regression control proves a non-excluded root marker is archived while no `.mosaic-test-work` path is archived.
- Final review: Codex code APPROVE confidence 0.93 and security risk NONE confidence 0.96, both with zero findings. Full installer tests, typecheck 45/45, lint 25/25, format, Bash syntax, ShellCheck, Woodpecker strict lint, and diff check pass with 11G free before/after.
- Full evidence: `docs/reports/verification/1050-c1-fix-round/11-round5-arm-coverage-and-purity.txt`.
## Round 6 — executable aggregation and production archive binding
- Reviews 114/115 independently reproduced the same caller-level mutant at `df705828`: ignoring `arms_status` in the workflow still left the helper-bounded focused suite green. The production OR was correct and was not rewritten; its aggregation moved byte-for-byte into `verify-greenfield-execution-coverage-gate.sh`, which the workflow invokes directly.
- The focused suite now exercises that exact helper across the complete truth table: cases PASS/arms FAIL RED; cases FAIL/arms PASS RED; both FAIL emits both outputs and is RED; both PASS GREEN. Local ignore-arms, ignore-cases, and always-RED mutants each make the suite RED.
- Review `rev-974` separately proved that deleting only the production `e2e-install-test.sh` `.mosaic-test-work` tar exclusion left the copied-selector control green. The control now structurally binds its semantic archive test to the production checkout-archive command; the delete-production-binding mutant is RED. The fenced production installer remains byte-unchanged.
- Final review: Codex code APPROVE confidence 0.94 and security risk NONE confidence 0.97, both with zero findings. Full installer tests, typecheck 45/45, lint 25/25, format, Bash syntax, ShellCheck, Woodpecker strict lint, focused truth-table and four mutant controls, and diff check pass with 11G free before validation.
- Full evidence: `docs/reports/verification/1050-c1-fix-round/12-round6-caller-coupling-and-archive-binding.txt`.
+27
View File
@@ -0,0 +1,27 @@
#!/usr/bin/env bash
set -euo pipefail
usage() {
echo "usage: $0 <coverage-checker> <case-manifest> <arm-manifest> <state-root> <run-id>" >&2
}
[[ "$#" -eq 5 ]] || { usage; exit 2; }
checker="$1"
case_manifest="$2"
arm_manifest="$3"
state_root="$4"
run_id="$5"
[[ -f "$checker" ]] \
|| { echo "[fixture-suite] coverage checker missing: $checker" >&2; exit 2; }
set +e
bash "$checker" check cases "$case_manifest" "$state_root" "$run_id"
cases_status=$?
bash "$checker" check arms "$arm_manifest" "$state_root" "$run_id"
arms_status=$?
set -e
if [ "$cases_status" -ne 0 ] || [ "$arms_status" -ne 0 ]; then
exit 1
fi
@@ -3,7 +3,9 @@ set -euo pipefail
ROOT="$(cd "$(dirname "${BASH_SOURCE[0]}")/.." && pwd)"
SUBJECT="$ROOT/tools/verify-greenfield-execution-coverage.sh"
GATE="${GREENFIELD_COVERAGE_GATE:-$ROOT/tools/verify-greenfield-execution-coverage-gate.sh}"
WORKFLOW="$ROOT/.woodpecker/greenfield-install.yml"
INSTALLER="${GREENFIELD_E2E_INSTALLER:-$ROOT/tools/e2e-install-test.sh}"
mkdir -p "$ROOT/.mosaic-test-work"
TMP="$(mktemp -d "$ROOT/.mosaic-test-work/execution-coverage.XXXXXX")"
selector_control="$ROOT/greenfield-coverage-selector-control.ran"
@@ -75,6 +77,85 @@ done
production_output="$(bash "$SUBJECT" check cases "$production_manifest" "$production_state" production-1)"
grep -qF '[fixture-suite] cases_defined=3 cases_executed=3' <<<"$production_output"
mark_all_cases() {
local root="$1" run="$2"
for case_name in next-git-present main-git-present next-git-absent; do
bash "$SUBJECT" mark cases "$production_manifest" "$root" "$run" "$case_name"
done
}
mark_all_arms() {
local root="$1" run="$2"
for arm_name in \
greenfield-git-present \
greenfield-main-git-present \
greenfield-remote-installer-contract \
greenfield-git-absent; do
bash "$SUBJECT" mark arms "$production_arms" "$root" "$run" "$arm_name"
done
}
gate_state="$TMP/gate-state"
gate_run="truth-table"
# Final aggregation truth table: cases PASS / arms FAIL is red.
bash "$SUBJECT" init cases "$production_manifest" "$gate_state" "$gate_run"
bash "$SUBJECT" init arms "$production_arms" "$gate_state" "$gate_run"
mark_all_cases "$gate_state" "$gate_run"
for arm_name in greenfield-git-present greenfield-main-git-present greenfield-git-absent; do
bash "$SUBJECT" mark arms "$production_arms" "$gate_state" "$gate_run" "$arm_name"
done
set +e
gate_cases_pass_output="$(bash "$GATE" "$SUBJECT" "$production_manifest" "$production_arms" "$gate_state" "$gate_run" 2>&1)"
gate_cases_pass_status=$?
