feat markdown rich

[80avin]

fixes #12648

• refactor(styled-text): replace StyledTextParagraph with ParagraphBlock enum
• fixup: ungate rich_text_content, remove dead append_paragraph
• fixup: handle unclosed heading at end of input
• feat(core): update StyledText wrapper for ParagraphBlock enum
• feat(renderer): add heading font-size scaling and horizontal rule rendering
• test: update tests for ParagraphBlock, headings, and horizontal rules
• test: add screenshot tests for headings and horizontal rules
• fixup: multiply rule spacing by scale factor for HiDPI support
• feat(markdown): add subscript and superscript support
• feat(markdown): support HTML inline tags <s>, <del>, <i>, <em>, <b>, <strong>
• feat(markdown): add footnotes, images, and math support
• feat(markdown): support HTML comments
• fix: heading spacing, list alignment, H1/H2 underline, table grid rendering
• fix: block quote indent, table alignments, table spacing
• feat(markdown): support <sub> and <sup> HTML tags
• chore: cleanup dead comments and fix InlineHtml in table cells
• chore: cargo fmt

How it looks at present:

[CLAassistant]

All committers have signed the CLA.

[tronical]

Thanks! The breakdown of the series looks overall quite reasonable. I'd very much appreciate though if you could also break it down in PRs. So just take the first refactoring and make a PR. Then we review it and get it in, then we continue from there. We're reviewing this code as humans :)

[80avin]

Thanks @tronical. I will cleanup and split into another PR before moving out of draft.
The intention behind the early draft is to collect opinions on the larger direction as well as blockers.
Right now, I see following as major blockers or points of discussion:

• Is the rendering style acceptable ? Note the horizontal line below H1/H2, style of block-quotes, the slight shift in first line of ordered list (the text after 1. starts early than other lines), etc.
• Blocker: superscript and subscript cannot be implemented using parley as it doesn't support baseline shifting. We may have to move to cosmic-text, which would rewrite/invalidate this PR.
• Major decision: How to dynamically load images ? Right now, the images are replaced with Alt text but a better support of images will demand wider architectural changes, possibly adding APIs allowing users to handle/load image urls.

Also, I'll be interested in knowing if the larger opinion is to skip these concerns and push a lean PR which only implements few text formatting elements.

[tronical]

Thanks @tronical. I will cleanup and split into another PR before moving out of draft. The intention behind the early draft is to collect opinions on the larger direction as well as blockers. Right now, I see following as major blockers or points of discussion:

• Is the rendering style acceptable ? Note the horizontal line below H1/H2, style of block-quotes, the slight shift in first line of ordered list (the text after 1. starts early than other lines), etc.

I suggest that we can fine-tune the style when we get to the PRs that implement style. They should come with screenshot tests, so that we can "lock" in the rendering.0

• Blocker: superscript and subscript cannot be implemented using parley as it doesn't support baseline shifting. We may have to move to cosmic-text, which would rewrite/invalidate this PR.

I don't see us moving to cosmic-text. I suggest to leave out superscript and subscript for now. Perhaps that could be implemented in parley first?

• Major decision: How to dynamically load images ? Right now, the images are replaced with Alt text but a better support of images will demand wider architectural changes, possibly adding APIs allowing users to handle/load image urls.

I'd say let's cross that bridge when we get to that PR. Each increment of better rich text support provides value as-is, and if one particular feature takes longer than another then that's fine.

Also, I'll be interested in knowing if the larger opinion is to skip these concerns and push a lean PR which only implements few text formatting elements.

I suggest to start with the basis, adjusting the data structures, then perhaps headings as a start, and then take it from there.

I suggest to use a tracking issue (that you can also edit as you like) so that we can all see what has landed and what's still missing.