Is there any way to prove that changed implementation does not introduce regression?

Question

There are two extremities:

• Test the whole system after any change of the code.
• Do not test at all.

Here under 'Testing' I mean running all not automated tests. User Acceptance tests to be more precise.

I'd like to have a solid understanding when it is absolutely safe not to perform manual acceptance tests.

100% code coverage is not sufficient here, I believe.

Answer 1

Well, it seems you mix terms. Breaking system's behavior does not mean system isn't passing acceptance tests, and also you can ruin UAT without breaking system, if you have a requirement of performance, or some visual, or UX stuff.

If you are talking about regression - that previously passed UAT will still pass, than they should be automated as much as possible. QA's always have test plans for regression on different environments, they can be automated to even comparing screenshots on different resolutions, like in facebook.

If you are talking about new functionality and it's UAT, than you can formalize and automate it before implementing, like cucumber approach.

The other way is to test on users, like yandex, or mail. You show users, or company employees knew version, and if you don't collect errors, or complains, you are probably fine. But that's not something you will do for each commit, and if it's an ap, or a desktop app, things can get more tricky