Add missing line breaks and tabulators to innerText

.gitignore

@@ -2,4 +2,5 @@
 .nyc_output
 coverage/
 node_modules/
-.idea
+.idea
+linkedom-*.tgz

.zed/settings.json

@@ -0,0 +1,8 @@
+// Folder-specific settings
+//
+// For a full list of overridable settings, and general information on folder-specific settings,
+// see the documentation: https://zed.dev/docs/configuring-zed#settings-files
+{
+    "tab_size": 2,
+    "format_on_save": "off"
+}

cjs/interface/element.js

@@ -4,6 +4,7 @@
 const {
   ATTRIBUTE_NODE,
   BLOCK_ELEMENTS,
+  TABLE_ELEMENTS,
   CDATA_SECTION_NODE,
   COMMENT_NODE,
   ELEMENT_NODE,
@@ -157,22 +158,59 @@ class Element extends ParentNode {
   // </specialGetters>
 
 
-  // <contentRelated>
-  get innerText() {
+  __getInnerText(customGetter) {
     const text = [];
     let {[NEXT]: next, [END]: end} = this;
+
+
+
     while (next !== end) {
+        if (typeof customGetter === "function") {
+          const custom = customGetter(next);
+          if (typeof custom === "string") {
+            text.push(custom)
+            next = next[NEXT];
+            continue;
+          }
+        }
+
+      // Add tabulators between table columns
+      if (
+          TABLE_ELEMENTS.has(next.tagName) &&
+          TABLE_ELEMENTS.has(next.previousElementSibling?.tagName)) {
+          text.push('\t');
+      }
+
       if (next.nodeType === TEXT_NODE) {
         text.push(next.textContent.replace(/\s+/g, ' '));
       } else if(
         text.length && next[NEXT] != end &&
         BLOCK_ELEMENTS.has(next.tagName)
+        && text.at(-1) !== '\n'
       ) {
         text.push('\n');
       }
+
+      // add line breaks on closing elements
+      if (
+        text.at(-1) !== '\n' &&
+        (!next.nextSibling || BLOCK_ELEMENTS.has(next.nextSibling?.tagName)) &&
+          (!next.previousSibling || BLOCK_ELEMENTS.has(next.previousSibling?.tagName)) &&
+          BLOCK_ELEMENTS.has(next.parentElement?.tagName)
+      ) {
+        text.push('\n');
+      }
+
       next = next[NEXT];
+
     }
     return text.join('');
+
+  }
+
+  // <contentRelated>
+  get innerText() {
+    return this.__getInnerText()
   }
 
   /**

cjs/shared/constants.js

@@ -22,9 +22,12 @@ const DOCUMENT_FRAGMENT_NODE = 11;
 exports.DOCUMENT_FRAGMENT_NODE = DOCUMENT_FRAGMENT_NODE;
 
 // Elements
-const BLOCK_ELEMENTS = new Set(['ARTICLE', 'ASIDE', 'BLOCKQUOTE', 'BODY', 'BR', 'BUTTON', 'CANVAS', 'CAPTION', 'COL', 'COLGROUP', 'DD', 'DIV', 'DL', 'DT', 'EMBED', 'FIELDSET', 'FIGCAPTION', 'FIGURE', 'FOOTER', 'FORM', 'H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'LI', 'UL', 'OL', 'P']);
+const BLOCK_ELEMENTS = new Set(['ARTICLE', 'ASIDE', 'BLOCKQUOTE', 'BODY', 'BR', 'BUTTON', 'CANVAS', 'CAPTION', 'COL', 'COLGROUP', 'DD', 'DIV', 'DL', 'DT', 'EMBED', 'FIELDSET', 'FIGCAPTION', 'FIGURE', 'FOOTER', 'FORM', 'H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'LI', 'UL', 'OL', 'P', 'TR', 'PRE', 'HR', 'ADDRESS']);
 exports.BLOCK_ELEMENTS = BLOCK_ELEMENTS;
 
+const TABLE_ELEMENTS = new Set(['TH', 'TD']);
+exports.TABLE_ELEMENTS = TABLE_ELEMENTS;
+
 // TreeWalker
 const SHOW_ALL = -1;
 exports.SHOW_ALL = SHOW_ALL;

esm/interface/element.js

@@ -3,6 +3,7 @@
 import {
   ATTRIBUTE_NODE,
   BLOCK_ELEMENTS,
+  TABLE_ELEMENTS,
   CDATA_SECTION_NODE,
   COMMENT_NODE,
   ELEMENT_NODE,
@@ -159,22 +160,59 @@ export class Element extends ParentNode {
   // </specialGetters>
 
