testing.yaml: test_integrity 规则/阈值声明（ADR-0035 / #86，P2-1）

[randypanding]

摘要

P2-1（#86 / ADR-0035）的 policy 面：governance/policy/testing.yaml 新增 test_integrity 节——CI-Workflows test-integrity.yml（配套 PR CI-Workflows#17）运行时拉取本节覆盖检测器内置缺省。

• 四条规则及严格度：TI-R1 测试文件删除 red / TI-R2 断言净下降 red / TI-R3 新增抑制标记 red / TI-R4 期望值改写 require_adr（调 red 则不可豁免）
• 模式（ERE）四组：test_file / assertion / suppression / non_source——与检测器内置缺省逐字节一致（双源同值校验已过，本地 python 比对 MATCH×4）
• escape_hatch：ADR-NNNN 引用式豁免+计数入账（P3-2 [自动合并 P3-2] C1 拆分 core/rest + ADR 实质校验 + 破玻璃自动 revert #96 落地后升级 scope 匹配）
• fail_closed: true——policy 拉取失败=红
• active_now 登记 T-12 test_integrity（placement: gate）

验证清单

• YAML 全量解析通过（safe_load；rules/patterns 结构断言）
• patterns 与 CI-Workflows scripts/test-integrity.sh 内置缺省逐字节一致（MISMATCH=0）
• 不触碰 governance/expected-state.json（留给主 agent 集成）
• 本仓 gate 绿（YAML 全量解析 + adr-required：ADR-0035 已合入 agent-registry）

[coderabbitai]

Warning

Review limit reached

@randypanding, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 8 minutes

Limit details: You’ve used all 10 included reviews currently available.

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

Wait for the limit to reset, then comment @coderabbitai review or push new commits to the PR.

An organization admin can change what happens after included review limits in Billing.

How do review limits work?

CodeRabbit enforces per-developer PR review limits within each organization.

For paid Pro and Pro+ reviews, CodeRabbit uses a developer's included PR review attempts over the past 7 days to set the current hourly allowance. At typical activity levels, the full plan allowance applies. Higher sustained activity can lower the allowance until earlier attempts leave the 7-day window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: af31f4f8-941b-4f8c-9f8d-d09beed0d039

📥 Commits

Reviewing files that changed from the base of the PR and between 1355046 and c9cc61b.

📒 Files selected for processing (1)

• governance/policy/testing.yaml

---

_{Comment @coderabbitai help to get the list of available commands.}

[qodo-code-review]

PR Summary by Qodo

Declare a fail-closed test integrity gate policy

✨ Enhancement ⚙️ Configuration changes 🕐 10-20 Minutes

AI Description

• Registers test integrity as an active pull-request gate.
• Defines four integrity rules with red or ADR-required enforcement.
• Supplies cross-language patterns and fails closed when policy retrieval fails.

Diagram

graph TD
  A["Testing policy"] --> B["CI workflow"] --> C["Integrity detector"] --> D{"Rule triggered?"}
  D -->|No| E["Gate passes"]
  D -->|Red| F["Gate fails"]
  D -->|ADR required| G["ADR waiver"] --> E
  B -->|Fetch failure| F

Loading

High-Level Assessment

The following are alternative approaches to this PR:

1. Use language-specific AST detectors

• ➕ Distinguishes semantic assertion changes more accurately than regular expressions
• ➕ Reduces false positives from comments, strings, and formatting-only edits
• ➖ Requires separate parsers and maintenance paths for every supported language
• ➖ Adds substantial complexity beyond the cross-language P2-1 scope
• ➖ Makes centralized rule portability and byte-for-byte fallback synchronization harder

Recommendation: Keep the declarative, language-neutral ERE policy for P2-1. It centralizes governance, supports the existing shell detector, and provides conservative fail-closed enforcement; AST-based detection is better treated as a later precision enhancement once false-positive data justifies the added complexity.

