• This is not a v0.x issue.
• I have searched the issues of this repository and believe that this is not a duplicate.

This was a previously #9492 however it does not actually provide a fix for the core problem which is that in testing environments we should be able to mount(<Component>) that are created from withStyles() and have snapshots be reliable without wrapping them.

Expected Behavior 🤔

Snapshots should create reliable labels for withStyles() used by both custom and defaut material-ui components like in the red portion of this Snapshot:

Current Behavior 😯

One can modify a custom component to have

withStyles(Styles, { generateClassName: (rule, styleSheet) => (`${styleSheet.options.classNamePrefix}-${rule.key}`)  
})

However, this only modifies overwritten styles in material UI, default ones are actually not modified at all:

1. Make a withStyles() that uses MUI components as well
2. Install a testing framework like jest
3. Run snapshots
4. Add a new test, and you will see how the snapshots will fail on prior ones because the ruleCounter increases.

Context 🔦

I would like to be able to mount() and call Jest's native .toMatchSnapshot() without worrying about the ruleCounter adding random numbers. It is not appropriate to wrap every object in a as stated in the docs of https://material-ui.com/customization/css-in-js/#creategenerateclassname-options-class-name-generator because tests should mirror their native implementation.

Your Environment 🌎

Ma