Compare commits

..

3 Commits

Author SHA1 Message Date
Gordon Lam (SH)
218079e596 test(AdvancedPaste): improve reflection helper robustness and rename test method
- Replace null-forgiving operators with explicit null checks and descriptive exception messages
- Rename CleanHtml_NestedGoogleSheetsTable_PreservesNestedContent to CleanHtml_ComplexGoogleSheetsTable_PreservesAllContent

Addresses PR review feedback for better test failure diagnostics and more accurate test naming.
2026-02-03 07:16:49 -08:00
Gordon Lam (SH)
fb19bb9966 fix(paste): address PR review comments
- Add null check for ParentNode in RemoveUnwantedElements to prevent potential NullReferenceException when wrapper is document root
- Remove unused HtmlAgilityPack import from test file
2026-01-31 09:35:25 -08:00
Gordon Lam (SH)
33142a556c fix(paste): handle Google Sheets HTML wrapper and remove style/colgroup elements 2026-01-31 09:29:59 -08:00
5 changed files with 205 additions and 125 deletions

View File

@@ -0,0 +1,171 @@
// Copyright (c) Microsoft Corporation
// The Microsoft Corporation licenses this file to you under the MIT license.
// See the LICENSE file in the project root for more information.
using System.Reflection;
using AdvancedPaste.Helpers;
using Microsoft.VisualStudio.TestTools.UnitTesting;
namespace AdvancedPaste.UnitTests.HelpersTests;
[TestClass]
public sealed class MarkdownHelperTests
{
/// <summary>
/// Helper method to invoke the private CleanHtml method for testing.
/// </summary>
private static string InvokeCleanHtml(string html)
{
var methodInfo = typeof(MarkdownHelper).GetMethod("CleanHtml", BindingFlags.NonPublic | BindingFlags.Static)
?? throw new InvalidOperationException($"Method 'CleanHtml' not found on {nameof(MarkdownHelper)}. The method may have been renamed or removed.");
return (string?)methodInfo.Invoke(null, [html])
?? throw new InvalidOperationException("CleanHtml returned null unexpectedly.");
}
/// <summary>
/// Helper method to invoke the private ConvertHtmlToMarkdown method for testing.
/// </summary>
private static string InvokeConvertHtmlToMarkdown(string html)
{
var methodInfo = typeof(MarkdownHelper).GetMethod("ConvertHtmlToMarkdown", BindingFlags.NonPublic | BindingFlags.Static)
?? throw new InvalidOperationException($"Method 'ConvertHtmlToMarkdown' not found on {nameof(MarkdownHelper)}. The method may have been renamed or removed.");
return (string?)methodInfo.Invoke(null, [html])
?? throw new InvalidOperationException("ConvertHtmlToMarkdown returned null unexpectedly.");
}
[TestMethod]
public void CleanHtml_GoogleSheetsWrapper_RemovesWrapperPreservesTable()
{
// Arrange - Google Sheets HTML with wrapper element
const string googleSheetsHtml = @"<google-sheets-html-origin>
<style type=""text/css""><!--td {border: 1px solid #cccccc;}--></style>
<table xmlns=""http://www.w3.org/1999/xhtml"" cellspacing=""0"" cellpadding=""0"" dir=""ltr"" border=""1"" data-sheets-root=""1"">
<colgroup><col width=""100""><col width=""100""></colgroup>
<tbody><tr><td>A</td><td>B</td></tr><tr><td>1</td><td>2</td></tr></tbody>
</table>
</google-sheets-html-origin>";
// Act
string cleanedHtml = InvokeCleanHtml(googleSheetsHtml);
// Assert - wrapper and style should be removed, table preserved
Assert.IsFalse(cleanedHtml.Contains("google-sheets-html-origin"), "Google Sheets wrapper should be removed");
Assert.IsFalse(cleanedHtml.Contains("<style"), "Style element should be removed");
Assert.IsFalse(cleanedHtml.Contains("<colgroup"), "Colgroup element should be removed");
Assert.IsTrue(cleanedHtml.Contains("<table"), "Table element should be preserved");
Assert.IsTrue(cleanedHtml.Contains("<td>A</td>"), "Table content should be preserved");
}
[TestMethod]
public void CleanHtml_GoogleSheetsHtml_ConvertsToMarkdownTable()
{
// Arrange - Google Sheets HTML with wrapper element
const string googleSheetsHtml = @"<google-sheets-html-origin>
<style type=""text/css""><!--td {border: 1px solid #cccccc;}--></style>
<table xmlns=""http://www.w3.org/1999/xhtml"" cellspacing=""0"" cellpadding=""0"" dir=""ltr"" border=""1"" data-sheets-root=""1"">
<colgroup><col width=""100""><col width=""100""></colgroup>
