Merge pull request 'Emitter 2c: the inline node directives, the directive marks and the carried whitespace' (#13) from inline-nodes into main

AGENTS.md

@@ -111,6 +111,11 @@ live Atlassian APIs; property-generated ADF trees; the CommonMark spec suite aga
   `Result<T>` — `{ ok: true; value } | { ok: false; error: ConvertError }` — nothing throws.
   `try/catch` only wrapped tightly around a call that genuinely throws, converted to a result on
   the spot.
+- Only the hard break's inline segment holds a raw newline — every other spelling escapes one or
+  refuses it — which is how the whitespace carry finds a line edge.
+- A readable spelling tried ahead of a general one — the image, the pipe table, a pipe cell —
+  returns `string | undefined`, never a `Result`: any failure is the fallback signal, and the
+  general form owns the refusal. Refusing there refuses a document the general form spells.
 - Nothing recurses unbounded: the guards walk iteratively, and blocks, marks and attribute values
   are all held to 500 levels, so a deep document is a `Result` rather than the stack overflow that
   waits near 2000.

README.md

@@ -3,8 +3,8 @@
 Lossless conversion between **Atlassian Document Format** (ADF), an extended markdown flavour, and
 an HTML dialect.
 
-**Status: pre-release — `adfToMarkdown` emits the CommonMark subset and the block nodes, nothing
+**Status: pre-release — `adfToMarkdown` emits every node the flavour spells but the opaque carry,
-else is built.**
+nothing else is built.**
 Plan: `todo.md`. Decisions: `AGENTS.md`. The flavour's grammar:
 [`spec/flavour.md`](spec/flavour.md).
 

package.json

@@ -13,7 +13,7 @@
     "node": ">=24"
   },
   "scripts": {
-    "test": "node --test --experimental-test-coverage --test-coverage-exclude=\"src/**/*.test.ts\" --test-coverage-branches=94 --test-coverage-functions=100 --test-coverage-lines=100 \"src/**/*.test.ts\"",
+    "test": "node --test --experimental-test-coverage --test-coverage-exclude=\"src/**/*.test.ts\" --test-coverage-branches=95 --test-coverage-functions=100 --test-coverage-lines=100 \"src/**/*.test.ts\"",
     "typecheck": "tsc --noEmit"
   },
   "devDependencies": {

spec/flavour.md

@@ -63,7 +63,9 @@ literal bracket. Whitespace at either edge of `[content]`, space or tab, is part
 survives inline parsing. Each section below says whether content is required. `:` opens a
 directive only when the name is followed immediately by `[` or `{`, and `{attrs}` must follow
 `]` (or the name) with no gap — anything else (`10:30`, `:smile:`, a stray `{…}` in text) is
-literal text.
+literal text. An inline directive binds as a unit before bracket matching, the way a code span
+does: a `]` or `(` inside its `{attrs}` is the directive's, never the enclosing content's, and a
+`(` after its closing `]` opens no link.
 
 **Container block**:
 
@@ -99,8 +101,9 @@ form, and omits empty `{attrs}` except where the `{` itself claims the directive
 (`:hardBreak{}`).
 
 **Escaping**: the emitter backslash-escapes whatever literal text would otherwise parse as
-directive syntax — the leading `:` of a would-be directive, `]` inside content; outside code
+directive syntax — the leading `:` of a would-be directive, `]` inside content, a `{` right
-spans and code blocks, a backslash before `:` in input yields a literal colon.
+after a directive's closing `]`, which would otherwise be read as the attributes it has none
+of; outside code spans and code blocks, a backslash before `:` in input yields a literal colon.
 
 **Malformed directives are error results**, named: an unclosed container at end of input, a body
 fence line of the container's length or longer, a bare colon-run line outside any container or
@@ -214,10 +217,10 @@ One header row plus plain inline cells is a pipe table; anything richer is the d
 (AGENTS.md §4). Precisely: a table emits as a pipe table exactly when the `table`, every row
 and every cell carry no attrs and no marks, the first row is all `tableHeader` and the rest all
 `tableCell`, every row has the header's cell count, and every cell holds exactly one attr-less,
-mark-less paragraph — an empty cell holds one empty paragraph — with no `|` in a code span,
+mark-less paragraph — an empty cell holds one empty paragraph — with no `|` anywhere the
-link destination or link title: backslash escapes are inert in everything the inline layer spells
+inline layer spells as syntax — a code span, a link destination or title, an inline directive's
-as syntax, so pipe form cannot spell that pipe and the table takes the directive form. A pipe table
+attributes: backslash escapes are inert there, so pipe form cannot spell that pipe and the table
-parses back to exactly that shape.
+takes the directive form. A pipe table parses back to exactly that shape.
 
 ```
 | Part | Qty |

src/adf-document.ts

@@ -4,6 +4,8 @@ export type AdfAttributes = { [key: string]: JsonValue }
 
 export type AttributeKind = 'boolean' | 'json' | 'number' | 'string'
 
+export type AttributeVocabulary = Readonly<Record<string, AttributeKind>>
+
 export type AdfMark = {
   attrs?: AdfAttributes
   type: string

src/adf-to-markdown.test.ts

@@ -29,7 +29,7 @@ test('names the node a refusal came from', () => {
   const unspellable: AdfNode = { attrs: { localId: 'a' }, type: 'paragraph' }
   const list: AdfNode = { content: [{ content: [paragraph({ text: 'x', type: 'text' })], type: 'listItem' }, { content: [unspellable], type: 'listItem' }], type: 'bulletList' }
   assert.deepEqual(path(adfToMarkdown(document(paragraph({ text: 'x', type: 'text' }), list))), ['content', 1, 'content', 1, 'content', 0])
-  assert.deepEqual(path(adfToMarkdown(document(paragraph({ text: 'x', type: 'text' }, { type: 'mention' })))), ['content', 0, 'content', 1])
+  assert.deepEqual(path(adfToMarkdown(document(paragraph({ text: 'x', type: 'text' }, { type: 'blockCard' })))), ['content', 0, 'content', 1])
   assert.deepEqual(path(adfToMarkdown({ type: 'doc', version: 2 })), [])
 })
 
@@ -93,17 +93,15 @@ test('refuses a link attribute no markdown spelling holds', () => {
 })
 
 test('refuses a mark the canonical spellings cannot nest', () => {
-  assert.equal(code(adfToMarkdown(document(paragraph({ marks: [{ type: 'underline' }], text: 'x', type: 'text' })))), 'unspellable-mark')
+  assert.equal(code(adfToMarkdown(document(paragraph({ marks: [{ type: 'annotation' }], text: 'x', type: 'text' })))), 'unspellable-mark')
   assert.equal(code(adfToMarkdown(document(paragraph({ marks: [{ type: 'code' }, { type: 'strong' }], text: 'x', type: 'text' })))), 'unspellable-mark')
   assert.equal(code(adfToMarkdown(document(paragraph({ marks: [{ attrs: { colour: 'red' }, type: 'em' }], text: 'x', type: 'text' })))), 'unspellable-mark')
 })
 
 test('refuses whitespace CommonMark cannot hold', () => {
-  assert.equal(code(adfToMarkdown(document(paragraph({ text: '  lead', type: 'text' })))), 'unspellable-whitespace')
-  assert.equal(code(adfToMarkdown(document(paragraph({ text: 'trail ', type: 'text' })))), 'unspellable-whitespace')
-  assert.equal(code(adfToMarkdown(document(paragraph({ text: 'a\nb', type: 'text' })))), 'unspellable-whitespace')
   assert.equal(code(adfToMarkdown(document(paragraph({ text: '\fa', type: 'text' })))), 'unspellable-whitespace')
-  assert.equal(code(adfToMarkdown(document(paragraph({ marks: [{ type: 'em' }], text: 'x ', type: 'text' }, { text: 'y', type: 'text' })))), 'unspellable-whitespace')
+  assert.equal(code(adfToMarkdown(document(paragraph({ text: 'a\rb', type: 'text' })))), 'unspellable-whitespace')
+  assert.equal(code(adfToMarkdown(document(paragraph({ marks: [{ type: 'em' }], text: 'x ', type: 'text' }, { text: 'y', type: 'text' })))), 'unspellable-mark')
 })
 
 test('refuses a line whose start block parsing would claim', () => {
@@ -118,7 +116,7 @@ test('refuses two adjacent lists of the same kind', () => {
 test('refuses a node type the canonical form does not cover', () => {
