HTML API: Backport updates from Core

[dmsnell]

What?

In this "blessed" patch we're backporting some udpates coming from Core in the Tag Processor, namely a major fix to a rare bug that occurs when making changes and then seeking to an earlier position in a document.

• Preserve internal cursor across updates #4371
• Update code example formatting to fix Developer Resources docs #4419
• Update whitespace formatting to pass Gutenberg linter #4433

Why?

The update fixes a defect in the Tag Processor and we want to avoid triggering that defect.

How?

See the original work in Core (linked above).

Testing Instructions

As a "blessed" PR this should only require basic confirmation of the changes (via code audit) and passing of the existing test suite. The code changes have been reviewed and approved and merged into WordPress.

[dmsnell]

@anton-vlasenko another example in this PR of code that was accepted in Core through the linting but is rejected in Gutenberg. mind taking a look?

[ockham]

another example in this PR of code that was accepted in Core through the linting but is rejected in Gutenberg.

Yeah, curious that this wasn't flagged by CI on WordPress/wordpress-develop#4371. I thought we were using the same WPCS rulesets on GB and wordpress-develop 🤔

[ockham]

BTW once we sort out the linter issue, there's already another formatting commit that we might consider backporting (as part of this PR, or separately).

[github-actions]

Flaky tests detected in 6e9a90d12ea7adfbff58de66f7e2c071941fa5c2.
Some tests passed with failed attempts. The failures may not be related to this commit but are still reported for visibility. See the documentation for more information.

🔍 Workflow run URL: https://github.com/WordPress/gutenberg/actions/runs/4882192923
📝 Reported issues:

• [Flaky Test] shouldn't conflict with server-side autosave #35176 in specs/editor/various/autosave.test.js

[anton-vlasenko]

@anton-vlasenko another example in this PR of code that was accepted in Core through the linting but is rejected in Gutenberg. mind taking a look?

Thanks for the ping, @dmsnell. Yes, this is expected.
It's because Gutenberg does use the WordPress-Docs ruleset, while Core does not.
This makes the backporting process less ideal than it could be.

[dmsnell]

@anton-vlasenko I support your linked suggestion to lift this demand in Gutenberg. also, the changes I've had to make are not particularly better after the lint rule, just different. The linter in this case forced a vertical misalignment (actually it did in the previous change I had to do as well) so from a robotic point of view it now meets all the rules, but from a human standpoint it's less comprehensible 🤷‍♂️

in any case it is frustrating to go through the linting process in Core and then bring back those changes into Gutenberg, which should need no further changes, and be rejected. I sense that you already understand this but wanted to share that's been my experience at least a couple times now.

thanks for looking into this and linking that conversation: I hope we can resolve this soon.