Remove span tag from the basic HTML format in the standard profile

Problem/Motivation

Out of the box, pasting content from Microsoft Word or Google Docs leads to unexpected span elements.

In ckeditor 4, these span elements are deeply nested. For example, a single h2, even after allowing ckeditor to detect and modify that the content was pasted from Word, leads to:

<h2><span><span><span><span><span><span><span><span>This is Heading 2</span></span></span></span></span></span></span></span></h2>

In ckeditor 5, it's better, but still adds a language attribute that in most cases isn't needed:

<h2>
    <span lang="EN-US">This is Heading 2</span>
</h2>

Looking back through the git and issue history, it looks like span was added at #1911884-30: Enable CKEditor in the Standard install profile.

The issue doesn't say, but I asked @quicksketch offline and he said:

I think it was primarily because it was needed for underline styling, since <u> tags were frowned upon at the time. Underlines are <span class="underline">.

As well, #1322906: Allow span tag by default for language changes? was marked as fixed with no commits to support span tags for marking language.

However, today in both Drupal 9 and Drupal 10:

1. The basic text format does not enable the underline button in ckeditor by default.
2. The language dropdown is not enabled either.

I've seen Lighthouse call this out as a performance issue on large pages due to the number of DOM nodes on the page. It's very confusing to look at a page as a developer debugging this, because visually the page might be relatively simple and it's easy to assume the problem is in the front-end theme or some JavaScript bug, and not that the content itself is saved this way.

Steps to reproduce

1. Install from the standard profile.
2. Create a paragraph of text in Google Docs or Word.
3. Paste it into an article or page body field.

Proposed resolution

If we remove span from the list of allowed tags, it is completely removed with pasting generating much simpler HTML. This wouldn't prevent adding the tag, or solve pasting from word processors when spans are allowed.

1. Remove span from core/profiles/standard/config/install/filter.format.basic_html.yml.
2. Ensure that span is added in ckeditor 5 configurations when adding the language selector.
3. If span is still required for underlines, do the same thing there.

Remaining tasks

User interface changes

API changes

None.

Data model changes

None.

Release notes snippet

The 'Basic HTML' format provided with the Standard profile no longer allows the <span> tag. In particular, CKEditor 4 is prone to generating many unnecessary <span> tags. This should simplify copying and pasting from MS Word, Google Docs and similar programs into either CKEditor 4 or CKEditor 5 on new installs. Existing installs may want to consider removing the tag from text formats and reviewing existing content.