@adobe/helix-docx2md 1.4.20 → 1.5.0

package/CHANGELOG.md

@@ -1,3 +1,17 @@
+# [1.5.0](https://github.com/adobe/helix-docx2md/compare/v1.4.21...v1.5.0) (2023-11-03)
+
+
+### Features
+
+* Handle emdash character for section breaks ([c2c8495](https://github.com/adobe/helix-docx2md/commit/c2c84957bf7a8f47acf70035d08318f84ea12de8)), closes [#343](https://github.com/adobe/helix-docx2md/issues/343)
+
+## [1.4.21](https://github.com/adobe/helix-docx2md/compare/v1.4.20...v1.4.21) (2023-10-28)
+
+
+### Bug Fixes
+
+* **deps:** update dependency unified to v11.0.4 ([#337](https://github.com/adobe/helix-docx2md/issues/337)) ([bb749d8](https://github.com/adobe/helix-docx2md/commit/bb749d8e1b435d0e5a8a707ab8e00cdbcd5fdf79))
+
 ## [1.4.20](https://github.com/adobe/helix-docx2md/compare/v1.4.19...v1.4.20) (2023-10-10)
 
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@adobe/helix-docx2md",
-  "version": "1.4.20",
+  "version": "1.5.0",
   "description": "Helix library that converts word documents to markdown",
   "type": "module",
   "main": "./src/index.js",
@@ -37,7 +37,7 @@
     "@adobe/helix-shared-process-queue": "3.0.0",
     "@adobe/mammoth": "1.5.1-bleeding.2",
     "@adobe/mdast-util-gridtables": "3.0.2",
-    "@adobe/remark-gridtables": "2.0.1",
+    "@adobe/remark-gridtables": "2.0.2",
     "dirname-filename-esm": "1.1.1",
     "github-slugger": "2.0.0",
     "mdast-util-to-markdown": "2.1.0",
@@ -45,26 +45,26 @@
     "remark-gfm": "4.0.0",
     "remark-parse": "11.0.0",
     "remark-stringify": "11.0.0",
-    "unified": "11.0.3",
+    "unified": "11.0.4",
     "unist-util-find": "3.0.0",
     "unist-util-visit": "5.0.0",
     "yauzl": "2.10.0"
   },
   "devDependencies": {
-    "@adobe/eslint-config-helix": "2.0.3",
-    "@adobe/helix-admin-support": "2.2.14",
-    "@adobe/helix-mediahandler": "2.2.18",
-    "@adobe/helix-onedrive-support": "10.8.5",
-    "@adobe/helix-shared-tokencache": "1.3.11",
+    "@adobe/eslint-config-helix": "2.0.4",
+    "@adobe/helix-admin-support": "2.2.20",
+    "@adobe/helix-mediahandler": "2.3.0",
+    "@adobe/helix-onedrive-support": "11.1.0",
+    "@adobe/helix-shared-tokencache": "1.3.14",
     "@semantic-release/changelog": "6.0.3",
     "@semantic-release/exec": "6.0.3",
     "@semantic-release/git": "10.0.1",
     "c8": "8.0.1",
     "dotenv": "16.3.1",
-    "eslint": "8.51.0",
+    "eslint": "8.52.0",
     "husky": "8.0.3",
-    "junit-report-builder": "3.0.1",
-    "lint-staged": "14.0.1",
+    "junit-report-builder": "3.1.0",
+    "lint-staged": "15.0.2",
     "mocha": "10.2.0",
     "mocha-multi-reporters": "1.5.1",
     "semantic-release": "22.0.5",

package/src/dast2mdast/handlers/paragraph.js

@@ -39,6 +39,11 @@ const HEADING_MAP = {
   'heading 8': 8,
 };
 
+function isHorizontalLine(nodes) {
+  const value = toString(nodes).trim();
+  return (value === '---' || value === '\u2014');
+}
+
 function getHeadingDepth(node) {
   const { styleId, styleName } = node;
   if (styleId in HEADING_MAP) {
@@ -265,8 +270,8 @@ export default function paragraph(h, node, parent, siblings) {
   }
 
   if (depth) {
-    // check if no --- in heading
-    if (toString(nodes).trim() === '---') {
+    // check if no horizontal line in heading
+    if (isHorizontalLine(nodes)) {
       return h('thematicBreak');
     }
     const heading = h('heading', { depth }, nodes);
@@ -310,10 +315,10 @@ export default function paragraph(h, node, parent, siblings) {
   // check for thematicBreaks and frontmatter. they need to be block elements
   const ret = [];
   let prevBreak;
-  let idx = nodes.findIndex((n) => toString(n).trim() === '---');
+  let idx = nodes.findIndex(isHorizontalLine);
   while (idx >= 0) {
     const brk = h('thematicBreak');
-    // remove the nodes up until the '---'
+    // remove the nodes up until the horizontal line
     const removed = nodes.splice(0, idx);
     // if last element is a break, ignore it.
     if (removed.length && removed[removed.length - 1].type === 'break') {
@@ -321,7 +326,7 @@ export default function paragraph(h, node, parent, siblings) {
       removed.pop();
     }
 
-    // remove '---'
+    // remove the horizontal line
     nodes.splice(0, 1);
 
     // also check if element following is a break
@@ -346,7 +351,7 @@ export default function paragraph(h, node, parent, siblings) {
       ret.push(brk);
       prevBreak = brk;
     }
-    idx = nodes.findIndex((n) => n.type === 'text' && n.value === '---');
+    idx = nodes.findIndex(isHorizontalLine);
   }
 
   // handle remaining nodes