 
-  // <contentRelated>
-  get innerText() {
+  __getInnerText(customGetter) {
     const text = [];
     let {[NEXT]: next, [END]: end} = this;
+
+
+
     while (next !== end) {
+        if (typeof customGetter === "function") {
+          const custom = customGetter(next);
+          if (typeof custom === "string") {
+            text.push(custom)
+            next = next[NEXT];
+            continue;
+          }
+        }
+
+      // Add tabulators between table columns
+      if (
+          TABLE_ELEMENTS.has(next.tagName) &&
+          TABLE_ELEMENTS.has(next.previousElementSibling?.tagName)) {
+          text.push('\t');
+      }
+
       if (next.nodeType === TEXT_NODE) {
         text.push(next.textContent.replace(/\s+/g, ' '));
       } else if(
         text.length && next[NEXT] != end &&
         BLOCK_ELEMENTS.has(next.tagName)
+        && text.at(-1) !== '\n'
       ) {
         text.push('\n');
       }
+
+      // add line breaks on closing elements
+      if (
+        text.at(-1) !== '\n' &&
+        (!next.nextSibling || BLOCK_ELEMENTS.has(next.nextSibling?.tagName)) &&
+          (!next.previousSibling || BLOCK_ELEMENTS.has(next.previousSibling?.tagName)) &&
+          BLOCK_ELEMENTS.has(next.parentElement?.tagName)
+      ) {
+        text.push('\n');
+      }
+
       next = next[NEXT];
+
     }
     return text.join('');
+
+  }
+
+  // <contentRelated>
+  get innerText() {
+    return this.__getInnerText()
   }
 
   /**

esm/shared/constants.js

@@ -12,7 +12,9 @@ export const DOCUMENT_TYPE_NODE = 10;
 export const DOCUMENT_FRAGMENT_NODE = 11;
 
 // Elements
-export const BLOCK_ELEMENTS = new Set(['ARTICLE', 'ASIDE', 'BLOCKQUOTE', 'BODY', 'BR', 'BUTTON', 'CANVAS', 'CAPTION', 'COL', 'COLGROUP', 'DD', 'DIV', 'DL', 'DT', 'EMBED', 'FIELDSET', 'FIGCAPTION', 'FIGURE', 'FOOTER', 'FORM', 'H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'LI', 'UL', 'OL', 'P']);
+export const BLOCK_ELEMENTS = new Set(['ARTICLE', 'ASIDE', 'BLOCKQUOTE', 'BODY', 'BR', 'BUTTON', 'CANVAS', 'CAPTION', 'COL', 'COLGROUP', 'DD', 'DIV', 'DL', 'DT', 'EMBED', 'FIELDSET', 'FIGCAPTION', 'FIGURE', 'FOOTER', 'FORM', 'H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'LI', 'UL', 'OL', 'P', 'TR', 'PRE', 'HR', 'ADDRESS']);
+
+export const TABLE_ELEMENTS = new Set(['TH', 'TD']);
 
 // TreeWalker
 export const SHOW_ALL = -1;

test/html/element.js

@@ -173,5 +173,56 @@ node.innerHTML = '<video src="" controls>';
 assert(node.innerHTML, '<video src="" controls></video>');
 
 node.innerHTML = '<div>The <strong>quick</strong> brown fox</div><div>Jumped over<br>The lazy\ndog</div>';
-assert(node.innerText, 'The quick brown fox\nJumped over\nThe lazy dog', 'innerText newlines');
+assert(v(node.innerText), v('\nThe quick brown fox\nJumped over\nThe lazy dog\n'), 'innerText newlines');
 assert(node.textContent, 'The quick brown foxJumped overThe lazy\ndog', 'textContent no newlines');
+
+function v(str) {
+  return str
+    .replace(/ /g, '·')   // Replace space with middle dot
+    .replace(/\t/g, '→')  // Replace tab with right arrow
+    .replace(/\n/g, '↵'); // Replace newline with down-left arrow
+}
+
+node.innerHTML = '<table><tr><th>Header1</th><th>Header2</th></tr><tr><td>Row1,Col1</td><td>Row1,Col2</td></tr><tr><td>Row2,Col1</td><td>Row2,Col2</td></tr></table>';
+assert(v(node.innerText), v('\nHeader1\tHeader2\nRow1,Col1\tRow1,Col2\nRow2,Col1\tRow2,Col2'));
+
+node.innerHTML = '<div>prefix<div>div</div>suffix</div>';
+assert(v(node.innerText), v('\nprefix\ndiv\nsuffix\n'));
+
+node.innerHTML = '<div>prefix<div>div1<div>div2</div></div>suffix</div>';
+assert(v(node.innerText), v('\nprefix\ndiv1\ndiv2\nsuffix\n'));
+
+node.innerHTML = '<div>prefix<div>foo<div>nested</div>bar</div>suffix</div>';
+assert(v(node.innerText), v('\nprefix\nfoo\nnested\nbar\nsuffix\n'));
+
+node.innerHTML = `
+
+  <div>
+    content
+    <div class="crap">
+      crap
+      <span>
+      morecrap
+      </span>
+
+    </div>
+    <span>
+    more content
+    </span>
+
+  </div>
+
+`
+console.log(node.__getInnerText(textNode =>{
+  if (textNode.nodeType !== 3) {
+    return;
+  }
+
+  const el = textNode.parentElement;
+
+  if (el.closest(".crap")){
+    return ""
+  }
+
+
+}))