   assert.equal(code(adfToMarkdown(document({ type: 'blockCard' }))), 'unsupported-node-type')
   assert.equal(code(adfToMarkdown(document({ type: 'toString' }))), 'unsupported-node-type')
-  assert.equal(code(adfToMarkdown(document(paragraph({ type: 'mention' })))), 'unsupported-node-type')
+  assert.equal(code(adfToMarkdown(document(paragraph({ type: 'blockCard' })))), 'unsupported-node-type')
 })
 
 test('refuses a node whose content model the canonical form cannot emit', () => {
@@ -230,6 +228,9 @@ test('escapes a literal delimiter that would merge with an emitted one', () => {
   assert.equal(emitted({ text: '`', type: 'text' }, marked('x', { type: 'code' })), '\\``x`\n')
   assert.equal(emitted(marked('x', { type: 'code' }), { text: '`', type: 'text' }), '`x`\\`\n')
   assert.equal(emitted({ text: '!', type: 'text' }, marked('x', { attrs: { href: 'https://example.com/' }, type: 'link' })), '\\![x](https://example.com/)\n')
+  assert.equal(emitted(marked('x', { type: 'underline' }), { text: '{}', type: 'text' }), ':underline[x]\\{}\n')
+  assert.equal(emitted({ attrs: { text: '' }, type: 'status' }, { text: '{color=red}', type: 'text' }), ':status[]\\{color=red}\n')
+  assert.equal(emitted(marked('x', { attrs: { href: 'https://example.com/' }, type: 'link' }), { text: '{}', type: 'text' }), '[x](https://example.com/){}\n')
 })
 
 test('escapes a hyphen underline a hard break would expose', () => {
@@ -372,10 +373,85 @@ test('spells a table as a pipe table only where every row and cell is plain', ()
   assert.ok(directive(adfToMarkdown(table(row(cell('tableHeader'))))))
   assert.ok(directive(adfToMarkdown(table(row(cell('tableHeader', text('a'), text('b')))))))
   assert.equal(code(adfToMarkdown(table(row(cell('tableHeader', { attrs: { localId: 'a' }, type: 'paragraph' }))))), 'unspelled-node-attribute')
+  assert.equal(code(adfToMarkdown(table(row(cell('tableHeader', { content: [{ type: 'blockCard' }], type: 'paragraph' }))))), 'unsupported-node-type')
+  assert.equal(code(adfToMarkdown(table(row(cell('tableHeader', { content: [{ text: '\fa', type: 'text' }], type: 'paragraph' }))))), 'unspellable-whitespace')
   assert.ok(directive(adfToMarkdown(table(row(cell('tableHeader', { attrs: { level: 1 }, type: 'heading' }))))))
-  assert.equal(code(adfToMarkdown(table(row(cell('tableHeader', { content: [{ text: ' a', type: 'text' }], type: 'paragraph' }))))), 'unspellable-whitespace')
+  assert.equal(markdown(adfToMarkdown(table(row(cell('tableHeader', { content: [{ text: ' a', type: 'text' }], type: 'paragraph' }))))), '| :text{text=" "}a |\n| --- |\n')
   const marked = (mark: AdfMark): AdfDocument => table(row(cell('tableHeader', { content: [{ marks: [mark], text: 'l', type: 'text' }], type: 'paragraph' })))
   assert.ok(directive(adfToMarkdown(marked({ attrs: { href: 'https://example.com/?x|y' }, type: 'link' }))))
   assert.ok(directive(adfToMarkdown(marked({ attrs: { href: 'https://example.com/', title: 'a|b' }, type: 'link' }))))
+  const fallsBack = (node: AdfNode): boolean => directive(adfToMarkdown(table(row(cell('tableHeader', { content: [node], type: 'paragraph' })))))
+  assert.ok(fallsBack({ marks: [{ type: 'code' }], text: 'a|b', type: 'text' }))
+  assert.ok(fallsBack({ attrs: { style: 'a|b' }, type: 'status' }))
   assert.equal(markdown(adfToMarkdown(marked({ attrs: { href: 'https://example.com/x' }, type: 'link' }))), '| [l](https://example.com/x) |\n| --- |\n')
 })
+
+test('spells an inline node as a directive with its content slot and attributes', () => {
+  const emitted = (node: AdfNode): string => markdown(adfToMarkdown(document(paragraph(node))))
+  assert.equal(emitted({ attrs: { timestamp: '1756080000000' }, type: 'date' }), ':date{timestamp=1756080000000}\n')
+  assert.equal(emitted({ type: 'mention' }), ':mention{}\n')
+  assert.equal(emitted({ attrs: { text: '' }, type: 'status' }), ':status[]\n')
+  assert.equal(emitted({ attrs: { color: 'yellow', text: 'In review' }, type: 'status' }), ':status[In review]{color=yellow}\n')
+  assert.equal(emitted({ attrs: { id: '1f389', text: 'a]b' }, type: 'emoji' }), ':emoji[a\\]b]{id=1f389}\n')
+  assert.equal(emitted({ attrs: { data: { url: 'https://example.com/' } }, type: 'inlineCard' }), ':inlineCard{data="{\\"url\\":\\"https://example.com/\\"}"}\n')
+  assert.equal(emitted({ attrs: { height: 24 }, type: 'mediaInline' }), ':mediaInline{height=24}\n')
+})
+
+test('refuses an inline node attribute no section spells', () => {
+  const refused = (node: AdfNode): string => code(adfToMarkdown(document(paragraph(node))))
+  assert.equal(refused({ attrs: { rounded: true }, type: 'status' }), 'unspelled-node-attribute')
+  assert.equal(refused({ attrs: { toString: 'x' }, type: 'status' }), 'unspelled-node-attribute')
+  assert.equal(refused({ attrs: { color: 4 }, type: 'status' }), 'unsupported-node-shape')
+  assert.equal(refused({ attrs: { width: '2' }, type: 'mediaInline' }), 'unsupported-node-shape')
+})
+
+test('refuses the content and slot an inline directive has no room for', () => {
+  const refused = (node: AdfNode): string => code(adfToMarkdown(document(paragraph(node, { text: 'y', type: 'text' }))))
+  assert.equal(refused({ content: [{ text: 'x', type: 'text' }], type: 'status' }), 'unsupported-node-shape')
+  assert.equal(refused({ text: 'x', type: 'status' }), 'unsupported-node-shape')
+  assert.equal(refused({ content: [{ text: 'x', type: 'text' }], type: 'hardBreak' }), 'unsupported-node-shape')
+  assert.equal(refused({ text: 'x', type: 'hardBreak' }), 'unsupported-node-shape')
+  assert.equal(refused({ attrs: { text: 4 }, type: 'status' }), 'unsupported-node-shape')
+  assert.equal(refused({ attrs: { text: 'a\nb' }, type: 'status' }), 'unspellable-whitespace')
+  assert.equal(refused({ attrs: { text: 'a\u0000b' }, type: 'status' }), 'unspellable-character')
+})
+
+test('spells the directive marks around the longest run they cover', () => {
+  const marked = (text: string, ...marks: AdfMark[]): AdfNode => ({ marks, text, type: 'text' })
+  const emitted = (...content: AdfNode[]): string => markdown(adfToMarkdown(document(paragraph(...content))))
+  const underline: AdfMark = { type: 'underline' }
+  assert.equal(emitted(marked('x', underline)), ':underline[x]\n')
+  assert.equal(emitted(marked('a', underline), marked('b', underline)), ':underline[ab]\n')
+  assert.equal(emitted(marked('x', { attrs: { type: 'sub' }, type: 'subsup' })), ':subsup[x]{type=sub}\n')
+  assert.equal(emitted(marked('x', { attrs: { color: '#ae2e24' }, type: 'textColor' })), ':textColor[x]{color="#ae2e24"}\n')
+  assert.equal(emitted(marked('x', { attrs: { color: '#091e42', size: 2 }, type: 'border' })), ':border[x]{color="#091e42" size=2}\n')
+  assert.equal(emitted(marked('x', { type: 'em' }, underline)), '_:underline[x]_\n')
+  assert.equal(emitted(marked('x', underline, { type: 'em' })), ':underline[_x_]\n')
+  assert.equal(emitted(marked('a', underline), { marks: [underline], type: 'hardBreak' }, marked('b', underline)), ':underline[a:hardBreak{}b]\n')
+  const link: AdfMark = { attrs: { href: 'https://example.com/' }, type: 'link' }
+  assert.equal(
+    emitted(marked('a', underline, link), { marks: [underline, link], type: 'hardBreak' }, marked('b', underline, link)),
+    ':underline[[a:hardBreak{}b](https://example.com/)]\n',
+  )
+  assert.equal(
+    emitted(marked('a', link), { marks: [link], type: 'hardBreak' }, marked('b', link)),
+    '[a\\\nb](https://example.com/)\n',
+  )
+})
+
+test('refuses a mark directive attribute no spelling holds', () => {
+  const refused = (mark: AdfMark): string => code(adfToMarkdown(document(paragraph({ marks: [mark], text: 'x', type: 'text' }))))
+  assert.equal(refused({ attrs: { width: 2 }, type: 'border' }), 'unspellable-mark')
+  assert.equal(refused({ attrs: { size: '2' }, type: 'border' }), 'unspellable-mark')
+})
+
+test('carries whitespace CommonMark strips in the reserved text directive', () => {
