feat(pre-push-safety-net): cover combined push failure paths

This commit is contained in:
Marius Ciepluch
2026-09-13 18:43:13 +02:00
parent fa074dd197
commit 4e3b19da5b
4 changed files with 173 additions and 7 deletions
@@ -0,0 +1,37 @@
# Group 8 implementation and qualification
The maintainer authorized group 8 on 2026-09-13. Fresh agent dispatch failed with
`agent thread limit reached`; the explicit manual Kiro fallback applies. Reviews
are main-context reviews, not independent reviewer evidence. Joint rollout gates
remain unchanged; implementation progress does not imply capability eligibility.
## Task 8.1 brief
Requirements 1.1–1.4, 2.2, 2.3, 4.2, 4.4, 5.2, 5.5, 7.1, 7.4 and 8.2;
design: Snapshot Adapter, Cache/Engine Boundary, Push runner and Result/Hook Adapter.
Combine immutable non-HEAD/tag/force/new-branch/deletion updates, unusual paths,
external boundaries, physical vendor exclusion, census failure and interrupted report
publication. Use production transaction/backend seams; preserve source/index/refs,
truthful scope and prior artifacts, and prove worker cleanup. Tests only: no production
behavior or feature toggle introduced. Existing focused suites retain the full cache,
missing-base, hook-exit and ordinary-scan regressions.
## Status Report
- STATUS: READY_FOR_REVIEW
- TASK: 8.1
- RED_PHASE_OUTPUT: N/A: tests-only integration coverage; no production behavior changed.
- TESTS_RUN: Three added combined controls pass. Full suite 1299 passed, 9 skipped in 117.83s; ruff, formatting and mypy pass.
## Review Verdict
- VERDICT: APPROVED
- TASK: 8.1
- MECHANICAL_RESULTS: actual new tests inspected, full regression suite passes, no placeholders or secrets; test-only integration boundary maintained.
- FINDINGS: mixed push control verifies all three exact comparisons and four dispositions, readable snapshot bytes, live state preservation and external-content exclusion. Failed census control inspects physically filtered source before returning missing answers. Interrupted writer control checks previous bytes and absence of its process PID.
- SUMMARY: Combined failures remain explicit and do not masquerade as clean checks.
## Verification Result
- STATUS: VERIFIED
- CLAIM_TYPE: TASK
- CLAIM: Task 8.1 integration regression coverage delivered.
- EVIDENCE: `/tmp/group8-1-focused.log`: 3 passed; `/tmp/group8-1-full.log`: 1299 passed, 9 skipped; static checks exit 0.
- GAPS: These controlled failures do not qualify real detector capabilities or representative latency.
+13 -6
View File
@@ -1,7 +1,7 @@
{
"feature_name": "pre-push-safety-net",
"created_at": "2026-09-12T15:35:00Z",
"updated_at": "2026-09-13T16:03:14.583834+00:00",
"updated_at": "2026-09-13T16:41:21.889949+00:00",
"language": "en",
"phase": "implementation",
"approvals": {
@@ -19,7 +19,7 @@
}
},
"ready_for_implementation": true,
"implementation_approval_note": "Groups 1\u20137 implemented and verified (23/27). Group 7 authorized as a whole; manual Kiro review fallback after fresh dispatch thread limit. Multi-ref input, explicit hook integration and optional bounded witness selection verified. Default capability registry remains empty and rollout remains NO-GO; group 8 pending.",
"implementation_approval_note": "Group 8 authorized and in progress; groups 1\u20137 verified. Capability admission and rollout remain NO-GO until the declared evidence passes.",
"depends_on": [
"contributor-scan"
],
@@ -75,6 +75,11 @@
"at": "2026-09-13T15:45:15.902603+00:00",
"source": "ok, continue with group 7 as a whole",
"action": "Authorized tasks 7.1\u20137.3; manual Kiro fallback after fresh dispatch hit thread limit. Group 8 and rollout remain pending."
},
{
"at": "2026-09-13T16:41:21.889949+00:00",
"source": "move on to group 8",
"action": "Authorized tasks 8.1\u20138.4 and their qualification work; preserve joint gates and record NO-GO wherever prerequisites or measurements do not pass. Fresh dispatch hit thread limit; manual Kiro fallback."
}
],
"task_review": {
@@ -110,10 +115,11 @@
"6.4",
"7.1",
"7.2",
"7.3"
"7.3",
"8.1"
],
"remaining_tasks": 4,
"verification_record": "implementation-group-7.md",
"remaining_tasks": 3,
"verification_record": "implementation-group-8.md",
"verification_records": [
"implementation-group-1.md",
"implementation-group-2.md",
@@ -121,7 +127,8 @@
"implementation-group-4.md",
"implementation-group-5.md",
"implementation-group-6.md",
"implementation-group-7.md"
"implementation-group-7.md",
"implementation-group-8.md"
],
"next_group": 8,
"blocked_tasks": [],
+1 -1
View File
@@ -212,7 +212,7 @@ independent capability qualification. Groups 7–8 remain pending; rollout remai
- _Requirements: 3.5, 4.1, 4.2, 7.1, 7.2, 7.3_
- [ ] 8. Validate usefulness, runtime and cross-language admission
- [ ] 8.1 Exercise complete push failure and regression paths
- [x] 8.1 Exercise complete push failure and regression paths
- Combine dirty/non-HEAD/multi-ref/new-branch/force/tag/deletion pushes with unusual paths, external content boundaries, failed census, missing comparisons and interrupted persistence.
