How did userEvent work before v14 and the #setup method?

[stephenpassero]

I'm installing RTL on a new project, and saw that as of v14, there's this new #setup method that is recommended to be used. From the documentation here, it looks this "appl[ies] workarounds and mock[s] the UI layer to simulate user interactions like they would happen in the browser."

This is all well and good, but how did userEvent v13 do this? Did it somehow just not mock the UI layer?

Also kinda a sidenote question, why is everything async now?

[ph-fritsche]

Previously there wasn't one place for such abstraction but multiple places where workarounds were applied trying to match the browser behavior.
The need for a centralized abstraction of a UI layer is explained #714 and further detailed in #744 .

As this is open source, you can always dig through the code (of the previous version).

Some APIs were already async before and some new implementations required asynchronous internals.
Always returning a Promise simplifies how we handle those and allows us to change how we handle things internally without further breaking changes. It also eliminates problems with frameworks applying changes asynchronously. For more information see #504 and the links in that thread.

[stephenpassero]

Ok, thanks! That really helped answer my questions