PayrollManagementSystem/backend/tests/unit/test_audit_serialization.py
liuwenxin c83287819a feat(system): 完成后端系统管理(审计/角色/操作日志/日志查询)
- 操作日志审计机制:audit 上下文(成功与失败都记、独立提交、操作人标量
  固化、commit 后查询前置约定),用户/角色全部写操作接入;request_data
  按键名的结构化脱敏 + UTF-8 字节边界截断
- 角色管理:CRUD 与启停;admin 禁删禁停、有绑定禁删(FOR UPDATE 锁行)、
  编码创建后不可改、停用角色不可再分配
- 日志查询:操作/登录两端点,列表与详情两段式(详情含已脱敏请求参数);
  时间参数统一转本地 naive(修混用时区 500 与偏移丢弃)
- 用户管理:补齐 reset_password 操作人、编辑空串归一 NULL、并发唯一键兜底 400
- 测试 301 例全绿(单元 223 / 集成 78)

Co-Authored-By: Claude Code <noreply@anthropic.com>
2026-10-10 16:28:04 +08:00

129 lines
5.5 KiB
Python
Raw Permalink Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

"""操作日志入库前处理单元测试。
守的是入库前的**截断与脱敏**:
- 截断:各列超长在 MySQL 严格模式下抛 1406,而「审计失败不影响业务」的约定
会把该异常吞掉 —— 整条日志静默消失;request_data 尤甚,中文一个字符
占 3 字节,必须按 **UTF-8 字节边界**切,否则会切出乱码残字
- 脱敏:密码是最容易从日志泄漏的凭据,request_data 入库前必经 scrub
"""
import json
from datetime import datetime
from app.models.operation_log import (
ERROR_MESSAGE_MAX_LEN,
REQUEST_DATA_MAX_BYTES,
SUMMARY_MAX_LEN,
)
from app.repositories.operation_log_repository import _clip, _serialize_request_data
class TestClip:
"""按字符数截断(长度常量与数据库列宽均按字符定义)。"""
def test_超长截断至上限(self) -> None:
result = _clip("x" * (SUMMARY_MAX_LEN + 50), SUMMARY_MAX_LEN)
assert result is not None
assert len(result) == SUMMARY_MAX_LEN
def test_未超长原样返回(self) -> None:
assert _clip("正常长度", SUMMARY_MAX_LEN) == "正常长度"
def test_空值返回None(self) -> None:
assert _clip(None, ERROR_MESSAGE_MAX_LEN) is None
assert _clip("", ERROR_MESSAGE_MAX_LEN) is None
def test_中文按字符数截断不切字(self) -> None:
"""str 切片天然按字符,不会切出半个汉字。"""
result = _clip("中" * 100, 10)
assert result == "中" * 10
class TestSerializeRequestData:
"""序列化:JSON 化 → 脱敏 → 字节边界截断。"""
def test_常规参数原样序列化(self) -> None:
text = _serialize_request_data({"status": 0, "roles": ["user"]})
assert text is not None
assert json.loads(text) == {"status": 0, "roles": ["user"]}
def test_空值返回None(self) -> None:
assert _serialize_request_data(None) is None
assert _serialize_request_data({}) is None
def test_密码值被抹除而键名保留(self) -> None:
"""键名保留便于审计阅读"提交过密码字段",值必须抹掉。"""
for payload in ({"password": "S3cret@2026"}, {"new_password": "S3cret@2026"}):
text = _serialize_request_data(payload) or ""
assert "S3cret@2026" not in text
assert "***" in text
def test_手机号被脱敏(self) -> None:
text = _serialize_request_data({"phone": "13812345678"}) or ""
assert "13812345678" not in text
assert "138****5678" in text
def test_超长中文按字节边界截断且无残字(self) -> None:
"""75000 字节(25000 个汉字)必然超限:
若按字符截断则字节数三倍超标;若按字节硬切可能切出半个字符。
正确结果应是原始 JSON 串的**干净前缀**。"""
payload = {"remark": "中" * 25000}
text = _serialize_request_data(payload)
assert text is not None
assert len(text.encode("utf-8")) <= REQUEST_DATA_MAX_BYTES
assert "<EFBFBD>" not in text # 无替换字符
raw = json.dumps(payload, ensure_ascii=False)
assert raw.startswith(text)
def test_未超长的中文完整保留(self) -> None:
payload = {"remark": "中" * 100}
text = _serialize_request_data(payload)
assert text is not None
assert json.loads(text) == payload
def test_不可JSON序列化的值降级为字符串(self) -> None:
text = _serialize_request_data({"at": datetime(2026, 10, 10, 12, 0, 0)})
assert text is not None
assert "2026-10-10" in text
class TestStructuredSanitization:
"""结构化抹除 —— 按键名整值替换,不依赖文本正则。
回归背景:此前用文本正则脱敏,有三种逃逸形态 —— 值以空白开头时
整体匹配失败(**完全不脱敏**);值含空格/逗号时只抹首段;值含引号时
既残留明文又破坏 JSON 结构。密码校验只查字节数、前后端都不 trim,
这些形态都是正常可用密码。
"""
def test_前导空白密码被完全抹除(self) -> None:
"""回归:此前 `" abc12345"` 原样明文入库。"""
text = _serialize_request_data({"new_password": " abc12345"}) or ""
assert "abc12345" not in text
assert json.loads(text) == {"new_password": "***"}
def test_含空格逗号引号均整体抹除且JSON完整(self) -> None:
for password in ("abc 12345", "abc,def", 'abc"def', "abc&def"):
text = _serialize_request_data({"password": password}) or ""
assert password not in text
assert json.loads(text) == {"password": "***"}
def test_嵌套结构与键名大小写(self) -> None:
text = (
_serialize_request_data({"user": {"New_Password": "s3"}, "list": [{"TOKEN": "t"}]})
or ""
)
assert "s3" not in text
assert json.loads(text) == {"user": {"New_Password": "***"}, "list": [{"TOKEN": "***"}]}
def test_非敏感键不受影响(self) -> None:
text = _serialize_request_data({"keyword": "a_b", "remark": "正常备注"}) or ""
assert json.loads(text) == {"keyword": "a_b", "remark": "正常备注"}
def test_孤立代理字符不再抛错(self) -> None:
"""回归:直接 encode('utf-8') 遇 \\ud800 抛 UnicodeEncodeError,
该请求的审计记录会被兜底逻辑吞掉、静默丢失。"""
text = _serialize_request_data({"username": "a\ud800b"})
assert text is not None