- Verify no orphan children, no overwritten live state, no vendor code in graphs, exact scope reports and no false clean result from skipped work.
- Done when the integration suite covers the critical flows through the real CLI/backend seams and preserves ordinary scan behavior.
+122
View File
@@ -0,0 +1,122 @@
"""Combined transaction controls, not detector qualification."""
import json
import multiprocessing
import os
import time
from test_push_runner import history
from openultrasast.config import PushConfig
from openultrasast.cpg.backend import NullBackend
from openultrasast.push.runner import push
def test_mixed_transaction_retains_exact_objects_and_external_boundaries(tmp_path):
root, base, head, git = history(tmp_path)
git("tag", "-a", "release", "-m", "release", head)
tag = git("rev-parse", "release")
# Non-HEAD push while the live tree contains unrelated unusual paths and a symlink.
path = root / "space ü\tname.js"
path.write_text("function route(req) { return req.body; }\n")
(root / "outside.js").symlink_to(tmp_path / "outside-secret")
(tmp_path / "outside-secret").write_text("must never be read by analysis")
git("add", "--", path.name, "outside.js")
git("commit", "-m", "external boundary")
tip = git("rev-parse", "HEAD")
before = (root / ".git/index").read_bytes(), git("show-ref"), git("status", "--porcelain")
zero = "0" * 40
lines = (
f"refs/tags/release {tag} refs/tags/release {base}\n"
f"refs/heads/rewind {base} refs/heads/rewind {head}\n"
f"refs/heads/new {tip} refs/heads/new {zero}\n"
f"(delete) {zero} refs/heads/deleted {head}\n"
)
artifact = tmp_path / "mixed.json"
result = push(
root,
updates=lines,
remote_name="origin",
remote_url="/local",
artifact=artifact,
backend=NullBackend(),
config=PushConfig(comparison_base=base),
)
data = json.loads(artifact.read_text())
assert len(data["resolution"]["updates"]) == 4
assert len(result.result.analyses) == 3
assert [(a.comparison.head_oid, a.comparison.base_oid) for a in result.result.analyses] == [(head, base), (base, head), (tip, base)]
assert result.result.coverage_status != "complete_within_scope"
assert any(m["boundaries"] for m in data["snapshots"])
assert all(m["bytes_read"] > 0 for m in data["snapshots"])
assert "must never be read by analysis" not in artifact.read_text()
assert ((root / ".git/index").read_bytes(), git("show-ref"), git("status", "--porcelain")) == before
def test_interrupted_writer_preserves_previous_artifact_and_reaps_child(tmp_path, monkeypatch):
import openultrasast.push.report as reporting
root, base, head, _ = history(tmp_path)
artifact = tmp_path / "result.json"
artifact.write_text("previous completed artifact")
pidfile = tmp_path / "writer.pid"
def interrupted(*args):
pidfile.write_text(str(os.getpid()))
time.sleep(10)
monkeypatch.setattr(reporting, "_write_artifact", interrupted)
before = {p.pid for p in multiprocessing.active_children()}
start = time.monotonic()
result = push(
root,
updates="",
remote_name="origin",
remote_url="/local",
artifact=artifact,
config=PushConfig(deadline_seconds=0.2, cancellation_allowance_seconds=0.4),
)
assert time.monotonic() - start < 0.8
assert result.error == "report_deadline_exhausted"
assert artifact.read_text() == "previous completed artifact"
assert result.result.coverage_status == "not_applicable"
assert {p.pid for p in multiprocessing.active_children()} == before
assert pidfile.exists()
assert not os.path.exists("/proc/" + pidfile.read_text())
def test_failed_census_after_physical_vendor_filter_is_not_clean(tmp_path):
from openultrasast.cpg.backend import CpgResult
root, base, head, git = history(tmp_path)
vendor = root / "node_modules/dependency"
vendor.mkdir(parents=True)
(vendor / "vendor.js").write_text("function vendor(req) { eval(req.body); }\n")
git("add", "--", "node_modules")
git("commit", "-m", "vendor boundary")
tip = git("rev-parse", "HEAD")
reads = []
class FailedCensus:
def build(self, source, **kwargs):
files = [p for p in source.rglob("*") if p.is_file()]
assert files
assert not any("node_modules" in p.parts for p in files)
reads.extend(len(p.read_bytes()) for p in files)
return CpgResult(tmp_path / "graph", lambda *_: None, lambda *_: {})
artifact = tmp_path / "census.json"
result = push(
root,
updates=f"refs/heads/main {tip} refs/heads/main {base}\n",
remote_name="origin",
remote_url="/local",
artifact=artifact,
backend=FailedCensus(),
)
data = json.loads(artifact.read_text())
assert reads and all(n > 0 for n in reads)
assert result.result.coverage_status in ("unavailable", "incomplete")
assert result.result.finding_status == "none"
assert data["admission"]["coverage_reasons"]
assert not any(a.completed_questions for a in result.result.analyses)