~bigbes/confluence-md-utilities

ref: 3c0272ac2232ee3a0edd0c259b71e9fd3384f211 confluence-md-utilities/CLAUDE.md -rw-r--r-- 9.8 KiB
3c0272ac — Eugene Blikh a month ago
fix: round-trip fidelity for heading comments, literal tags, and emphasis

Surfaced by `verify` on two real documents that aren't in the justfile set.

- Heading-only HTML comment (a TOC macro marker, `<h1><!-- ac:toc ... --></h1>`)
  was dropped by xml2md — a CommentNode is neither text nor element, so it fell
  through walkChildren and vanished, collapsing `# <!-- ... -->` to `#`.
- Literal angle-bracket text stored as `&lt;table&gt;` (e.g. `box.backup.start(<table>)`)
  round-tripped into a phantom `<table>` element that swallowed the rest of the
  document: xml2md now backslash-escapes a tag-like `<` in both the TextNode and
  walkChildrenInline paths, and renderText unescapes only `\<` (never a blanket
  UnescapePunctuations, which would eat the deliberate `\|` table-pipe escaping).
- `<strong> x </strong>` (whitespace inside the markers) rendered as literal
  `** x **`; writeEmphasis now moves the whitespace outside the markers.
- normalizeForVerify canonicalizes values Markdown can't carry — task ids,
  <ul>/<ol> bullet styling, and a trailing nbsp before </ac:task-body>.

Adds regression tests for each and documents the new gotchas in CLAUDE.md.
f6c846a1 — Eugene Blikh a month ago
fix: preserve inline spacing in fmt and code-block round-trip

Three related round-trip/formatting bugs, each pinned by a regression test
that fails on the prior code:

- fmt dropped the significant space between two inline elements
  (</strong> <code>): a lone-space text node was always treated as
  indentation. Keep it mid-line; drop it only at a block boundary.
- fmt's line-wrapper dropped a word's trailing space before a tag
  (strings.Fields discards it) and broke inside <code>...</code>,
  injecting indent that xml2md collapses into a spurious in-span space.
  Rewrote wrapLine around an atom model with a per-atom spaceBefore flag
  and keeps an inline code span whole.
- verify.normalizeForVerify's reBareTextAfterHeading used \S for its
  "visible char" token; \S matches '<', so the capture swallowed a
  following <ac:structured-macro> and wrapped it in <p>...</p>, splitting
  a code macro from its body and breaking round-trip for a no-language
  fenced code block after a heading. Use [^\s<] instead.

Full RFC-0010 document now verifies as lossless.
442bf598 — Eugene Blikh 2 months ago
fix: round-trip fidelity for tables, page links, and nested lists

Four real-document failures surfaced via `mdcx verify`:

- Pipes inside `<code>` in GFM table cells were treated as column
  separators, splitting cells. `writeTableRow` now escapes `|` as `\|`
  per the GFM spec; the parser strips the backslash before inline
  parsing so the pipe survives inside the resulting code span.
- Two `<code>` elements separated by whitespace text were being merged
  because `isPrev/NextSiblingCode` skipped whitespace nodes. Adjacency
  is now strict.
- `<ac:link><ri:page …/></ac:link>` was dropped because goldmark cannot
  parse `<ac:link>` as raw inline HTML (the colon disqualifies it).
  Preserved via a `<span data-page-link …>` placeholder; the inverse is
  driven by a new `pageLinkDepth` counter in the renderer.
- Nested `<ol>` collapsed to a flat list and went loose because the
  fixed `2*(depth-1)` indent didn't satisfy CommonMark's column rule
  for ordered markers and a stray `\n` after each `<li>` made the list
  loose. Indent is now a per-level stack (3 for ol, 2 for ul/task) and
  list-item terminators use `ensureSingleNewline`.

Regression tests in converter/md2xml_test.go pin each fix; CLAUDE.md
documents the gotchas.
b3d83904 — Eugene Blikh 2 months ago
feat: add pull --with-comments sidecar

`mdcx pull --with-comments` fetches inline + page comments via
`/rest/api/content/{id}/child/comment` (paginated, two-level reply
expansion) and writes them to `<output>.comments.json` alongside the
Markdown file. Inline comments correlate to the markdown's
`data-inline-comment="UUID"` spans through their `marker_ref` field.

Comment bodies are stored as raw Confluence storage XML; pushing
comments back is not implemented. Requires `--output`.
46926399 — Eugene Blikh 4 months ago
Move CLI to cmd/mdcx/, add justfile and CLAUDE.md

Restructure to standard Go project layout: merge package cmd
into package main under cmd/mdcx/. Fix .gitignore to use /mdcx
so it only ignores the binary at repo root. Update install path
in README.md.

false