Not sure what applications you tested with RTL, but in a large microfrontends enterprise web application that thing needs so many mocks to work, it’s worse than enzyme. Not to mention having to mock service calls. Add to that, that half the time it throws errors because it thinks something is too slow or too fast. It’s sold like saving tons of development time, but what it actually did, it doubled development time. The documentation is so terrible, Kent had to write another set of docs just to explain how not to use the library wrong. People should just stick to what works: good unit tests with solid E2E. This middle layer of the pyramid Kent is so enamoured with is helping no one.