@peaceroad/markdown-it-numbering-ul-regarded-as-ol 0.4.0 → 0.4.2

package/README.md

@@ -45,6 +45,7 @@ The plugin supports the following marker types.
     - filled-squared-upper-latin: `🅰`, `🅱`, `🅲`, pattern: enclosed
     - fullwidth-lower-roman: `ⅰ`, `ⅱ`, `ⅲ`, pattern: fullwidth
     - fullwidth-upper-roman: `Ⅰ`, `Ⅱ`, `Ⅲ`, pattern: fullwidth
+    - fullwidth-decimal: `０`, `１`, `２`, pattern: fullwidth
     - japanese-informal: `一`, `二`, `三`, pattern: fullwidth
     - katakana: `ア`, `イ`, `ウ`, pattern: fullwidth
     - katakana-iroha: `イ`, `ロ`, `ハ`, pattern: fullwidth
@@ -90,10 +91,11 @@ You can customize the conversion using options.
 - `hasListStyleNone` (boolean) — When the plugin emits `role="list"`, also add `style="list-style: none;"` to the `<ol>`.
 - `omitMarkerMetadata` (boolean) — If `true`, omit the `data-marker-prefix` / `data-marker-suffix` attributes.
 - `addMarkerStyleToClass` (boolean) — When `true`, append suffix-style information to the generated class name (e.g. `ol-decimal-with-round-round`). When `false` (default) the class stays as `ol-decimal`.
-- `enableLiteralNumberingFix` (boolean) — Enable literal nested list recovery (for example, nested lists starting with 2 or greater). This is opt-in; it only applies inside list items, evaluates indentation relative to the parent list marker (marker width + 0–3 spaces), and does not convert code blocks (indent >= marker width + 4).
-  - Compatibility note: switching the default from `false` to `true` changes rendered HTML by design. On the current test corpus (`394` markdown cases), `18` cases change.
-  - Changed files in that comparison: `examples-default-14-repeated-numbers.txt` (`7`), `examples-option-literal-numbering-attrs.txt` (`1`), `examples-option-literal-numbering-fix-disabled.txt` (`2`), `examples-option-literal-numbering-fix.txt` (`3`), `examples-option-literal-numbering-indent.txt` (`4`).
-  - Recommendation: keep the default as `false` for patch/minor releases; if changing the default to `true`, treat it as a breaking change and release a major version.
+- `enableLiteralNumberingFix` (boolean) — Enable literal nested list recovery (for example, nested lists starting with 2 or greater). Default is `false` (legacy-compatible); set it to `true` to normalize literal nested numbering. It only applies inside list items, evaluates indentation relative to the parent list marker (marker width + 0–3 spaces), and does not convert code blocks (indent >= marker width + 4).
+  - Compatibility note: compared with the default/legacy mode (`enableLiteralNumberingFix: false`), enabling `true` changes rendered HTML by design. On the current test corpus (`399` markdown cases), `19` cases change.
+  - Changed files in that comparison: `examples-default-14-repeated-numbers.txt` (`7`), `examples-option-literal-numbering-attrs-disabled.txt` (`1`), `examples-option-literal-numbering-attrs.txt` (`1`), `examples-option-literal-numbering-fix-disabled.txt` (`2`), `examples-option-literal-numbering-fix.txt` (`4`), `examples-option-literal-numbering-indent.txt` (`4`).
+  - Migration tip: if you enable `enableLiteralNumberingFix: true`, treat it as a breaking-output change for snapshots/downstream HTML.
+  - Detailed notes: `docs/enable-literal-numbering-fix.md`.
 