+  const emitted = (...content: AdfNode[]): string => markdown(adfToMarkdown(document(paragraph(...content))))
+  assert.equal(emitted({ text: '  lead', type: 'text' }), ':text{text="  "}lead\n')
+  assert.equal(emitted({ text: 'trail ', type: 'text' }), 'trail:text{text=" "}\n')
+  assert.equal(emitted({ text: 'a\nb', type: 'text' }), 'a:text{text="\\n"}b\n')
+  assert.equal(emitted({ text: '\t', type: 'text' }), ':text{text="\\t"}\n')
+  assert.equal(emitted({ text: 'a ', type: 'text' }, { type: 'hardBreak' }, { text: ' b', type: 'text' }), 'a:text{text=" "}\\\n:text{text=" "}b\n')
+  assert.equal(emitted({ marks: [{ type: 'em' }], text: ' a ', type: 'text' }), '_:text{text=" "}a:text{text=" "}_\n')
+  assert.equal(markdown(adfToMarkdown(document({ attrs: { level: 1 }, content: [{ text: 'x ', type: 'text' }], type: 'heading' }))), '# x:text{text=" "}\n')
+})

src/adf-to-markdown.ts

@@ -2,9 +2,9 @@ import type { AdfDocument, AdfNode } from './adf-document.ts'
 import type { BlockDirective } from './block-directives.ts'
 import type { JsonValue } from './json-value.ts'
 import { blockDirective, spellDirectiveHeader } from './block-directives.ts'
-import { emitImage } from './markdown-image.ts'
+import { tryImage } from './markdown-image.ts'
 import { emitInlineLine } from './markdown-inline.ts'
-import { emitPipeTable } from './markdown-pipe-table.ts'
+import { tryPipeTable } from './markdown-pipe-table.ts'
 import { failure, success, type ConvertErrorPath, type Result } from './result.ts'
 import { holdsNullCharacter, isThematicBreak } from './commonmark-grammar.ts'
 import { isAdfDocument } from './adf-document.ts'
@@ -98,7 +98,11 @@ function emitBlock(node: AdfNode, path: ConvertErrorPath, depth: number): Result
 
 function commonMarkLine(text: Result<string>): Result<EmittedBlock> {
   if (!text.ok) return text
-  return success({ fenceColons: 0, spelling: 'commonmark', text: text.value })
+  return success(commonMarkText(text.value))
+}
+
+function commonMarkText(text: string): EmittedBlock {
+  return { fenceColons: 0, spelling: 'commonmark', text }
 }
 
 function commonMarkContainer(body: Result<EmittedBody>): Result<EmittedBlock> {
@@ -131,15 +135,15 @@ function emitInlineBody(content: readonly AdfNode[], path: ConvertErrorPath): Re
 }
 
 function emitMediaSingle(node: AdfNode, directive: BlockDirective, path: ConvertErrorPath, depth: number): Result<EmittedBlock> {
-  const image = emitImage(node, path)
+  const image = tryImage(node, path)
   if (image === undefined) return emitDirectiveBlock(node, directive, path, depth)
-  return commonMarkLine(image)
+  return success(commonMarkText(image))
 }
 
 function emitTable(node: AdfNode, directive: BlockDirective, path: ConvertErrorPath, depth: number): Result<EmittedBlock> {
-  const pipe = emitPipeTable(node, path)
+  const pipe = tryPipeTable(node, path)
   if (pipe === undefined) return emitDirectiveBlock(node, directive, path, depth)
-  return commonMarkLine(pipe)
+  return success(commonMarkText(pipe))
 }
 
 function emitBlockquote(node: AdfNode, path: ConvertErrorPath, depth: number): Result<EmittedBody> {

src/block-directives.ts

@@ -1,15 +1,15 @@
-import type { AdfMark, AdfNode, AttributeKind } from './adf-document.ts'
+import type { AdfMark, AdfNode, AttributeVocabulary } from './adf-document.ts'
 import type { JsonValue } from './json-value.ts'
 import { failure, success, type ConvertErrorPath, type Result } from './result.ts'
-import { isBareToken, spellAttributeValue, spellAttributes, spellJsonAttribute } from './directive-attributes.ts'
+import { attributeFailure, isBareToken, spellAttributes, spellJsonAttribute, vocabularyPairs } from './directive-attributes.ts'
 
 export type BlockDirective = {
   argument?: string
-  attributes: Readonly<Record<string, AttributeKind>>
+  attributes: AttributeVocabulary
   body: 'block' | 'inline' | 'none'
 }
 
-const cellAttributes: Readonly<Record<string, AttributeKind>> = {
+const cellAttributes: AttributeVocabulary = {
   background: 'string',
   colspan: 'number',
   colwidth: 'json',
@@ -18,9 +18,9 @@ const cellAttributes: Readonly<Record<string, AttributeKind>> = {
   valign: 'string',
 }
 
-const expandAttributes: Readonly<Record<string, AttributeKind>> = { localId: 'string', title: 'string' }
+const expandAttributes: AttributeVocabulary = { localId: 'string', title: 'string' }
 
-const extensionAttributes: Readonly<Record<string, AttributeKind>> = {
+const extensionAttributes: AttributeVocabulary = {
   extensionKey: 'string',
   extensionType: 'string',
   layout: 'string',
@@ -29,9 +29,9 @@ const extensionAttributes: Readonly<Record<string, AttributeKind>> = {
   text: 'string',
 }
 
-const itemAttributes: Readonly<Record<string, AttributeKind>> = { localId: 'string' }
+const itemAttributes: AttributeVocabulary = { localId: 'string' }
 
-const mediaAttributes: Readonly<Record<string, AttributeKind>> = {
+const mediaAttributes: AttributeVocabulary = {
   alt: 'string',
   collection: 'string',
   height: 'number',
@@ -43,7 +43,7 @@ const mediaAttributes: Readonly<Record<string, AttributeKind>> = {
   width: 'number',
 }
 
-const syncBlockAttributes: Readonly<Record<string, AttributeKind>> = { localId: 'string', resourceId: 'string' }
+const syncBlockAttributes: AttributeVocabulary = { localId: 'string', resourceId: 'string' }
 
 const blockDirectives: Readonly<Record<string, BlockDirective>> = {
   blockTaskItem: { argument: 'state', attributes: itemAttributes, body: 'block' },
@@ -81,30 +81,23 @@ export function blockDirective(type: string): BlockDirective | undefined {
 }
 
 export function spellDirectiveHeader(node: AdfNode, directive: BlockDirective, path: ConvertErrorPath): Result<string> {
-  const pairs: [string, string][] = []
+  const argument = spellArgument(node, directive, path)
-  let argument = ''
+  if (!argument.ok) return argument
-  for (const [key, value] of Object.entries(node.attrs ?? {})) {
+  const spelled = vocabularyPairs(node.attrs ?? {}, directive.attributes, directive.argument)
-    if (key === directive.argument) {
+  if (spelled.fault !== undefined) return attributeFailure(node.type, spelled.fault, path)
-      if (typeof value !== 'string' || !isBareToken(value)) {
-        return failure('unspelled-node-attribute', `the ${node.type} attribute ${key} holds no bare token the arg slot spells`, path)
-      }
-      argument = ` ${value}`
-      continue
-    }
-    const kind = attributeKind(directive, key)
-    if (kind === undefined) return failure('unspelled-node-attribute', `the ${node.type} attribute ${key} has no canonical markdown spelling`, path)
-    const spelled = spellAttributeValue(value, kind)
-    if (spelled === undefined) return failure('unsupported-node-shape', `the ${node.type} attribute ${key} holds no ${kind}`, path)
-    pairs.push([key, spelled])
-  }
   const marks = node.marks ?? []
-  if (marks.length > 0) pairs.push(['marks', spellJsonAttribute(markValues(marks))])
+  if (marks.length > 0) spelled.pairs.push(['marks', spellJsonAttribute(markValues(marks))])
-  const attributes = spellAttributes(pairs)
+  const attributes = spellAttributes(spelled.pairs)
-  return success(`${node.type}${argument}${attributes === '' ? '' : ` ${attributes}`}`)
+  return success(`${node.type}${argument.value}${attributes === '' ? '' : ` ${attributes}`}`)
 }
 
-function attributeKind(directive: BlockDirective, key: string): AttributeKind | undefined {
+function spellArgument(node: AdfNode, directive: BlockDirective, path: ConvertErrorPath): Result<string> {
-  return Object.hasOwn(directive.attributes, key) ? directive.attributes[key] : undefined
+  const value = directive.argument === undefined ? undefined : node.attrs?.[directive.argument]
+  if (value === undefined) return success('')
+  if (typeof value !== 'string' || !isBareToken(value)) {
+    return failure('unspelled-node-attribute', `the ${node.type} attribute ${directive.argument} holds no bare token the arg slot spells`, path)
