[RFC] feat(test-runner-visual-regression): make "update" strict and add "buildCache" for more repeatable tests

[Westbrook]

What I did

1. Make the update property in the VRT config "strict". It will no only update if set to true, allowing newly added regressions (tests without a baseline) to fail the test pass.
2. Add the buildCache property to the VRT config. This has the test save all of the current test output in one place so that it is easier to create a new cache from updated tests listings, including when tests have been intentionally removed.

Questions

1. Using update strictly means that a user with new screenshots can't add those shots while testing if existing tests fail. Is this a common use case? Should we add an update: 'auto' option? Or possibly a strict flag so that the current functionality can be accessed without breaks, but users can alter their workflow to the new version if desired?
2. buildCache currently leverages the "failed" workflow. In that way should we update the naming conventions throughout to the saveCurrent or similar to clarify the usage of the code throughout?
3. Placing "current" next to "failed" may make it more difficult for a user to parse the output of their tests. Should these two outputs be treated differently?

[changeset-bot]

🦋 Changeset detected

Latest commit: 805e731

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 4 packages

Name	Type
@web/test-runner-core	Patch
@web/test-runner-visual-regression	Minor
@web/test-runner	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[LarsDenBakker]

Looks good to me!