I'm sure we would all want to work on a code base with 100% test coverage, but it's not always possible for reasons of budget, legacy laziness, or people not understanding the benefits.
I'm currently re-working a major part of a project at the moment, and am very glad we have high test coverage in it (even if poorly implemented in places). This has let me identify the areas I would be causing to break without having to manually test everything.
The project has a minimum threshold of 95% test coverage, which we're slowly edging down to (we're on about 98 at the moment), but will catch up again in the new year (there's technical debt tasks for them already).
It's accepted that aiming for 100% gets to a point of diminishing returns over a large code base, so what level of coverage do you aim for?
Top comments (0)