 ## Description lists conversion behavior
 
@@ -145,7 +147,7 @@ Custom marker conversion example:
 - ③ Third
 
 [HTML]
-<ol role="list" class="ol-filled-circled-decimal">
+<ol role="list" class="ol-circled-decimal">
 <li><span class="li-num" aria-hidden="true">①</span> First</li>
 <li><span class="li-num" aria-hidden="true">②</span> Second</li>
 <li><span class="li-num" aria-hidden="true">③</span> Third</li>
@@ -268,9 +270,9 @@ When description lists are enabled the plugin can convert the following patterns
 
 ```
 [Markdown]
-- **Term 1**  
+- **Term 1**
 Description text for term 1
-- **Term 2**  
+- **Term 2**
 Description text for term 2
 
 [HTML]
@@ -349,3 +351,5 @@ md.use(mditNumberingUl)
 const html = md.render(`- a. First\n- b. Second`)
 console.log(html)
 ```
+
+Register this plugin only once per `markdown-it` instance. A second registration fails fast because running the conversion pipeline twice on the same token stream would corrupt already-converted lists.

package/index.js

@@ -7,7 +7,17 @@ import { processHtmlBlocks } from './src/phase4-html-blocks.js'
 import { generateSpans } from './src/phase5-spans.js'
 import { normalizeLiteralOrderedLists } from './src/preprocess-literal-lists.js'
 
+const INSTALL_FLAG = Symbol.for('@peaceroad/markdown-it-numbering-ul-regarded-as-ol/installed')
+
 const mditNumberingUl = (md, option) => {
+  if (md[INSTALL_FLAG]) {
+    throw new Error('@peaceroad/markdown-it-numbering-ul-regarded-as-ol is already registered on this markdown-it instance')
+  }
+  Object.defineProperty(md, INSTALL_FLAG, {
+    value: true,
+    configurable: false
+  })
+
   const opt = {
     // Core options
     descriptionList: false,       // Convert **Term** patterns to <dl>/<dt>/<dd>
@@ -21,11 +31,19 @@ const mditNumberingUl = (md, option) => {
     useCounterStyle: false,       // true=users will use @counter-style; suppress marker spans and role attr
     addMarkerStyleToClass: false, // true=append -with-* marker style suffix to class names
     enableLiteralNumberingFix: false, // true=normalize nested lists that don't start at 1 (opt-in)
-    
+
     // Override with user options
     ...option
   }
 
+  const normalizedMarkerSpanClass = typeof opt.markerSpanClass === 'string'
+    ? opt.markerSpanClass.trim()
+    : ''
+  opt.markerSpanClass = normalizedMarkerSpanClass || 'li-num'
+  opt.descriptionListDivClass = typeof opt.descriptionListDivClass === 'string'
+    ? opt.descriptionListDivClass.trim()
+    : ''
+
   const addRuleAfter = (ruler, afterName, ruleName, fn) => {
     try {
       ruler.after(afterName, ruleName, fn)
@@ -35,9 +53,6 @@ const mditNumberingUl = (md, option) => {
   }
 
   const dlProcessor = (state) => {
-    if (!state.env) {
-      state.env = {}
-    }
     if (!opt.descriptionList && !opt.descriptionListWithDiv) {
       return true
     }
@@ -46,54 +61,74 @@ const mditNumberingUl = (md, option) => {
   }
 
   const listProcessor = (state) => {
-    // Initialize state.env
-    if (!state.env) {
-      state.env = {}
-    }
-
     const tokens = state.tokens
+    let hasAnyList = false
+    for (let i = 0; i < tokens.length; i++) {
+      const type = tokens[i]?.type
+      if (type === 'bullet_list_open' || type === 'ordered_list_open') {
+        hasAnyList = true
+        break
+      }
+    }
+    if (!hasAnyList) {
+      return true
+    }
 