<tbody><tr><td>A</td><td>B</td></tr><tr><td>1</td><td>2</td></tr></tbody>
</table>
</google-sheets-html-origin>";
// Act
string cleanedHtml = InvokeCleanHtml(googleSheetsHtml);
string markdown = InvokeConvertHtmlToMarkdown(cleanedHtml);
// Assert - should produce valid Markdown table
Assert.IsTrue(markdown.Contains("|"), "Markdown should contain table pipes");
Assert.IsTrue(markdown.Contains("A") && markdown.Contains("B"), "Markdown should contain table content");
Assert.IsTrue(markdown.Contains("1") && markdown.Contains("2"), "Markdown should contain table data");
Assert.IsFalse(markdown.Contains("<google-sheets-html-origin>"), "Markdown should not contain HTML wrapper");
}
[TestMethod]
public void CleanHtml_ExcelHtml_ConvertsToMarkdownTable()
{
// Arrange - Typical Excel HTML (for regression testing)
const string excelHtml = @"<table border=""0"" cellpadding=""0"" cellspacing=""0"" width=""192"">
<tbody><tr height=""20""><td height=""20"" width=""64"">Name</td><td width=""64"">Value</td><td width=""64"">Status</td></tr>
<tr height=""20""><td height=""20"">Item1</td><td>100</td><td>Active</td></tr>
<tr height=""20""><td height=""20"">Item2</td><td>200</td><td>Inactive</td></tr>
</tbody></table>";
// Act
string cleanedHtml = InvokeCleanHtml(excelHtml);
string markdown = InvokeConvertHtmlToMarkdown(cleanedHtml);
// Assert - Excel HTML should still convert correctly
Assert.IsTrue(markdown.Contains("|"), "Markdown should contain table pipes");
Assert.IsTrue(markdown.Contains("Name"), "Markdown should contain header content");
Assert.IsTrue(markdown.Contains("Item1") && markdown.Contains("Item2"), "Markdown should contain row data");
}
[TestMethod]
public void CleanHtml_StyleElement_IsRemoved()
{
// Arrange
const string htmlWithStyle = @"<html><head><style>body { color: red; }</style></head><body><p>Text</p></body></html>";
// Act
string cleanedHtml = InvokeCleanHtml(htmlWithStyle);
// Assert
Assert.IsFalse(cleanedHtml.Contains("<style"), "Style element should be removed");
Assert.IsTrue(cleanedHtml.Contains("<p>Text</p>") || cleanedHtml.Contains("Text"), "Content should be preserved");
}
[TestMethod]
public void CleanHtml_ColgroupElement_IsRemoved()
{
// Arrange
const string htmlWithColgroup = @"<table><colgroup><col width=""100""></colgroup><tbody><tr><td>Data</td></tr></tbody></table>";
// Act
string cleanedHtml = InvokeCleanHtml(htmlWithColgroup);
// Assert
Assert.IsFalse(cleanedHtml.Contains("<colgroup"), "Colgroup element should be removed");
Assert.IsTrue(cleanedHtml.Contains("<table"), "Table element should be preserved");
Assert.IsTrue(cleanedHtml.Contains("Data"), "Table content should be preserved");
}
[TestMethod]
public void CleanHtml_ScriptElement_IsRemoved()
{
// Arrange
const string htmlWithScript = @"<html><body><script>alert('test');</script><p>Content</p></body></html>";
// Act
string cleanedHtml = InvokeCleanHtml(htmlWithScript);
// Assert
Assert.IsFalse(cleanedHtml.Contains("<script"), "Script element should be removed");
Assert.IsTrue(cleanedHtml.Contains("Content"), "Content should be preserved");
}
[TestMethod]
public void CleanHtml_ComplexGoogleSheetsTable_PreservesAllContent()
{
// Arrange - More complex Google Sheets HTML
const string complexGoogleSheetsHtml = @"<google-sheets-html-origin>
<style type=""text/css"">td {border: 1px solid #ccc;}</style>
<table data-sheets-root=""1"">
<colgroup><col width=""100""><col width=""150""><col width=""100""></colgroup>
<tbody>
<tr><td>Header1</td><td>Header2</td><td>Header3</td></tr>
<tr><td>Row1Col1</td><td>Row1Col2</td><td>Row1Col3</td></tr>
<tr><td>Row2Col1</td><td>Row2Col2</td><td>Row2Col3</td></tr>
</tbody>
</table>
</google-sheets-html-origin>";
// Act
string cleanedHtml = InvokeCleanHtml(complexGoogleSheetsHtml);
string markdown = InvokeConvertHtmlToMarkdown(cleanedHtml);
// Assert
Assert.IsFalse(markdown.Contains("<google-sheets-html-origin>"), "Wrapper should be removed from markdown");
Assert.IsTrue(markdown.Contains("Header1"), "Headers should be preserved");
Assert.IsTrue(markdown.Contains("Row1Col1"), "Row data should be preserved");
Assert.IsTrue(markdown.Contains("Row2Col3"), "All cells should be preserved");
}
}

