知识卡片
代表性回放结合模糊输入
少数用户特定输入必现的错误可被总体低错误率掩盖。历史输入回放保留真实使用分布,需覆盖地区、语言、时间与规模;模糊测试探索尚未出现的边界,并可生成满足结构约束的随机数据。样本永远不是穷尽证明,收集时还需脱敏与权限控制,不能把历史未见当作永不发生。
这张卡片看不懂?
这张卡片的反馈已满 10 条,暂时无法继续提交。
已有反馈 · 0 / 10 条
还没有人提交反馈。
在整书笔记中的位置
原书出处 · 1 处
- Chapter 16. Common Failure Patterns · 查看原书依据
High-quality coverage in testing only comes when your test data is representative of all of your users. The second solution is to use fully randomized or “fuzzed” data for your inputs. These tests can be significantly more comprehensive, but they are also much harder to create. For a
参考来源
- 位置:《Designing Distributed Systems, 2nd Edition》第16章《Chapter 16. Common Failure Patterns》(来源项目 book-sources;源文件:extracted/Designing Distributed Systems, 2nd Edition/_epub-src/OEBPS/ch16.html)
- 结论依据:原文区域字符、时段与随机合法输入说明两种覆盖手段的互补。
- 原始内容:High-quality coverage in testing only comes when your test data is representative of all of your users. The second solution is to use fully randomized or “fuzzed” data for your inputs. These tests can be significantly more comprehensive, but they are also much harder to create. For a