     // Normalize literal nested ordered lists (markdown-it only creates nested lists when they start at 1)
-    normalizeLiteralOrderedLists(tokens, opt)
-    
+    if (opt.enableLiteralNumberingFix) {
+      normalizeLiteralOrderedLists(tokens, opt)
+    }
+
     // ===== PHASE 1: List Structure Analysis =====
     // Analyze marker detection and structure without token conversion
-    const listInfos = analyzeListStructure(tokens, opt)
-    
+    const listInfos = analyzeListStructure(tokens)
+
     // ===== PHASE 2: Token Conversion =====
     // Convert bullet_list to ordered_list based on Phase1 analysis
     // Note: simplifyNestedBulletLists removes tokens, changing indices
     convertLists(tokens, listInfos, opt)
-    
-    // ===== PHASE 3: Add Attributes =====
-    // Add type, class, data-* attributes to converted lists
-    // Use markerInfo stored on list tokens (safe after Phase2 mutations)
-    addAttributes(tokens, opt)
-    
-    // ===== PHASE 4: HTML Block Processing =====
-    // Remove indents from HTML blocks in lists and normalize line breaks
+
+    let hasOrderedList = false
     let hasNestedHtmlBlock = false
     for (let i = 0; i < tokens.length; i++) {
       const token = tokens[i]
-      if (token.type === 'html_block' && token.level > 0) {
+      if (!hasOrderedList && token.type === 'ordered_list_open') {
+        hasOrderedList = true
+      }
+      if (!hasNestedHtmlBlock && token.type === 'html_block' && token.level > 0) {
         hasNestedHtmlBlock = true
+      }
+      if (hasOrderedList && hasNestedHtmlBlock) {
         break
       }
     }
+
+    // ===== PHASE 3: Add Attributes =====
+    // Add type, class, data-* attributes to converted lists
+    // Use markerInfo stored on list tokens (safe after Phase2 mutations)
+    let closeMap = null
+    if (hasOrderedList) {
+      closeMap = addAttributes(tokens, opt)
+    }
+
+    // ===== PHASE 4: HTML Block Processing =====
+    // Remove indents from HTML blocks in lists and normalize line breaks
     if (hasNestedHtmlBlock) {
       processHtmlBlocks(state)
     }
-    
+
     // ===== PHASE 5: Span Generation =====
     // Generate marker spans in alwaysMarkerSpan mode
-    generateSpans(tokens, opt)
-    
+    if (hasOrderedList && !opt.useCounterStyle) {
+      generateSpans(tokens, opt, closeMap?.listCloseByOpen || null)
+    }
+
     return true
   }
 
   md.core.ruler.before('inline', 'numbering_dl_parser', dlProcessor)
   md.core.ruler.after('numbering_dl_parser', 'numbering_ul_phases', listProcessor)
-  
+
   // Description list: Move paragraph attributes to dl and add custom renderers
   if (opt.descriptionList || opt.descriptionListWithDiv) {
     // Move paragraph attributes to dl (after inline and any attribute plugins)
@@ -101,7 +136,7 @@ const mditNumberingUl = (md, option) => {
       moveParagraphAttributesToDL(state.tokens)
       return true
     }
-    
+
     addRuleAfter(md.core.ruler, 'curly_attributes', 'numbering_dl_attrs', dlAttrProcessor)
   }
 }

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@peaceroad/markdown-it-numbering-ul-regarded-as-ol",
-  "version": "0.4.0",
+  "version": "0.4.2",
   "description": "This markdown-it plugin regard ul element with numbering lists as ol element.",
   "main": "index.js",
   "type": "module",
@@ -27,7 +27,8 @@
   },
   "homepage": "https://github.com/peaceroad/p7d-markdown-it-numbering-ul-regarded-as-ol#readme",
   "devDependencies": {
-    "@peaceroad/markdown-it-strong-ja": "^0.7.2",
+    "@peaceroad/markdown-it-cjk-breaks-mod": "^0.1.10",
+    "@peaceroad/markdown-it-strong-ja": "^0.9.0",
     "markdown-it": "^14.1.0",
     "markdown-it-attrs": "^4.3.1",
     "markdown-it-deflist": "^3.0.0"

