I keep seeing bugs described as regressions. Does that just mean an old bug came back, and how do teams try to stop that from happening?
What does a regression mean in software testing?
Answers
A regression is when something that used to work stops working after a change. That can be a previously fixed bug returning, but it can also be a new problem in a feature that worked before.
For example, a change to a checkout screen might accidentally break an existing discount calculation. The change doesn't have to be related to the discount on purpose.
Reply to this answer
When a team fixes a bug, they can add a regression test that reproduces the original conditions and checks for the correct result. If a later change brings the problem back, the test should fail.
Those checks are often automated and run alongside other tests before a release. The old bug report can help describe the setup and expected behavior, though the test still needs to be maintained as the software changes.
Reply to this answer
A test doesn't have to be automated to be useful. A team might keep a particular item, account setup, or other unusual piece of test data around because it once exposed a bug. Someone can check it again after an update.
That kind of safeguard is a reminder to exercise a case that might otherwise be easy to miss. It only catches the problem if someone actually checks it, though.
Reply to this answer
Passing regression tests doesn't prove an update is bug free. Tests cover the cases people thought to check, and changes can cause trouble somewhere nobody expected.
Automation can repeat known checks quickly. Human testers can also explore unexpected behavior and judge whether the checks still make sense. Tools that use AI may help with testing, but they don't make reviewing the results unnecessary.
Reply to this answer
Add your answer
Answers are reviewed before publication.
GBAcartSlot+Sardine:
So it doesn't have to be a bug the team has seen before. That distinction helps.