Not really a fan of the name here, what about

• prefer-testing-library-methods
• disallow-raw-locators
• ...

Honestly not sure I like those options either, hmm.

@mskelton What about disallow-page-locator-selector? As this is essentially what the rule is doing. It checks for the usage of page.locator() every other method is allowed.

Would definitely use this rule! the prefer- prefix makes sense to me.

Problem with disallow-page-locator-selector is that there is also frame.locator, locator.locator, etc.

I think I would like to proceed with disallow-raw-locators, it makes the most sense of all the options given.

@mskelton That's a good point you're making. I would then change the name of the rule to your suggested disallow-raw-locators. And to be sure that I'm understanding you correctly. Should I also modify the rule in a way that will disallow the usage of frame.locator and locator.locator?

Yes, the rule should detect all usages of .locator, not just page.locator.

Talking about the naming again. I think it would make sense to align it with other rules like no-page-pause or no-useless-await. So the rule would be named no-raw-selector.

Another thing: should this rule be included in the recommended config and with which level of severity should this rule report?

Yeah, good call, no-raw-selector is better.

This should definitely NOT be in the recommended config. While the other methods are better when possible to use, there are still plenty of valid use cases for .locator() and I do not want that warning users all the time unnecessarily.

Ok, I've made the changes you've requested.

🎉 This PR is included in version 0.16.0 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀

Should this rule be added to the list of rules in the readme?

@seanlaff Indeed it should, I missed that before. If you are willing to open a PR, that'd be appreciated!

@seanlaff It's in the readme now