[ButtonBase] Unfocus does not clear ripple

[siriwatknp]

keyboard tab between ButtonBase component does not clear the focus ripple. I can confirm that revert the code on this commit does not have this issue (in ButtonBase.js). https://github.com/mui-org/material-ui/commit/5f30983bfa16195237fde55a78d5e43b151a29fa

cc @michaldudak can you help take a look?

• [x] The issue is present in the latest release.
• [x] I have searched the issues of this repository and believe that this is not a duplicate.

Current Behavior 😯

Expected Behavior 🤔

it should clear the ripple element

Steps to Reproduce 🕹

Steps:

1. open https://next.material-ui.com/components/buttons/#main-content
2. tab through the docs

Context 🔦

Your Environment 🌎

`npx @mui/envinfo`

``` Don't forget to mention which browser you used. Output from `npx @mui/envinfo` goes here. ```

[michaldudak]

Yes, I've noticed that and I'm already working on a fix.

[eps1lon]

I think I've talked repeatedly about not doing the abstraction. The ButtonBase is integral to the codebase and shouldn't be jeopardized by the unstyled effort. The unstyled effort can be a separate experiment that doesn't need to affect the core product.

[oliviertassinari]

The ButtonBase is integral to the codebase and shouldn't be jeopardized by the unstyled effort

@eps1lon If you could expand on the why, I think that it would be great.

Looking at this regression, it seems that one of the values of using unstyled in the material design components is that it allows surfacing problems earlier. For instance, maybe we missed a test case, or maybe we didn't use the best abstraction for useButton or maybe it's something completely unrelated.

[eps1lon]

@eps1lon If you could expand on the why, I think that it would be great.

Why do I need to explain that a fundamental component should not be affected by experiments? Like what would you expect to hear?

[michaldudak]

I think I've talked repeatedly about not doing the abstraction.

You mean not to implement Button with useButton? If so, I suppose I may have missed that. In the useButton PR, both @mnajdova and @oliviertassinari suggested using the new hook in the ButtonBase and I haven't seen any objections from your side.

As for the issue itself, I already pinpointed it and I'm adding few tests to ButtonBase to help to avoid such problems in the future.

[siriwatknp]

I remember we talked about this (not sure exactly which meeting) that if the unstyled abstraction does not introduce breaking change to the core, we will do it in the core.

[michaldudak]

As discussed today in the planning, we will restore the original implementation and create a separate private package with Material components (no matter if they have breaking changes or not) built using @mui/core.