+  }
+  return success(` ${value}`)
 }
 
 function markValues(marks: readonly AdfMark[]): JsonValue {

src/corpus.test.ts

@@ -13,8 +13,7 @@ const corpusRoot = join(dirname(fileURLToPath(import.meta.url)), '..', 'corpus')
 const roundTripRoot = join(corpusRoot, 'round-trip')
 const unspellableRoot = join(corpusRoot, 'unspellable')
 
-const emittingDirectories = ['block-nodes', 'commonmark-subset']
+const emittingDirectories = ['block-nodes', 'commonmark-subset', 'inline-nodes']
-const pendingDirectories = ['inline-nodes']
 
 function directoryNames(root: string): string[] {
   return readdirSync(root, { withFileTypes: true })
@@ -41,21 +40,16 @@ function corpusJsonPaths(): string[] {
     .sort()
 }
 
-test('every round-trip directory is either emitting or explicitly pending', () => {
+test('every round-trip directory emits', () => {
-  assert.deepEqual(directoryNames(roundTripRoot), [...emittingDirectories, ...pendingDirectories].sort())
+  assert.deepEqual(directoryNames(roundTripRoot), [...emittingDirectories].sort())
 })
 
-for (const directory of [...emittingDirectories, ...pendingDirectories].sort()) {
-  test(`${directory} pairs every .json with a .md`, () => {
-    assert.deepEqual(fixtureNames(directory, '.json'), fixtureNames(directory, '.md'))
-  })
-}
-
 for (const directory of emittingDirectories) {
   const names = [...new Set([...fixtureNames(directory, '.json'), ...fixtureNames(directory, '.md')])].sort()
 
-  test(`${directory} holds fixtures`, () => {
+  test(`${directory} pairs every .json with a .md`, () => {
     assert.ok(names.length > 0, `${directory} is expected to emit but holds no fixture pairs`)
+    assert.deepEqual(fixtureNames(directory, '.json'), fixtureNames(directory, '.md'))
   })
 
   for (const name of names) {

src/directive-attributes.ts

@@ -1,13 +1,36 @@
-import type { AttributeKind } from './adf-document.ts'
+import type { AdfAttributes, AttributeKind, AttributeVocabulary } from './adf-document.ts'
 import type { JsonValue } from './json-value.ts'
+import { failure, type ConvertErrorPath, type Result } from './result.ts'
 import { serializeCanonicalJson } from './canonical-json.ts'
 
+export type AttributeFault = { key: string; kind: AttributeKind | undefined }
+
+export type SpelledPairs = { fault: AttributeFault; pairs?: undefined } | { fault?: undefined; pairs: [string, string][] }
+
 const bareToken = /^[A-Za-z0-9_-]+$/
 
 export function isBareToken(text: string): boolean {
   return bareToken.test(text)
 }
 
+export function attributeFailure<T>(type: string, fault: AttributeFault, path: ConvertErrorPath): Result<T> {
+  if (fault.kind === undefined) return failure('unspelled-node-attribute', `the ${type} attribute ${fault.key} has no canonical markdown spelling`, path)
+  return failure('unsupported-node-shape', `the ${type} attribute ${fault.key} holds no ${fault.kind}`, path)
+}
+
+export function vocabularyPairs(attrs: AdfAttributes, vocabulary: AttributeVocabulary, slot: string | undefined): SpelledPairs {
+  const pairs: [string, string][] = []
+  for (const [key, value] of Object.entries(attrs)) {
+    if (key === slot) continue
+    const kind = Object.hasOwn(vocabulary, key) ? vocabulary[key] : undefined
+    if (kind === undefined) return { fault: { key, kind: undefined } }
+    const spelled = spellAttributeValue(value, kind)
+    if (spelled === undefined) return { fault: { key, kind } }
+    pairs.push([key, spelled])
+  }
+  return { pairs }
+}
+
 export function spellAttributes(pairs: readonly (readonly [string, string])[]): string {
   if (pairs.length === 0) return ''
   const spelled = [...pairs].sort(([left], [right]) => (left < right ? -1 : 1)).map(([key, value]) => `${key}=${value}`)
@@ -17,15 +40,15 @@ export function spellAttributes(pairs: readonly (readonly [string, string])[]):
 export function spellAttributeValue(value: JsonValue, kind: AttributeKind): string | undefined {
   if (kind === 'json') return spellJsonAttribute(value)
   if (kind === 'boolean') return typeof value === 'boolean' ? `${value}` : undefined
-  if (kind === 'number') return typeof value === 'number' ? spell(JSON.stringify(value)) : undefined
+  if (kind === 'number') return typeof value === 'number' ? spellStringAttribute(JSON.stringify(value)) : undefined
-  return typeof value === 'string' ? spell(value) : undefined
+  return typeof value === 'string' ? spellStringAttribute(value) : undefined
 }
 
 export function spellJsonAttribute(value: JsonValue): string {
   return quote(serializeCanonicalJson(value, 'compact'))
 }
 
-function spell(text: string): string {
+export function spellStringAttribute(text: string): string {
   return isBareToken(text) ? text : quote(text)
 }
 

src/inline-directives.ts

@@ -0,0 +1,62 @@
+import type { AdfMark, AdfNode, AttributeVocabulary } from './adf-document.ts'
+import type { AttributeFault } from './directive-attributes.ts'
+import { failure, success, type ConvertErrorPath, type Result } from './result.ts'
+import { attributeFailure, spellAttributes, vocabularyPairs } from './directive-attributes.ts'
+
+export type InlineDirective = {
+  attributes: AttributeVocabulary
+  slot?: string
+}
+
+const inlineDirectives: Readonly<Record<string, InlineDirective>> = {
+  date: { attributes: { localId: 'string', timestamp: 'string' } },
+  emoji: { attributes: { id: 'string', localId: 'string', shortName: 'string' }, slot: 'text' },
+  inlineCard: { attributes: { data: 'json', localId: 'string', url: 'string' } },
+  mediaInline: {
+    attributes: {
+      alt: 'string',
+      collection: 'string',
+      data: 'json',
+      height: 'number',
+      id: 'string',
+      localId: 'string',
+      occurrenceKey: 'string',
+      type: 'string',
+      width: 'number',
+    },
+  },
+  mention: { attributes: { accessLevel: 'string', id: 'string', localId: 'string', userType: 'string' }, slot: 'text' },
+  status: { attributes: { color: 'string', localId: 'string', style: 'string' }, slot: 'text' },
+}
+
+const markDirectives: Readonly<Record<string, AttributeVocabulary>> = {
+  border: { color: 'string', size: 'number' },
+  subsup: { type: 'string' },
+  textColor: { color: 'string' },
+  underline: {},
+}
+
+export function inlineDirective(type: string): InlineDirective | undefined {
+  return Object.hasOwn(inlineDirectives, type) ? inlineDirectives[type] : undefined
+}
+
+export function markDirective(type: string): AttributeVocabulary | undefined {
+  return Object.hasOwn(markDirectives, type) ? markDirectives[type] : undefined
+}
+
+export function spellInlineNodeAttributes(node: AdfNode, directive: InlineDirective, path: ConvertErrorPath): Result<string> {
+  const spelled = vocabularyPairs(node.attrs ?? {}, directive.attributes, directive.slot)
+  if (spelled.fault !== undefined) return attributeFailure(node.type, spelled.fault, path)
+  return success(spellAttributes(spelled.pairs))
+}
+
+export function spellMarkAttributes(mark: AdfMark, vocabulary: AttributeVocabulary, path: ConvertErrorPath): Result<string> {
+  const spelled = vocabularyPairs(mark.attrs ?? {}, vocabulary, undefined)
+  if (spelled.fault !== undefined) return markFailure(mark.type, spelled.fault, path)
+  return success(spellAttributes(spelled.pairs))
+}
+
+function markFailure<T>(type: string, fault: AttributeFault, path: ConvertErrorPath): Result<T> {
+  if (fault.kind === undefined) return failure('unspellable-mark', `the ${type} spelling holds no ${fault.key} attribute`, path)
+  return failure('unspellable-mark', `the ${type} attribute ${fault.key} holds no ${fault.kind}`, path)
+}

src/markdown-escaping.ts

@@ -1,8 +1,12 @@
 import { escapesLineClaim, isUnicodeWhitespace, opensBracketedAutolink, startsEntityReference, type LinePosition } from './commonmark-grammar.ts'
 
+export type EmphasisRole = 'close' | 'open'
+
+export type InlineEscaping = 'attribute' | 'backslash' | 'bracketed' | 'none'
+
 export type InlineSegment =