Files changed (1) +28 / -0

Other (1) +28 / -0

testing.yamlDefine the T-12 test integrity gate policy +28/-0

Define the T-12 test integrity gate policy

• Registers test integrity as an active gate and declares four rules covering test deletion, assertion decline, suppression additions, and expectation rewrites. Adds detector-compatible ERE patterns, ADR-based waiver semantics, and fail-closed policy retrieval behavior.

governance/policy/testing.yaml

[qodo-code-review]

Code Review by Qodo

🐞 Bugs (1) 📘 Rule violations (0) 📜 Skill insights (0)

1. Uppercase assertions evade TI-R2 🐞 Bug ≡ Correctness

Description

TI-R2 的大小写敏感正则无法匹配 ASSERT_EQ、EXPECT_EQ、REQUIRE 等常见全大写 C/C++ 断言宏。删除这些断言不会减少检测器统计值，因此可绕过声明为 red
的断言净下降检查。

Code

governance/policy/testing.yaml[87]

+    assertion: '\bassert|\bexpect\(|\bexpect\.|\brequire\(|\brequire\.|\bshould\b|\bt\.Error|\bt\.Fatal|\bself\.assert|\bfail_if\(|\bAssert|\bExpect\(|\bRequire\(|\bSo\('

Relevance

●●● Strong

This is a deterministic correctness gap in a machine-enforced regex that permits assertion-count
bypasses.

PR-#50
PR-#19

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The policy declares regex-level language-independent detection and makes a negative assertion net
red, but the expression contains only lowercase or initial-capital alternatives. Since ERE matching
is case-sensitive here, conventional all-uppercase assertion macros are not counted.

governance/policy/testing.yaml[69-69]
governance/policy/testing.yaml[79-80]
governance/policy/testing.yaml[85-87]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
TI-R2's assertion ERE omits conventional all-uppercase assertion macros, allowing assertion deletion to evade the red gate.

## Issue Context
The policy describes the detector as language-independent and defines any negative assertion net as red. Preserve compatibility with GNU ERE and synchronize the corrected expression with the CI-Workflows built-in default and its tests.

## Fix Focus Areas
- governance/policy/testing.yaml[79-87]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

2. T-12 severity contradicts policy ⊘ Outdated 🐞 Bug ≡ Correctness

Description

The active_now entry states that all four tampering forms fail directly red, while TI-R4 is
explicitly require_adr and therefore ADR-waivable. This gives policy readers or consumers of the
summary entry a stricter rule than the detailed machine-executable policy actually enforces.

Code

governance/policy/testing.yaml[25]

+     placement: gate, rule: "测试篡改四形态直接红（ADR-0035）", note: "P2-1 第一必需门"}

Relevance

●●● Strong

Recent accepted governance reviews prioritize aligning summaries, executable policy, and waiver
semantics.

PR-#48
PR-#19

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The newly added T-12 summary says all four forms are directly red, but the same added policy
documents ADR-based waivers and assigns TI-R4 require_adr, unlike the red severity assigned to
TI-R1 through TI-R3.

governance/policy/testing.yaml[24-25]
governance/policy/testing.yaml[70-72]
governance/policy/testing.yaml[77-84]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The T-12 active-policy summary says all four rules are directly red, contradicting TI-R4's `require_adr` severity and documented escape hatch.

## Issue Context
Keep the active registry description consistent with the authoritative machine-executable rule semantics so humans and downstream consumers receive the same policy.

## Fix Focus Areas
- governance/policy/testing.yaml[24-25]
- governance/policy/testing.yaml[70-84]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

Context sources

✅ Compliance rules (platform): 8 rules

✅ Web pages:

  🌐 .github/workflows/core-integrity.yml

  🌐 .github/workflows/core-integrity-call.yml

  🌐 .github/workflows/ci.yml

  +2 more

Review mode: ⚖️ Balanced: 这会改变 gate 的机器可执行策略、阈值与 fail-closed 行为，属于有实际行为和 CI 守门影响的配置变更；范围虽小但不适合降为 lite，且逻辑集中、无需 extended 冗余审查。

Tip of the day

💡 Did you know, you can copy the agent prompt from any finding and feed it to your IDE agent

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

[qodo-code-review]

1. Uppercase assertions evade ti-r2 🐞 Bug ≡ Correctness

TI-R2 的大小写敏感正则无法匹配 ASSERT_EQ、EXPECT_EQ、REQUIRE 等常见全大写 C/C++ 断言宏。删除这些断言不会减少检测器统计值，因此可绕过声明为 red
的断言净下降检查。

Agent Prompt

## Issue description
TI-R2's assertion ERE omits conventional all-uppercase assertion macros, allowing assertion deletion to evade the red gate.

## Issue Context
The policy describes the detector as language-independent and defines any negative assertion net as red. Preserve compatibility with GNU ERE and synchronize the corrected expression with the CI-Workflows built-in default and its tests.

## Fix Focus Areas
- governance/policy/testing.yaml[79-87]

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools