Article 4 of 5
Bug Report
A failed result that someone else can act on.
A failure is only useful if a developer can reproduce it. The bug report is where “it does not work” becomes an address: these steps, this input, this environment, this difference between expected and actual.
In this system a bug report starts as the comment you write when you mark a case Failed. Bugs are created after the test run: when it is finished, Report bug on a failed result lets you either report a new bug, pre-filled from the case and your comment, or link a bug that already exists. Either way the bug stays linked to the result. Write it as if the reader has never seen the case.
What it is made of
- Summary
- What is broken and where: “Insufficient funds error shows generic text on checkout”.
- Steps to reproduce
- The shortest path to the failure, numbered.
- Expected vs Actual
- The two lines that make it a bug rather than an opinion.
- Environment
- Build, browser, OS, account, test data.
- Severity
- How badly it breaks the product: blocker (nothing works, testing stops), critical, major, minor, trivial.
- Priority
- How soon it must be fixed: P0 now, P1 in this release, P2 when there is time. Severity and priority are different questions and often have different answers.
- Status
- Where the bug is in its life: new until someone deals with it, then fixed or rejected. A fixed or rejected bug that comes back is reopened, and goes on to fixed or rejected again.
Rules that save you later
- One bug per report. Two bugs in one report means one of them gets forgotten.
- Attach evidence: screenshot, log, request id. Evidence turns an argument into a fact.
- Check it reproduces before reporting it. “Happened once” is a note, not a bug report yet.
- Search before you report. If the bug already exists, link the failed result to it instead of filing a duplicate.