-  | { kind: 'emphasis-close' | 'emphasis-open'; mark: string; text: string }
+  | { emphasis: EmphasisRole; escaping: 'none'; mark: string; text: string }
-  | { kind: 'link-text' | 'literal' | 'syntax'; text: string }
+  | { emphasis?: undefined; escaping: InlineEscaping; text: string }
 
 export type AssembledLine = { line: string; unspellableMark: string | undefined }
 
@@ -14,7 +18,7 @@ const delimiters = ['*', '_', '`', '~']
 
 const asciiPunctuation = /[!"#$%&'()*+,\-./:;<=>?@[\\\]^_`{|}~]/
 const htmlConstructs = [/^<[!?]/, /^<\/?[A-Za-z][A-Za-z0-9-]*(?:[\s/>]|$)/, /^<[^\s<>@]+@[^\s<>@]+>/]
-const inlineDirective = /^:[a-z][A-Za-z0-9]*[[{]/
+const inlineDirectiveOpener = /^:[a-z][A-Za-z0-9]*[[{]/
 const linkOpener = /\](?=[([:])/
 const unicodePunctuation = /[\p{P}\p{S}]/u
 
@@ -36,8 +40,8 @@ function resolveEmphasis(segments: readonly InlineSegment[]): InlineSegment[] {
   for (let index = 0; index < resolved.length; index += 1) {
     const segment = resolved[index]
     if (segment === undefined) continue
-    if (segment.kind === 'emphasis-open') open.push(index)
+    if (segment.emphasis === 'open') open.push(index)
-    if (segment.kind !== 'emphasis-close') continue
+    if (segment.emphasis !== 'close') continue
     const openerIndex = open.pop()
     const opener = openerIndex === undefined ? undefined : resolved[openerIndex]
     if (openerIndex === undefined || opener === undefined) continue
@@ -53,15 +57,15 @@ function resolveEmphasis(segments: readonly InlineSegment[]): InlineSegment[] {
 
 function escape(segments: readonly InlineSegment[], container: LineContainer): AssembledLine {
   const scan = segments.map((segment) => segment.text).join('')
-  const kinds: InlineSegment['kind'][] = []
+  const escapings: InlineEscaping[] = []
-  for (const segment of segments) for (let index = 0; index < segment.text.length; index += 1) kinds.push(segment.kind)
+  for (const segment of segments) for (let index = 0; index < segment.text.length; index += 1) escapings.push(segment.escaping)
   const escaped = new Set<number>()
   const placements: number[] = []
   let output = ''
   for (let index = 0; index < scan.length; index += 1) {
-    const kind = kinds[index]
+    const escaping = escapings[index]
-    const escapable = kind === 'literal' || kind === 'link-text'
+    const escapable = escaping === 'backslash' || escaping === 'bracketed'
-    if (escapable && (mergesWithSyntax(scan, kinds, index) || opensConstruct(scan, index, kind === 'link-text', container, escaped))) {
+    if (escapable && (mergesWithSyntax(scan, escapings, index) || opensConstruct(scan, index, escaping === 'bracketed', container, escaped))) {
       output += '\\'
       escaped.add(index)
     }
@@ -87,8 +91,8 @@ function delimiterRuns(segments: readonly InlineSegment[], placements: readonly
   for (const segment of segments) {
     const start = placements[cursor] ?? 0
     cursor += segment.text.length
-    if (segment.kind !== 'emphasis-close' && segment.kind !== 'emphasis-open') continue
+    if (segment.emphasis === undefined) continue
-    const closes = segment.kind === 'emphasis-close'
+    const closes = segment.emphasis === 'close'
     const end = start + segment.text.length
     const previous = runs[runs.length - 1]
     if (previous !== undefined && previous.end === start && previous.character === segment.text.charAt(0)) {
@@ -108,28 +112,29 @@ function delimiterRuns(segments: readonly InlineSegment[], placements: readonly
   return runs
 }
 
-function mergesWithSyntax(scan: string, kinds: readonly (InlineSegment['kind'] | undefined)[], index: number): boolean {
+function mergesWithSyntax(scan: string, escapings: readonly (InlineEscaping | undefined)[], index: number): boolean {
   const character = scan.charAt(index)
-  if (character === '!') return scan.charAt(index + 1) === '[' && isSyntax(kinds[index + 1])
+  if (character === '!') return scan.charAt(index + 1) === '[' && isSyntax(escapings[index + 1])
+  if (character === '{') return scan.charAt(index - 1) === ']' && isSyntax(escapings[index - 1])
   if (!delimiters.includes(character)) return false
-  return touchesSyntax(scan, kinds, index, -1) || touchesSyntax(scan, kinds, index, 1)
+  return touchesSyntax(scan, escapings, index, -1) || touchesSyntax(scan, escapings, index, 1)
 }
 
-function touchesSyntax(scan: string, kinds: readonly (InlineSegment['kind'] | undefined)[], index: number, step: number): boolean {
+function touchesSyntax(scan: string, escapings: readonly (InlineEscaping | undefined)[], index: number, step: number): boolean {
   const character = scan.charAt(index)
   let cursor = index + step
-  while (scan.charAt(cursor) === character && !isSyntax(kinds[cursor])) cursor += step
+  while (scan.charAt(cursor) === character && !isSyntax(escapings[cursor])) cursor += step
-  return scan.charAt(cursor) === character && isSyntax(kinds[cursor])
+  return scan.charAt(cursor) === character && isSyntax(escapings[cursor])
 }
 
-function isSyntax(kind: InlineSegment['kind'] | undefined): boolean {
+function isSyntax(escaping: InlineEscaping | undefined): boolean {
-  return kind === 'emphasis-close' || kind === 'emphasis-open' || kind === 'syntax'
+  return escaping === 'attribute' || escaping === 'none'
 }
 
-function opensConstruct(scan: string, index: number, inLinkText: boolean, container: LineContainer, escaped: ReadonlySet<number>): boolean {
+function opensConstruct(scan: string, index: number, inBrackets: boolean, container: LineContainer, escaped: ReadonlySet<number>): boolean {
   if (container === 'heading' && closesHeading(scan, index)) return true
   if (container === 'paragraph' && claimsLineStart(scan, index)) return true
-  return claimsCharacter(scan, index, inLinkText, container, escaped)
+  return claimsCharacter(scan, index, inBrackets, container, escaped)
 }
 
 function claimsLineStart(scan: string, index: number): boolean {
@@ -145,15 +150,15 @@ function closesHeading(scan: string, index: number): boolean {
   return index === 0 || /[ \t]/.test(scan.charAt(index - 1))
 }
 
-function claimsCharacter(scan: string, index: number, inLinkText: boolean, container: LineContainer, escaped: ReadonlySet<number>): boolean {
+function claimsCharacter(scan: string, index: number, inBrackets: boolean, container: LineContainer, escaped: ReadonlySet<number>): boolean {
   const character = scan.charAt(index)
   const rest = scan.slice(index)
-  if (inLinkText && (character === '[' || character === ']')) return true
+  if (inBrackets && (character === '[' || character === ']')) return true
   if (character === '|') return container === 'table-cell'
   if (character === '\\') return asciiPunctuation.test(scan.charAt(index + 1))
   if (character === '&') return startsEntityReference(rest)
   if (character === '<') return opensBracketedAutolink(rest) || htmlConstructs.some((construct) => construct.test(rest))
-  if (character === ':') return inlineDirective.test(rest)
+  if (character === ':') return inlineDirectiveOpener.test(rest)
   if (character === '[') return linkOpener.test(rest)
   if (character === '`') return opensCodeSpan(scan, index, escaped)
   if (character === '*' || character === '_' || character === '~') return opensEmphasis(scan, index, escaped)

src/markdown-image.ts

@@ -1,16 +1,15 @@
 import type { AdfNode } from './adf-document.ts'
-import type { ConvertErrorPath, Result } from './result.ts'
+import type { ConvertErrorPath } from './result.ts'
-import { emitImageLine } from './markdown-inline.ts'
 import { serializeCanonicalJson } from './canonical-json.ts'
+import { tryImageLine } from './markdown-inline.ts'
 
 const centeredMediaSingle = '{"layout":"center"}'
 const imageAttributes = ['alt', 'type', 'url']
 
-export function emitImage(node: AdfNode, path: ConvertErrorPath): Result<string> | undefined {
+export function tryImage(node: AdfNode, path: ConvertErrorPath): string | undefined {
   const image = imageShape(node)
   if (image === undefined) return undefined
-  const line = emitImageLine(image.alt, image.url, [...path, 'content', 0])
+  return tryImageLine(image.alt, image.url, [...path, 'content', 0])
-  return line.ok ? line : undefined
 }
 
 function imageShape(node: AdfNode): { alt: string | undefined; url: string } | undefined {

src/markdown-inline.ts

@@ -1,37 +1,53 @@
 import type { AdfMark, AdfNode } from './adf-document.ts'
-import { assembleInlineLine, type InlineSegment, type LineContainer } from './markdown-escaping.ts'
+import type { InlineDirective } from './inline-directives.ts'
+import { assembleInlineLine, type InlineEscaping, type InlineSegment, type LineContainer } from './markdown-escaping.ts'
+import { inlineDirective, markDirective, spellInlineNodeAttributes, spellMarkAttributes } from './inline-directives.ts'
 import { largestNesting } from './nesting.ts'
 import { claimsLine, holdsControlCharacter, holdsEntityReference, holdsNullCharacter, isAutolink, isUnicodeWhitespace } from './commonmark-grammar.ts'
 import { failure, success, type ConvertErrorPath, type Result } from './result.ts'
 import { longestBacktickRun } from './backtick-runs.ts'
 import { serializeCanonicalJson } from './canonical-json.ts'
+import { spellAttributes, spellStringAttribute } from './directive-attributes.ts'
 
 type InlineContext = {
   atBlockEnd: boolean
-  container: LineContainer
+  bracketed: boolean
-  inLinkText: boolean
   path: ConvertErrorPath
+  spansLines: boolean
 }
 
 type InlineRun = { index: number; kind: 'marked'; mark: AdfMark; nodes: AdfNode[] } | { index: number; kind: 'plain'; node: AdfNode }
 
+const emphasisSpellings: Readonly<Record<string, string>> = { em: '_', strike: '~~', strong: '**' }
 const linkAttributes = ['href', 'title']
 
 export function emitInlineLine(nodes: readonly AdfNode[], container: LineContainer, path: ConvertErrorPath): Result<string> {
-  const segments = emitRun(nodes, 0, 0, { atBlockEnd: true, container, inLinkText: false, path })
+  const segments = lineSegments(nodes, container, path)
   if (!segments.ok) return segments
   return finishLine(segments.value, container, path)
 }
 
-export function emitImageLine(alt: string | undefined, href: string, path: ConvertErrorPath): Result<string> {
+export function tryPipeCell(nodes: readonly AdfNode[], path: ConvertErrorPath): string | undefined {
-  if (alt !== undefined && /^[ \t]|[ \t]$|[\n\r]/.test(alt)) {
+  const segments = lineSegments(nodes, 'table-cell', path)
-    return failure('unspellable-whitespace', 'a media alt holds whitespace no image description spells', path)
+  if (!segments.ok) return undefined
+  if (segments.value.some((segment) => segment.escaping === 'none' && segment.text.includes('|'))) return undefined
+  const line = finishLine(segments.value, 'table-cell', path)
+  return line.ok ? line.value : undefined
 }
-  if (alt !== undefined && holdsNullCharacter(alt)) return failure('unspellable-character', 'a media alt holds a null character CommonMark replaces', path)
+
+export function tryImageLine(alt: string | undefined, href: string, path: ConvertErrorPath): string | undefined {
+  if (alt !== undefined && (/^[ \t]|[ \t]$|[\n\r]/.test(alt) || holdsNullCharacter(alt))) return undefined
   const destination = spellDestination(href, path)
-  if (!destination.ok) return destination
+  if (!destination.ok) return undefined
-  const description: InlineSegment[] = alt === undefined ? [] : [{ kind: 'link-text', text: alt }]
+  const description: InlineSegment[] = alt === undefined ? [] : [{ escaping: 'bracketed', text: alt }]
-  return finishLine([{ kind: 'syntax', text: '![' }, ...description, { kind: 'syntax', text: `](${destination.value})` }], 'paragraph', path)
+  const line = finishLine([syntax('!['), ...description, syntax(`](${destination.value})`)], 'paragraph', path)
+  return line.ok ? line.value : undefined
+}
+
+function lineSegments(nodes: readonly AdfNode[], container: LineContainer, path: ConvertErrorPath): Result<InlineSegment[]> {
+  const segments = emitRun(nodes, 0, 0, { atBlockEnd: true, bracketed: false, path, spansLines: container === 'paragraph' })
+  if (!segments.ok) return segments
+  return success(carryStrippedWhitespace(segments.value))
 }
 
 function finishLine(segments: readonly InlineSegment[], container: LineContainer, path: ConvertErrorPath): Result<string> {
@@ -51,6 +67,51 @@ function finishLine(segments: readonly InlineSegment[], container: LineContainer
   return success(line)
 }
 
+// spec/flavour.md, Inline nodes.
+function carryStrippedWhitespace(segments: readonly InlineSegment[]): InlineSegment[] {
+  const carried: InlineSegment[] = []
+  for (const [index, segment] of segments.entries()) {
+    const previous = segments[index - 1]
+    const next = segments[index + 1]
+    const leading = previous === undefined || previous.text.includes('\n')
+    const trailing = next === undefined || next.text.includes('\n')
+    carried.push(...carryEdges(segment, leading, trailing))
+  }
+  return carried
+}
+
+function carryEdges(segment: InlineSegment, leading: boolean, trailing: boolean): InlineSegment[] {
+  if (segment.escaping !== 'backslash' && segment.escaping !== 'bracketed') return [segment]
+  const head = leading ? (/^[ \t]+/.exec(segment.text)?.[0] ?? '') : ''
+  const body = segment.text.slice(head.length)
+  const tail = trailing ? (/[ \t]+$/.exec(body)?.[0] ?? '') : ''
+  const middle = body.slice(0, body.length - tail.length)
+  const edges: InlineSegment[] = []
+  if (head !== '') edges.push(carriedText(head))
+  if (middle !== '') edges.push({ escaping: segment.escaping, text: middle })
+  if (tail !== '') edges.push(carriedText(tail))
+  return edges
+}
+
+function carriedText(text: string): InlineSegment {
+  return { escaping: 'attribute', text: spellLeafDirective('text', spellAttributes([['text', spellStringAttribute(text)]])) }
+}
+
+function spellLeafDirective(name: string, attributes: string): string {
+  return `:${name}${attributes === '' ? '{}' : attributes}`
+}
+
+function syntax(text: string): InlineSegment {
+  return { escaping: 'none', text }
+}
+
+function refuseContentAndText(node: AdfNode, path: ConvertErrorPath): Result<null> {
+  if ((node.content ?? []).length > 0 || node.text !== undefined) {
+    return failure('unsupported-node-shape', `a ${node.type} node holds neither content nor text`, path)
+  }
+  return success(null)
+}
+
 function emitRun(nodes: readonly AdfNode[], depth: number, firstIndex: number, context: InlineContext): Result<InlineSegment[]> {
   if (depth > largestNesting) {
     return failure('unsupported-node-shape', `the marks nest deeper than the ${largestNesting} levels the emitter carries`, context.path)
@@ -88,38 +149,85 @@ function nodePath(context: InlineContext, index: number): ConvertErrorPath {
 
 function emitLeaf(node: AdfNode, context: InlineContext, index: number): Result<InlineSegment[]> {
   const path = nodePath(context, index)
-  if (node.type !== 'hardBreak' && node.type !== 'text') {
-    return failure('unsupported-node-type', `the canonical form spells no inline node of type ${node.type}`, path)
-  }
-  const unspelled = Object.keys(node.attrs ?? {})[0]
-  if (unspelled !== undefined) {
-    return failure('unspelled-node-attribute', `the ${node.type} attribute ${unspelled} has no canonical markdown spelling`, path)
-  }
   const types = (node.marks ?? []).map((mark) => mark.type)
   if (new Set(types).size !== types.length) return failure('unsupported-node-shape', `a ${node.type} node carries one mark type twice`, path)
-  if (node.type === 'hardBreak') {
+  if (node.type === 'text') return emitText(node, context, path)
-    if (context.container === 'paragraph' && !context.atBlockEnd) return success([{ kind: 'syntax', text: '\\\n' }])
+  if (node.type === 'hardBreak') return emitHardBreak(node, context, path)
-    return success([{ kind: 'syntax', text: ':hardBreak{}' }])
+  const directive = inlineDirective(node.type)
+  if (directive === undefined) return failure('unsupported-node-type', `the canonical form spells no inline node of type ${node.type}`, path)
+  return emitInlineDirective(node, directive, path)
 }
-  if (typeof node.text !== 'string' || node.text === '') return failure('unsupported-node-shape', 'a text node carries no text', path)
+
-  if ((node.content ?? []).length > 0) return failure('unsupported-node-shape', 'a text node carries content', path)
+function emitHardBreak(node: AdfNode, context: InlineContext, path: ConvertErrorPath): Result<InlineSegment[]> {
-  if (/[\n\r]/.test(node.text)) return failure('unspellable-whitespace', 'a text node holds a newline CommonMark cannot spell', path)
+  const unspelled = Object.keys(node.attrs ?? {})[0]
+  if (unspelled !== undefined) return failure('unspelled-node-attribute', `the hardBreak attribute ${unspelled} has no canonical markdown spelling`, path)
+  const empty = refuseContentAndText(node, path)
+  if (!empty.ok) return empty
+  if (context.spansLines && !context.atBlockEnd) return success([syntax('\\\n')])
+  return success([syntax(spellLeafDirective('hardBreak', ''))])
+}
+
+function emitInlineDirective(node: AdfNode, directive: InlineDirective, path: ConvertErrorPath): Result<InlineSegment[]> {
+  const empty = refuseContentAndText(node, path)
+  if (!empty.ok) return empty
+  const attributes = spellInlineNodeAttributes(node, directive, path)
+  if (!attributes.ok) return attributes
+  const slot = directive.slot === undefined ? undefined : node.attrs?.[directive.slot]
+  if (slot === undefined) return success([syntax(spellLeafDirective(node.type, attributes.value))])
+  if (typeof slot !== 'string') return failure('unsupported-node-shape', `the ${node.type} attribute ${directive.slot} holds no string`, path)
+  if (/[\n\r]/.test(slot)) return failure('unspellable-whitespace', `a ${node.type} content slot holds a newline no inline directive spans`, path)
+  if (holdsNullCharacter(slot)) return failure('unspellable-character', `a ${node.type} content slot holds a null character CommonMark replaces`, path)
+  const content: InlineSegment[] = slot === '' ? [] : [{ escaping: 'bracketed', text: slot }]
+  return success([syntax(`:${node.type}[`), ...content, syntax(`]${attributes.value}`)])
+}
+
+function emitText(node: AdfNode, context: InlineContext, path: ConvertErrorPath): Result<InlineSegment[]> {
+  const unspelled = Object.keys(node.attrs ?? {})[0]
+  if (unspelled !== undefined) return failure('unspelled-node-attribute', `the text attribute ${unspelled} has no canonical markdown spelling`, path)
+  if (typeof node.text !== 'string' || node.text === '') return failure('unsupported-node-shape', 'a text node holds text', path)
+  if ((node.content ?? []).length > 0) return failure('unsupported-node-shape', 'a text node holds no content', path)
+  if (/\r/.test(node.text)) return failure('unspellable-whitespace', 'a text node holds a carriage return CommonMark rewrites', path)
   if (holdsNullCharacter(node.text)) return failure('unspellable-character', 'a text node holds a null character CommonMark replaces', path)
-  return success([{ kind: context.inLinkText ? 'link-text' : 'literal', text: node.text }])
+  const escaping: InlineEscaping = context.bracketed ? 'bracketed' : 'backslash'
+  const parts = node.text.split(/(\n+)/).filter((part) => part !== '')
+  return success(parts.map((part) => (part.startsWith('\n') ? carriedText(part) : { escaping, text: part })))
 }
 
 function emitMarkedRun(nodes: readonly AdfNode[], mark: AdfMark, depth: number, index: number, context: InlineContext): Result<InlineSegment[]> {
   if (mark.type === 'code') return emitCodeSpan(nodes, depth, nodePath(context, index))
   if (mark.type === 'link') return emitLink(nodes, mark, depth, index, context)
   const path = nodePath(context, index)
-  const spelling = mark.type === 'em' ? '_' : mark.type === 'strike' ? '~~' : mark.type === 'strong' ? '**' : undefined
+  const spelling = Object.hasOwn(emphasisSpellings, mark.type) ? emphasisSpellings[mark.type] : undefined
-  if (spelling === undefined) return failure('unspellable-mark', `no markdown spelling holds the ${mark.type} mark`, path)
+  if (spelling !== undefined) return emitEmphasis(nodes, mark, spelling, depth, index, context, path)
+  const vocabulary = markDirective(mark.type)
+  if (vocabulary === undefined) return failure('unspellable-mark', `no markdown spelling holds the ${mark.type} mark`, path)
+  const attributes = spellMarkAttributes(mark, vocabulary, path)
+  if (!attributes.ok) return attributes
+  const inner = emitRun(nodes, depth + 1, index, { ...context, bracketed: true, spansLines: false })
+  if (!inner.ok) return inner
+  return success([syntax(`:${mark.type}[`), ...inner.value, syntax(`]${attributes.value}`)])
+}
+
+function emitEmphasis(
+  nodes: readonly AdfNode[],
+  mark: AdfMark,
+  spelling: string,
+  depth: number,
+  index: number,
+  context: InlineContext,
+  path: ConvertErrorPath,
+): Result<InlineSegment[]> {
   if (Object.keys(mark.attrs ?? {}).length > 0) return failure('unspellable-mark', `the ${mark.type} spelling holds no attributes`, path)
   const inner = emitRun(nodes, depth + 1, index, context)
   if (!inner.ok) return inner
-  const text = inner.value.map((segment) => segment.text).join('')
+  const carried = carryStrippedWhitespace(inner.value)
+  const text = carried.map((segment) => segment.text).join('')
   if (holdsEdgeWhitespace(text)) return failure('unspellable-whitespace', `the ${mark.type} spelling cannot open or close beside whitespace`, path)
-  return success([{ kind: 'emphasis-open', mark: mark.type, text: spelling }, ...inner.value, { kind: 'emphasis-close', mark: mark.type, text: spelling }])
+  return success([
+    { emphasis: 'open', escaping: 'none', mark: mark.type, text: spelling },
+    ...carried,
+    { emphasis: 'close', escaping: 'none', mark: mark.type, text: spelling },
+  ])
 }
 
 function emitCodeSpan(nodes: readonly AdfNode[], depth: number, path: ConvertErrorPath): Result<InlineSegment[]> {
@@ -133,7 +241,7 @@ function emitCodeSpan(nodes: readonly AdfNode[], depth: number, path: ConvertErr
   if (holdsNullCharacter(text)) return failure('unspellable-character', 'a code span holds a null character CommonMark replaces', path)
   const fence = '`'.repeat(longestBacktickRun(text) + 1)
   const padded = needsPadding(text) ? ` ${text} ` : text
-  return success([{ kind: 'syntax', text: `${fence}${padded}${fence}` }])
+  return success([syntax(`${fence}${padded}${fence}`)])
 }
 
 function holdsEdgeWhitespace(text: string): boolean {
@@ -155,14 +263,14 @@ function emitLink(nodes: readonly AdfNode[], mark: AdfMark, depth: number, index
   if (title !== undefined && typeof title !== 'string') return failure('unsupported-node-shape', 'a link title is no string', path)
   const node = nodes[0]
   const bare = nodes.length === 1 && node !== undefined && node.type === 'text' && node.text === href && (node.marks ?? []).length === depth + 1
-  if (bare && title === undefined && isAutolink(href) && !holdsEntityReference(href)) return success([{ kind: 'syntax', text: `<${href}>` }])
+  if (bare && title === undefined && isAutolink(href) && !holdsEntityReference(href)) return success([syntax(`<${href}>`)])
   const destination = spellDestination(href, path)
   if (!destination.ok) return destination
   const spelledTitle = title === undefined ? success('') : spellTitle(title, path)
   if (!spelledTitle.ok) return spelledTitle
-  const inner = emitRun(nodes, depth + 1, index, { ...context, inLinkText: true })
+  const inner = emitRun(nodes, depth + 1, index, { ...context, bracketed: true })
   if (!inner.ok) return inner
-  return success([{ kind: 'syntax', text: '[' }, ...inner.value, { kind: 'syntax', text: `](${destination.value}${spelledTitle.value})` }])