set -e
[[ "$gate_cases_pass_status" -eq 1 ]]
grep -qF '[fixture-suite] cases_defined=3 cases_executed=3' <<<"$gate_cases_pass_output"
grep -qF '[fixture-suite] arms_defined=4 arms_executed=3' <<<"$gate_cases_pass_output"
grep -qF '[fixture-suite] missing_arm=greenfield-remote-installer-contract' <<<"$gate_cases_pass_output"
# Final aggregation truth table: cases FAIL / arms PASS is red.
bash "$SUBJECT" init cases "$production_manifest" "$gate_state" "$gate_run"
bash "$SUBJECT" init arms "$production_arms" "$gate_state" "$gate_run"
for case_name in next-git-present main-git-present; do
bash "$SUBJECT" mark cases "$production_manifest" "$gate_state" "$gate_run" "$case_name"
done
mark_all_arms "$gate_state" "$gate_run"
set +e
gate_arms_pass_output="$(bash "$GATE" "$SUBJECT" "$production_manifest" "$production_arms" "$gate_state" "$gate_run" 2>&1)"
gate_arms_pass_status=$?
set -e
[[ "$gate_arms_pass_status" -eq 1 ]]
grep -qF '[fixture-suite] cases_defined=3 cases_executed=2' <<<"$gate_arms_pass_output"
grep -qF '[fixture-suite] missing_case=next-git-absent' <<<"$gate_arms_pass_output"
grep -qF '[fixture-suite] arms_defined=4 arms_executed=4' <<<"$gate_arms_pass_output"
# Final aggregation truth table: both FAIL emits both failures and is red.
bash "$SUBJECT" init cases "$production_manifest" "$gate_state" "$gate_run"
bash "$SUBJECT" init arms "$production_arms" "$gate_state" "$gate_run"
for case_name in next-git-present main-git-present; do
bash "$SUBJECT" mark cases "$production_manifest" "$gate_state" "$gate_run" "$case_name"
done
for arm_name in greenfield-git-present greenfield-main-git-present greenfield-git-absent; do
bash "$SUBJECT" mark arms "$production_arms" "$gate_state" "$gate_run" "$arm_name"
done
set +e
gate_both_fail_output="$(bash "$GATE" "$SUBJECT" "$production_manifest" "$production_arms" "$gate_state" "$gate_run" 2>&1)"
gate_both_fail_status=$?
set -e
[[ "$gate_both_fail_status" -eq 1 ]]
grep -qF '[fixture-suite] missing_case=next-git-absent' <<<"$gate_both_fail_output"
grep -qF '[fixture-suite] missing_arm=greenfield-remote-installer-contract' <<<"$gate_both_fail_output"
# Final aggregation truth table: both PASS is green.
bash "$SUBJECT" init cases "$production_manifest" "$gate_state" "$gate_run"
bash "$SUBJECT" init arms "$production_arms" "$gate_state" "$gate_run"
mark_all_cases "$gate_state" "$gate_run"
mark_all_arms "$gate_state" "$gate_run"
gate_both_pass_output="$(bash "$GATE" "$SUBJECT" "$production_manifest" "$production_arms" "$gate_state" "$gate_run")"
grep -qF '[fixture-suite] cases_defined=3 cases_executed=3' <<<"$gate_both_pass_output"
grep -qF '[fixture-suite] arms_defined=4 arms_executed=4' <<<"$gate_both_pass_output"
# A future manifest case changes the derived set and cannot silently escape coverage.
printf '%s\n' $'future-case\texit\t1' >> "$manifest"
set +e
@@ -121,7 +202,24 @@ done
arm_complete_output="$(bash "$SUBJECT" check arms "$production_arms" "$production_state" production-1)"
grep -qF '[fixture-suite] arms_defined=4 arms_executed=4' <<<"$arm_complete_output"
# The exact checkout-archive selector excludes coverage state under .mosaic-test-work.
# Bind archive purity to the production checkout-archive selector before proving its effect.
python3 - "$INSTALLER" <<'PY'
import re
import sys
from pathlib import Path
text = Path(sys.argv[1]).read_text(encoding="utf-8")
archive = re.search(
r'(?ms)^ tar -C "\$repo_parent" \\\n(?P<selectors>.*?)^ -czf "\$checkout_archive" "\$repo_name"$',
text,
)
assert archive, "production checkout-archive command not found"
assert "--exclude='*/.mosaic-test-work'" in archive.group("selectors"), (
"production checkout archive does not exclude .mosaic-test-work"
)
PY
# The production-bound checkout-archive selector excludes coverage state under .mosaic-test-work.
: > "$selector_control"
archive="$TMP/checkout.tar.gz"
archive_list="$TMP/checkout.list"
@@ -180,8 +278,9 @@ assert ".greenfield-case-state" not in text
assert ".mosaic-test-work/greenfield-execution-coverage" in text
final = steps["greenfield-case-denominator"]
assert "status: [success, failure]" in final
assert "check cases" in final and "check arms" in final
assert "cases_status" in final and "arms_status" in final
assert "bash tools/verify-greenfield-execution-coverage-gate.sh" in final
assert "check cases" not in final and "check arms" not in final
assert "cases_status" not in final and "arms_status" not in final
for dependency in arm_steps:
assert f" - {dependency}\n" in final, f"final gate can race {dependency}"
PY