[chaliy]

Requested by Михайло · project thread

What changed

HTML whose blocks are separated by source indentation (almost every real web page) now renders with blank lines between blocks and no leading spaces. Before, everything after the first block ran together with a leading space, which also stopped later headings and tables from being markdown at all.

The parser now normalizes block-context node lists (document body, blockquote, list item): whitespace-only inline runs between blocks are dropped, and loose inline text next to blocks becomes its own paragraph. Pure-inline content is untouched, so spaces between inline siblings survive.

Why

Found during the maintenance smoke test of examples/basic.rs. The golden fixtures had recorded the broken output, so tests passed.

Before / After

$ printf '<body>\n  <h1>A</h1>\n  <p>B</p>\n</body>' | markkit extract -
# before
# A
 B
# after
# A

B

examples/basic.rs now prints separated heading, paragraph, list and link paragraph. All 9 golden fixtures regenerated; each diff only adds blank lines and removes the leading space (reviewed one by one).

Tests: 7 new integration tests (indented blocks, nested transparent containers, blockquote paragraphs, adjacent lists, loose text, inline-only unchanged) and a property test asserting indented and compact documents render identically. The property test fails on the old parser.

Known limitation, documented rather than fixed here: after an unclosed <strong>, html5ever wraps the following paragraphs in it, so they render inside ** (see malformed fixture).

Risk

• Low / Medium
• Output changes for any multi-block HTML (intended). Consumers comparing exact output will see the added blank lines.

Checklist

• Tests added or updated
• Backward compatibility considered (output change is the fix; CHANGELOG Fixed entry)

---

Generated by Claude Code