types/esm/interface/element.d.ts

@@ -32,6 +32,7 @@ export class Element extends ParentNode implements globalThis.Element {
     get tabIndex(): number;
     set slot(value: any);
     get slot(): any;
+    __getInnerText(customGetter: any): string;
     get innerText(): string;
     set textContent(text: string);
     /**

types/esm/interface/image.d.ts

@@ -121,6 +121,7 @@ export function ImageClass(ownerDocument: any): {
         readonly style: any;
         tabIndex: number;
         slot: any;
+        __getInnerText(customGetter: any): string;
         readonly innerText: string;
         textContent: string;
         innerHTML: string;

types/esm/shared/constants.d.ts

@@ -8,6 +8,7 @@ export const DOCUMENT_NODE: 9;
 export const DOCUMENT_TYPE_NODE: 10;
 export const DOCUMENT_FRAGMENT_NODE: 11;
 export const BLOCK_ELEMENTS: Set<string>;
+export const TABLE_ELEMENTS: Set<string>;
 export const SHOW_ALL: -1;
 export const SHOW_ELEMENT: 1;
 export const SHOW_TEXT: 4;

worker.js

@@ -3839,7 +3839,9 @@ const DOCUMENT_TYPE_NODE = 10;
 const DOCUMENT_FRAGMENT_NODE = 11;
 
 // Elements
-const BLOCK_ELEMENTS = new Set(['ARTICLE', 'ASIDE', 'BLOCKQUOTE', 'BODY', 'BR', 'BUTTON', 'CANVAS', 'CAPTION', 'COL', 'COLGROUP', 'DD', 'DIV', 'DL', 'DT', 'EMBED', 'FIELDSET', 'FIGCAPTION', 'FIGURE', 'FOOTER', 'FORM', 'H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'LI', 'UL', 'OL', 'P']);
+const BLOCK_ELEMENTS = new Set(['ARTICLE', 'ASIDE', 'BLOCKQUOTE', 'BODY', 'BR', 'BUTTON', 'CANVAS', 'CAPTION', 'COL', 'COLGROUP', 'DD', 'DIV', 'DL', 'DT', 'EMBED', 'FIELDSET', 'FIGCAPTION', 'FIGURE', 'FOOTER', 'FORM', 'H1', 'H2', 'H3', 'H4', 'H5', 'H6', 'LI', 'UL', 'OL', 'P', 'TR', 'PRE', 'HR', 'ADDRESS']);
+
+const TABLE_ELEMENTS = new Set(['TH', 'TD']);
 
 // TreeWalker
 const SHOW_ALL = -1;
@@ -7739,22 +7741,59 @@ let Element$1 = class Element extends ParentNode {
   // </specialGetters>
 
 
-  // <contentRelated>
-  get innerText() {
+  __getInnerText(customGetter) {
     const text = [];
     let {[NEXT]: next, [END]: end} = this;
+
+
+
     while (next !== end) {
+        if (typeof customGetter === "function") {
+          const custom = customGetter(next);
+          if (typeof custom === "string") {
+            text.push(custom);
+            next = next[NEXT];
+            continue;
+          }
+        }
+
+      // Add tabulators between table columns
+      if (
+          TABLE_ELEMENTS.has(next.tagName) &&
+          TABLE_ELEMENTS.has(next.previousElementSibling?.tagName)) {
+          text.push('\t');
+      }
+
       if (next.nodeType === TEXT_NODE) {
         text.push(next.textContent.replace(/\s+/g, ' '));
       } else if(
         text.length && next[NEXT] != end &&
         BLOCK_ELEMENTS.has(next.tagName)
+        && text.at(-1) !== '\n'
       ) {
         text.push('\n');
       }
+
+      // add line breaks on closing elements
+      if (
+        text.at(-1) !== '\n' &&
+        (!next.nextSibling || BLOCK_ELEMENTS.has(next.nextSibling?.tagName)) &&
+          (!next.previousSibling || BLOCK_ELEMENTS.has(next.previousSibling?.tagName)) &&
+          BLOCK_ELEMENTS.has(next.parentElement?.tagName)
+      ) {
+        text.push('\n');
+      }
+
       next = next[NEXT];
+
     }
     return text.join('');
+
+  }
+
+  // <contentRelated>
+  get innerText() {
+    return this.__getInnerText()
   }
 
   /**