From f6e8f3a2a319468fd08572d767a3cf8d953f0be2 Mon Sep 17 00:00:00 2001
From: Alf Eaton <alf.eaton@overleaf.com>
Date: Thu, 14 Sep 2023 13:20:41 +0100
Subject: [PATCH] Merge pull request #14830 from overleaf/ae-office-paste

[visual] Improve handling of content pasted from desktop Office applications

GitOrigin-RevId: 10ce77bae008ad078afa22c8fac2cb48af3d9cb4
---
 .../extensions/visual/paste-html.ts           | 47 ++++++++++++++---
 ...demirror-editor-visual-paste-html.spec.tsx | 50 +++++++++++++++++++
 2 files changed, 91 insertions(+), 6 deletions(-)

diff --git a/services/web/frontend/js/features/source-editor/extensions/visual/paste-html.ts b/services/web/frontend/js/features/source-editor/extensions/visual/paste-html.ts
index e88a583290..ac75d867e4 100644
--- a/services/web/frontend/js/features/source-editor/extensions/visual/paste-html.ts
+++ b/services/web/frontend/js/features/source-editor/extensions/visual/paste-html.ts
@@ -135,8 +135,13 @@ const htmlToLaTeX = (documentElement: HTMLElement) => {
     return ''
   }
 
-  // normalise multiple newlines
-  return text.replaceAll(/\n{2,}/g, '\n\n')
+  return (
+    text
+      // remove zero-width spaces (e.g. those added by Powerpoint)
+      .replaceAll('​', '')
+      // normalise multiple newlines
+      .replaceAll(/\n{2,}/g, '\n\n')
+  )
 }
 
 const processWhitespace = (documentElement: HTMLElement) => {
@@ -274,8 +279,34 @@ const processLists = (element: HTMLElement) => {
   }
 }
 
+const removeNonContentTextNodes = (table: HTMLTableElement) => {
+  // remove text nodes that are direct children of non-content table elements
+  const containers = table.querySelectorAll('thead,tbody,tr')
+  for (const element of [table, ...containers]) {
+    for (const childNode of element.childNodes) {
+      if (childNode.nodeType === Node.TEXT_NODE) {
+        element.removeChild(childNode)
+      }
+    }
+  }
+
+  // remove whitespace-only text nodes at the start or end of table cells
+  for (const element of table.querySelectorAll('th,td')) {
+    for (const childNode of [element.firstChild, element.lastChild]) {
+      if (
+        childNode?.nodeType === Node.TEXT_NODE &&
+        childNode.textContent?.trim() === ''
+      ) {
+        element.removeChild(childNode)
+      }
+    }
+  }
+}
+
 const processTables = (element: HTMLElement) => {
   for (const table of element.querySelectorAll('table')) {
+    removeNonContentTextNodes(table)
+
     // create a wrapper element for the table and the caption
     const container = document.createElement('div')
     container.className = 'ol-table-wrap'
@@ -670,11 +701,13 @@ const selectors = [
     selector: 'tr > td, tr > th',
     start: (element: HTMLTableCellElement) => {
       let output = ''
-      if (element.getAttribute('colspan')) {
+      const colspan = element.getAttribute('colspan')
+      if (colspan && Number(colspan) > 1) {
         output += startMulticolumn(element)
       }
       // NOTE: multirow is nested inside multicolumn
-      if (element.getAttribute('rowspan')) {
+      const rowspan = element.getAttribute('rowspan')
+      if (rowspan && Number(rowspan) > 1) {
         output += startMultirow(element)
       }
       return output
@@ -682,10 +715,12 @@ const selectors = [
     end: element => {
       let output = ''
       // NOTE: multirow is nested inside multicolumn
-      if (element.getAttribute('rowspan')) {
+      const rowspan = element.getAttribute('rowspan')
+      if (rowspan && Number(rowspan) > 1) {
         output += '}'
       }
-      if (element.getAttribute('colspan')) {
+      const colspan = element.getAttribute('colspan')
+      if (colspan && Number(colspan) > 1) {
         output += '}'
       }
       const row = element.parentElement as HTMLTableRowElement
diff --git a/services/web/test/frontend/features/source-editor/components/codemirror-editor-visual-paste-html.spec.tsx b/services/web/test/frontend/features/source-editor/components/codemirror-editor-visual-paste-html.spec.tsx
index 540b59ee4b..bf9c32afe3 100644
--- a/services/web/test/frontend/features/source-editor/components/codemirror-editor-visual-paste-html.spec.tsx
+++ b/services/web/test/frontend/features/source-editor/components/codemirror-editor-visual-paste-html.spec.tsx
@@ -237,6 +237,26 @@ describe('<CodeMirrorEditor/> paste HTML in Visual mode', function () {
     )
   })
 
+  it('ignores rowspan="1" and colspan="1"', function () {
+    mountEditor()
+
+    const data = [
+      `<table><tbody>`,
+      `<tr><td colspan="1" rowspan="1">test</td><td>test</td></tr>`,
+      `<tr><td>test</td><td>test</td><td>test</td></tr>`,
+      `</tbody></table>`,
+    ].join('')
+
+    const clipboardData = new DataTransfer()
+    clipboardData.setData('text/html', data)
+    cy.get('@content').trigger('paste', { clipboardData })
+
+    cy.get('@content').should(
+      'have.text',
+      '\\begin{tabular}{l l l}test & test ↩test & test & test ↩\\end{tabular}'
+    )
+  })
+
   it('handles a pasted table with adjacent borders and merged cells', function () {
     mountEditor()
 
@@ -543,6 +563,18 @@ describe('<CodeMirrorEditor/> paste HTML in Visual mode', function () {
     cy.get('@content').should('have.text', 'foo bar')
   })
 
+  it('removes all zero-width spaces', function () {
+    mountEditor()
+
+    const data = 'foo\u200bbar'
+
+    const clipboardData = new DataTransfer()
+    clipboardData.setData('text/html', data)
+    cy.get('@content').trigger('paste', { clipboardData })
+
+    cy.get('@content').should('have.text', 'foobar')
+  })
+
   it('ignores HTML pasted from VS Code', function () {
     mountEditor()
 
@@ -588,4 +620,22 @@ describe('<CodeMirrorEditor/> paste HTML in Visual mode', function () {
     cy.get('.ol-cm-character').should('have.length', 2)
     cy.get('.ol-cm-command-verb').should('have.length', 1)
   })
+
+  it('tidies whitespace in pasted tables', function () {
+    mountEditor()
+
+    const data = `<table>
+ <tr>
+  <td>
+  <p><b>test</b></p>
+  </td>
+</tr>
+</table>`
+
+    const clipboardData = new DataTransfer()
+    clipboardData.setData('text/html', data)
+    cy.get('@content').trigger('paste', { clipboardData })
+
+    cy.get('.cm-line').should('have.length', 8)
+  })
 })