+  return success([syntax('['), ...inner.value, syntax(`](${destination.value}${spelledTitle.value})`)])
 }
 
 function spellDestination(href: string, path: ConvertErrorPath): Result<string> {

src/markdown-pipe-table.ts

@@ -1,9 +1,8 @@
 import type { AdfNode } from './adf-document.ts'
-import type { JsonValue } from './json-value.ts'
+import { tryPipeCell } from './markdown-inline.ts'
-import { emitInlineLine } from './markdown-inline.ts'
+import type { ConvertErrorPath } from './result.ts'
-import { success, type ConvertErrorPath, type Result } from './result.ts'
 
-export function emitPipeTable(node: AdfNode, path: ConvertErrorPath): Result<string> | undefined {
+export function tryPipeTable(node: AdfNode, path: ConvertErrorPath): string | undefined {
   const rows = pipeRows(node)
   if (rows === undefined) return undefined
   const lines: string[] = []
@@ -11,14 +10,14 @@ export function emitPipeTable(node: AdfNode, path: ConvertErrorPath): Result<str
     const cells: string[] = []
     for (const [cellIndex, paragraph] of row.entries()) {
       const content = paragraph.content ?? []
-      const line = content.length === 0 ? success('') : emitInlineLine(content, 'table-cell', [...path, 'content', rowIndex, 'content', cellIndex, 'content', 0])
+      const line = content.length === 0 ? '' : tryPipeCell(content, [...path, 'content', rowIndex, 'content', cellIndex, 'content', 0])
-      if (!line.ok) return line
+      if (line === undefined) return undefined
-      cells.push(line.value)
+      cells.push(line)
     }
     lines.push(`| ${cells.join(' | ')} |`)
     if (rowIndex === 0) lines.push(`| ${cells.map(() => '---').join(' | ')} |`)
   }
-  return success(lines.join('\n'))
+  return lines.join('\n')
 }
 
 function pipeRows(node: AdfNode): AdfNode[][] | undefined {
@@ -49,18 +48,5 @@ function plainParagraph(cell: AdfNode): AdfNode | undefined {
   const content = cell.content ?? []
   const paragraph = content[0]
   if (paragraph === undefined || content.length !== 1 || paragraph.type !== 'paragraph' || !isPlain(paragraph)) return undefined
-  return (paragraph.content ?? []).some(spellsPipeAsSyntax) ? undefined : paragraph
+  return paragraph
-}
-
-// A pipe the inline layer emits as syntax takes no backslash, so the cell has no pipe spelling.
-function spellsPipeAsSyntax(child: AdfNode): boolean {
-  return (child.marks ?? []).some((mark) => {
-    if (mark.type === 'code') return (child.text ?? '').includes('|')
-    if (mark.type !== 'link') return false
-    return holdsPipe(mark.attrs?.['href']) || holdsPipe(mark.attrs?.['title'])
-  })
-}
-
-function holdsPipe(value: JsonValue | undefined): boolean {
-  return typeof value === 'string' && value.includes('|')
 }

todo.md

@@ -71,13 +71,13 @@ detail is settled at its own milestone.
         answer on a CommonMark block beside a directive block (1d) drops into the same seam. Give
         the emitter's refusals a corpus home while the directories grow: `corpus/unspellable/`,
         a `.json` beside the `ConvertErrorCode` it must return, the emitter half of `corpus/errors/`.
-  - [ ] **2c — Inline nodes and marks.** `inline-nodes/` green. `InlineSegment.kind` splits into
+  - [ ] **2c — Inline nodes and marks.** `inline-nodes/` green. `InlineSegment` splits into its
-        its two axes here — escapability (`attribute` for `:text{text="…"}`, `backslash`, `none`)
+        two axes — escapability (`attribute` for `:text{text="…"}`, `backslash`, `bracketed`,
-        and the emphasis role — rather than gaining a third value that means one of each. A lone
+        `none`) and the emphasis role. A lone surrogate in a text node emits verbatim and becomes
-        surrogate in a text node emits verbatim and becomes U+FFFD on any UTF-8 encode, a §2 break
+        U+FFFD on any UTF-8 encode, a §2 break plain text still holds open — attribute values
-        plain text still holds open — attribute values already escape it. The inline directives
+        already escape it. The pipe form's fallback reads the emitted segments rather than naming
-        arriving here emit their attribute values as syntax, so `spellsPipeAsSyntax` grows with
+        the nodes whose attribute values spell a pipe as syntax, so 2e's `\u007c` narrows it in
-        them alongside the `|` escaping `spec/flavour.md` already mandates in a pipe cell.
+        one place.
   - [ ] **2d — The opaque carry** (§3). Fixtures and emitter together, into
         `corpus/round-trip/opaque-carry/`: an unknown node in both positions, the reserved `adf`
         info string, and the `codeBlock` whose language is `adf`.
@@ -90,6 +90,13 @@ detail is settled at its own milestone.
         paragraph line inside a container body shaped like a closing fence (`:::`, `::: x`).
         Guard `fenceNestingFault`'s bare-run pop here too — a run shorter than the open fence is a
         fault, not a close — which today's emitter cannot reach.
+        Two moves land before the attribute spelling changes. One mark vocabulary:
+        `emphasisSpellings`, `linkAttributes` and the `code`/`link` names join
+        `inline-directives.ts`, which holds four of the nine marks while the rest are branch
+        literals in the emitter — and the parser (3) needs every name to make `:em[x]` the named
+        error `spec/flavour.md` promises. And `escaping: 'attribute'` earns its keep at the
+        `\u007c` rule or collapses into `none`: nothing the escaper does tells the two apart
+        today, since a carried segment holds only spaces, tabs and newlines.
         The gate gains the collision property here: no two corpus documents may emit the same
         bytes — one spelling for two documents is a round-trip break no parser can undo, and it is
         provable without one. It also settles the emitter's one known approximation: delimiter
@@ -108,11 +115,17 @@ detail is settled at its own milestone.
       assumes — CommonMark flanking, as for `*` — which `spec/flavour.md` does not yet pin.
       `src/` gets its hierarchy at the same split — `adf/`,
       `markdown/`, `html/`, the grammar module shared inside `markdown/` — while the rename is
-      still mechanical. `block-directives.ts` is the one file that does not move whole: the node
+      still mechanical. Three files do not move whole: `block-directives.ts` and
-      table is ADF knowledge milestones 6-7 need too and belongs in `adf/`, `spellDirectiveHeader`
+      `inline-directives.ts` each hold a node table milestones 6-7 need in `adf/` beside a
-      in `markdown/`. `AttributeKind` stays above both — it is the vocabulary a string-typed
+      markdown spelling that belongs in `markdown/`, and `directive-attributes.ts` fuses the
-      attribute grammar needs, which is why HTML will want it too, not a markdown spelling. The table is a second copy of `spec/flavour.md`'s prose with no drift guard,
+      format-neutral conformance walk (`vocabularyPairs`) with the markdown value spelling HTML
-      and a mistyped attribute name degrades into a false refusal no test catches.
+      has no use for. `spellDestination`, `spellTitle` and `balanced` leave `markdown-inline.ts`
+      here too — CommonMark destination spelling `emitLink` and `tryImageLine` share, and the six
+      concerns that file carries are one fewer for it. `AttributeKind` and `AttributeVocabulary`
+      stay above all of it — the vocabulary a string-typed attribute grammar needs, which is why
+      HTML will want them too, not a markdown spelling. Both node tables are a second copy of
+      `spec/flavour.md`'s prose with no drift guard, and a mistyped attribute name degrades into a
+      false refusal no test catches.
 - [ ] **4 — Round-trip property tests** over the corpus, both ways — the thing that proves 2 and
       3. Editor-normal (§2) gets its implementation here — `toEditorNormal(doc)` and the equality
       the round-trip asserts, which over normalized input is the canonical serializer's compact