package/src/list-helpers.js

@@ -49,11 +49,13 @@ export function findListItemEnd(tokens, startIndex) {
 /**
  * Build close-index maps for list and list_item tokens.
  * @param {Array} tokens - Token array
- * @returns {{ listCloseByOpen: number[], listItemCloseByOpen: number[] }}
+ * @returns {{ listCloseByOpen: Int32Array, listItemCloseByOpen: Int32Array }}
  */
 export function buildListCloseIndexMap(tokens) {
-  const listCloseByOpen = new Array(tokens.length).fill(-1)
-  const listItemCloseByOpen = new Array(tokens.length).fill(-1)
+  const listCloseByOpen = new Int32Array(tokens.length)
+  listCloseByOpen.fill(-1)
+  const listItemCloseByOpen = new Int32Array(tokens.length)
+  listItemCloseByOpen.fill(-1)
   const bulletStack = []
   const orderedStack = []
   const listItemStack = []

package/src/phase0-description-list.js

@@ -14,6 +14,8 @@ const LEADING_ATTR_BLOCK_REGEX = /^[ \t]*\{([^}]+)\}/
 const STANDALONE_ATTR_BLOCK_REGEX = /^\{([^}]+)\}$/
 const TRAILING_TWO_SPACES_REGEX = / {2,}$/
 const TRAILING_BACKSLASH_BREAK_REGEX = /\\\s*$/
+const DL_TERM_INLINE_REGEX = /^\*\*(.*?)\*\*(.*)/s
+const TRAILING_LIST_ATTR_LINE_REGEX = /\n\s*\{([^}]+)\}\s*$/
 
 const parseAttrString = (attrStr) => {
   if (!attrStr || typeof attrStr !== 'string') {
@@ -123,9 +125,11 @@ const hasMeaningfulDescriptionContent = (text) => {
   if (typeof text !== 'string' || text.trim().length === 0) {
     return false
   }
+  if (!text.includes('{')) {
+    return true
+  }
 
   const lines = text.split('\n')
-  const renderedLines = []
   for (const line of lines) {
     const trimmed = line.trim()
     if (!trimmed) {
@@ -138,9 +142,9 @@ const hasMeaningfulDescriptionContent = (text) => {
         continue
       }
     }
-    renderedLines.push(trimmed)
+    return true
   }
-  return renderedLines.join('\n').trim().length > 0
+  return false
 }
 
 const copyMap = (target, source) => {
@@ -159,10 +163,24 @@ export const processDescriptionList = (tokens, opt) => {
   if (!opt.descriptionList && !opt.descriptionListWithDiv) {
     return
   }
+  if (!Array.isArray(tokens) || tokens.length === 0) {
+    return
+  }
+
+  let firstBulletIndex = -1
+  for (let i = 0; i < tokens.length; i++) {
+    if (tokens[i].type === 'bullet_list_open') {
+      firstBulletIndex = i
+      break
+    }
+  }
+  if (firstBulletIndex === -1) {
+    return
+  }
 
   // Find bullet_lists and check if they match DL pattern
   // Process in single pass: check and convert together to avoid duplicate scanning
-  let i = 0
+  let i = firstBulletIndex
   while (i < tokens.length) {
     if (tokens[i].type === 'bullet_list_open') {
       const listEnd = findListEnd(tokens, i)
@@ -253,6 +271,9 @@ const checkAndConvertToDL = (tokens, listStart, listEnd, opt) => {
   // First pass: validate all items match DL pattern
   let hasAnyDLItem = false
   const itemRanges = collectListItemRanges(tokens, listStart, listEnd)
+  if (itemRanges.length === 0) {
+    return { nextIndex: listEnd + 1 }
+  }
 
   for (const range of itemRanges) {
     const itemStart = range.open
@@ -329,9 +350,12 @@ const checkAndConvertToDL = (tokens, listStart, listEnd, opt) => {
  */
 const isDLPattern = (content) => {
   if (!content) return { isMatch: false, afterStrong: null }
+  if (content.length < 4 || content[0] !== '*' || content[1] !== '*') {
+    return { isMatch: false, afterStrong: null }
+  }
   
