Indented (4-space) code blocks are parsed as paragraphs, and re-emitted unindented

[rundel]

CommonMark indented code blocks (lines starting with four spaces) are not recognized by pampa's parser. The indented content is parsed as a series of Para blocks (with the indentation stripped), so lines like - A and - B become Para [Str "-", Space, Str "A"] and Para [Str "-", Space, Str "B"]. The writer then emits each paragraph at column 0, producing - A\n\n- B, which the reader on the round trip recognises as a real BulletList.

$ printf 'Before.\n\n    categories:\n      - A\n      - B\n\nAfter.\n'
Before.

    categories:
      - A
      - B

After.

$ printf 'Before.\n\n    categories:\n      - A\n      - B\n\nAfter.\n' | cargo run --bin pampa --
[ Para [Str "Before."], Para [Str "categories:"], Para [Str "-", Space, Str "A"], Para [Str "-", Space, Str "B"], Para [Str "After."] ]

$ printf 'Before.\n\n    categories:\n      - A\n      - B\n\nAfter.\n' | cargo run --bin pampa -- -t qmd
Before.

categories:

- A

- B

After.

$ printf 'Before.\n\n    categories:\n      - A\n      - B\n\nAfter.\n' | cargo run --bin pampa -- -t qmd | cargo run --bin pampa --
[ Para [Str "Before."], Para [Str "categories:"], BulletList [[Para [Str "A"]], [Para [Str "B"]]], Para [Str "After."] ]

From quarto-web:

• https://github.com/quarto-dev/quarto-web/blob/8b1be24df1c9301f2700c77e41e7a9f789535717/docs/output-formats/hugo.qmd#L51

[cscheid]

This should be a parse error for us - we're not going to support 4-space code blocks.