Possible regression in 7.2

Describe the bug One of our tests started to fail when trying to upgrade this library to 7.2.0, then 7.2.1 and it kept failing. The test ensures that a field is in error state when an error condition triggers, then turn back to normal state when error condition is cleared.

To Reproduce Nothing really fancy, except that we manually call setError(field) / clearErrors(field), i.e.: manual validation.

Codesandbox link (Required) I will try to provide a codesandbox later, but I do provide all the details above.

V7

• JS Template
• TS Template

V6

• useForm Template
• useFieldArray Template
• Controller Template

Expected behavior The test does not fail.

Desktop (please complete the following information):

• OS: Linux
• Browser: jsdom (jest tests)
• Version: latest

Additional context I see there’s a change regarding validation in 7.2, this might be related.

[ddolcimascolo]

@bluebill1049 Sorry for the delay, I managed to reproduce in a CSB https://codesandbox.io/s/condescending-volhard-3xj87

Submit with an empty field, the field turns to error state. Type-in some characters, the error state does not automatically reset. If you change the version of RHF to 7.1.1 it resets.

I’m a bit confused now by our call to clearErrors that might not be required at all, but that’s another story 😕

David

[NickR49]

@bluebill1049 I’ve fixed the issue I was having. It was where my component was doing an unnecessary call to clearErrors(name) before then calling setError(name) in the next render (my component in this case was meant to start in an error state). I didn’t think it would matter but this was causing my component to not be listed in the errors object. Removing the unnecessary call to clearErrors (even though it happened before setError) meant that it was correctly in the errors object.

CSB - https://codesandbox.io/s/dreamy-galileo-opyt3?file=/src/App.js

Update - oddly it’s indeterminate. Refresh the CSB a few times and sometimes the error shows correctly and sometimes it doesn’t