   // Match **Term** pattern (allow spaces inside for markdown-it-strong-ja compatibility)
-  const match = content.match(/^\*\*(.*?)\*\*(.*)/s)  // s flag for including newlines
+  const match = content.match(DL_TERM_INLINE_REGEX)
   if (!match) return { isMatch: false, afterStrong: null }
   
   const afterStrong = match[2]  // Text after closing **
@@ -450,7 +474,7 @@ const convertListItemToDtDd = (tokens, itemStart, itemEnd, parentLevel, opt) =>
   let descStart = ''
   
   const content = inlineToken.content
-  const match = content.match(/^\*\*(.*?)\*\*(.*)/s)  // s flag for including newlines
+  const match = content.match(DL_TERM_INLINE_REGEX)
   
   if (match) {
     // Trim to avoid leading space when **Term** starts with whitespace (e.g. "** *term*")
@@ -471,7 +495,7 @@ const convertListItemToDtDd = (tokens, itemStart, itemEnd, parentLevel, opt) =>
     // Pattern B: {.attrs} on last line (e.g., "Description\n{.attrs}")
     // This will be processed by markdown-it-attrs and applied to list, not paragraph
     // We need to remove it from description content and save for list-level attrs
-    const lastLineAttrsMatch = afterStrong.match(/\n\s*\{([^}]+)\}\s*$/)
+    const lastLineAttrsMatch = afterStrong.match(TRAILING_LIST_ATTR_LINE_REGEX)
     if (lastLineAttrsMatch) {
       // Parse attributes
       const attrString = lastLineAttrsMatch[1]
@@ -479,7 +503,7 @@ const convertListItemToDtDd = (tokens, itemStart, itemEnd, parentLevel, opt) =>
       if (parsedAttrs.length > 0) {
         listAttrs.push(...parsedAttrs)
         // Remove {.attrs} line from afterStrong
-        afterStrong = afterStrong.replace(/\n\s*\{[^}]+\}\s*$/, '')
+        afterStrong = afterStrong.replace(TRAILING_LIST_ATTR_LINE_REGEX, '')
       }
     }
     
@@ -495,14 +519,12 @@ const convertListItemToDtDd = (tokens, itemStart, itemEnd, parentLevel, opt) =>
   if (!term) return { tokens: result, listAttrs }
   
   // Create div_open if descriptionListWithDiv is enabled
-  if (opt && opt.descriptionListWithDiv) {
+  if (opt.descriptionListWithDiv) {
     const divOpen = new tokens[firstPara].constructor('div_open', 'div', 1)
     divOpen.level = parentLevel + 1
     divOpen.block = true
     copyMap(divOpen, tokens[firstPara])
-    const divClass = typeof opt.descriptionListDivClass === 'string'
-      ? opt.descriptionListDivClass.trim()
-      : ''
+    const divClass = opt.descriptionListDivClass
     if (divClass) {
       divOpen.attrs = [['class', divClass]]
     }
@@ -542,7 +564,7 @@ const convertListItemToDtDd = (tokens, itemStart, itemEnd, parentLevel, opt) =>
   result.push(ddOpen)
   
   // First paragraph in dd (if description exists)
-  if (descStart.trim()) {
+  if (descStart) {
     const pOpen = new tokens[firstPara].constructor('paragraph_open', 'p', 1)
     pOpen.level = parentLevel + 2
     pOpen.block = true  // IMPORTANT: Enable block mode for proper newline rendering
@@ -554,7 +576,7 @@ const convertListItemToDtDd = (tokens, itemStart, itemEnd, parentLevel, opt) =>
     pInline.level = parentLevel + 3
     pInline.block = true  // IMPORTANT: Enable block mode
     const pText = new tokens[firstPara].constructor('text', '', 0)
-    pText.content = descStart.trim()
+    pText.content = descStart
     pInline.children = [pText]
     result.push(pInline)
     
@@ -612,7 +634,7 @@ const convertListItemToDtDd = (tokens, itemStart, itemEnd, parentLevel, opt) =>
   result.push(ddClose)
   
   // Create div_close if descriptionListWithDiv is enabled
-  if (opt && opt.descriptionListWithDiv) {
+  if (opt.descriptionListWithDiv) {
     const divClose = new tokens[firstPara].constructor('div_close', 'div', -1)
     divClose.level = parentLevel + 1
     divClose.block = true