Fix: regression caused by triple click fix

[bytrangle]

fixes #4490

Description
Introducing fix for triple-click selection in this PR caused some regressions that I didn't expect.

This new PR will alleviate that.

Issue
Some of the issue caused by the triple click fix include, but not limited to:

• Crash when selecting void nodes Crash during mouse selection on Chrome over multiple nodes #4490
• Shift + down arrow then delete text doesn't delete it on Chrome, shift-arrow down then delete doesn't delete text #4496

Context
There are two problems with my triple-click fix:

• The condition for checking if an action is a triple-click is too broad.
• The logic for resetting focusNode of a selection is not fail-proof. The building block of the content component in a Slate editor are HTML element with attribute data-slate-node="element". They may contain child elements data-slate-string, but not always. The current logic assumes that Element always contains String elements, which leads to the crash when selecting void nodes.

Checks

• The new code matches the existing patterns and styles.
• [] The tests pass with yarn test.
• The linter passes with yarn lint. (Fix errors with yarn fix.)
• The relevant examples still work. (Run examples with yarn start.)
• You've added a changeset if changing functionality. (Add one with yarn changeset add.)

[changeset-bot]

🦋 Changeset detected

Latest commit: 8ebf4c7

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

This PR includes changesets to release 1 package

Name	Type
slate-react	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

[bytrangle]

Thank you @dylans for the edit. Are the failed tests acceptable? They passed on my local machine before the addition of changeset.

[dylans]

Thank you @dylans for the edit. Are the failed tests acceptable? They passed on my local machine before the addition of changeset.

They appear to be timeouts in the CI/CD testing environment, so I’m not going to block based on that.

[TheSpyder]

Thanks for doing that quickly after it was discovered! I have a small request.

This is a fix for a change that wasn't released, so it doesn't need a changeset. The original fix did need a changeset but one wasn't added. As such, your one entry in the changelog now looks a bit weird.

Would you mind creating a new PR that removes this changeset and adds one for the #4455 fix? I would do it myself but then you'd lose credit.