Serializer: Remove HTML beautification, preserve whitespace

[aduth]

Supersedes #6716
Fixes #4456

This pull request seeks to remove in-block HTML beautification. Per intent with #633, there are still newlines between blocks. Within a block, however, the default serialization will not inject any of its own automated whitespace or tabbing. Existing intentional whitespace will be preserved (e.g. code blocks, shortcode blocks).

This has a few benefits:

• Performance
  • js-beautify accounts for 48kb (18.1kb gzipped), now removed
  • No longer need to parse / rewrite HTML when saving a post during serialization stage
• Accuracy
  • children parse would include (often wrongly) text nodes not always intended but automatically injected by beautifier
  • No more collapsing whitespace in shortcode blocks
    • Verified in updated adding-blocks E2E test which would have previously failed in master (see bd78d55)
  • Surfaces potential issues with raw handler not stripping newlines where stripping may be intended (see new test/integration/fixtures output files, e.g. apple-out.html table block)

Implementation notes:

The vast majority of affected file changes are fixture updates. The only meaningful changes are those made within blocks/api/serializer.js, though fixtures verification would be good as well.

Testing instructions:

Verify there are no issues with serializing and restoring block contents.

Examples:

• Save a Demo post and reload it, ensure no validation errors
• Save a post including a Columns block and reload it, ensure content preserved and no validation errors

[danielbachhuber]

Can you include a test specifically for #4456? My understanding is that we'd need to change the Shortcode Block <input> to <textarea> to properly display multi-line shortcodes.

Also, it'd be great if you merged in #7846 to incorporate those tests.

[aduth]

Rebased against master.

@danielbachhuber The tests added in 6ae7a11 specifically target multi-line shortcode (the shortcode block is already a textarea field). Is this what you had in mind; or something different?

[danielbachhuber]

(the shortcode block is already a textarea field)

Ok, great.

Is this what you had in mind; or something different?

Could you add one more test for segmentHTMLToShortcodeBlock?

[aduth]

I don't think that falls within the scope of what this pull request is targeting. Raw handling is a different feature than the serialization beautifier.

[danielbachhuber]

I don't think that falls within the scope of what this pull request is targeting. Raw handling is a different feature than the serialization beautifier.

Ok. I still think it would be a helpful test to have to truly fix #4456. Would you like me to add the test when I have a moment?

[aduth]

That's fine. I guess I'm unclear about what's specifically not covered already to address #4456 . The classic editor content issue is already fixed (on master), and this one resolves issues with shortcode block (with tests). There's nothing in the original report that touches segmentHTMLToShortcodeBlock .

[danielbachhuber]

@aduth Ok. I'll defer to your better judgement for this PR, and follow up with a second PR for anything I think of.

[mtias]

And such was the way beauty left our world...