View File

@@ -125,17 +125,7 @@ namespace AdvancedPaste
public void SetFocus()
{
// Set initial focus based on AI enabled state:
// - If AI is enabled, focus the prompt textbox
// - If AI is disabled, focus the paste options list for keyboard navigation
if (_optionsViewModel.IsCustomAIServiceEnabled)
{
MainPage.CustomFormatTextBox.InputTxtBox.Focus(FocusState.Programmatic);
}
else
{
MainPage.SetInitialFocusToPasteOptions();
}
MainPage.CustomFormatTextBox.InputTxtBox.Focus(FocusState.Programmatic);
}
public void ClearInputText()

View File

@@ -163,13 +163,11 @@
</Grid.ColumnDefinitions>
<controls:ClipboardHistoryItemPreviewControl Height="48" ClipboardItem="{x:Bind ViewModel.CurrentClipboardItem, Mode=OneWay}" />
<Button
x:Name="ClipboardHistoryButton"
x:Uid="ClipboardHistoryButton"
Grid.Column="1"
Margin="0,0,4,0"
VerticalAlignment="Center"
IsEnabled="{x:Bind ViewModel.ClipboardHistoryEnabled, Mode=TwoWay}"
KeyDown="ClipboardHistoryButton_KeyDown"
Style="{StaticResource SubtleButtonStyle}"
Visibility="{x:Bind ViewModel.ShowClipboardHistoryButton, Converter={StaticResource BoolToVisibilityConverter}, Mode=OneWay}">
<ToolTipService.ToolTip>
@@ -183,19 +181,15 @@
Glyph="&#xE81C;" />
<Button.Flyout>
<Flyout
x:Name="ClipboardHistoryFlyout"
Closed="ClipboardHistoryFlyout_Closed"
FlyoutPresenterStyle="{StaticResource PaddingLessFlyoutPresenterStyle}"
Placement="Right"
ShouldConstrainToRootBounds="False">
<ItemsView
x:Name="ClipboardHistoryItemsView"
Width="320"
Margin="8,8,8,0"
IsItemInvokedEnabled="True"
ItemInvoked="ClipboardHistory_ItemInvoked"
ItemsSource="{x:Bind clipboardHistory, Mode=OneWay}"
KeyDown="ClipboardHistoryItemsView_KeyDown"
SelectionMode="None">
<ItemsView.Layout>
<StackLayout Orientation="Vertical" Spacing="8" />
@@ -323,13 +317,10 @@
ItemContainerTransitions="{x:Null}"
ItemTemplateSelector="{StaticResource PasteFormatTemplateSelector}"
ItemsSource="{x:Bind ViewModel.StandardPasteFormats, Mode=OneWay}"
KeyDown="PasteOptionsListView_KeyDown"
ScrollViewer.VerticalScrollBarVisibility="Disabled"
ScrollViewer.VerticalScrollMode="Disabled"
SelectionMode="Single"
SingleSelectionFollowsFocus="True"
TabIndex="1"
XYFocusKeyboardNavigation="Enabled" />
SelectionMode="None"
TabIndex="1" />
<Rectangle
Grid.Row="1"
Height="1"
@@ -346,13 +337,10 @@
ItemContainerTransitions="{x:Null}"
ItemTemplateSelector="{StaticResource PasteFormatTemplateSelector}"
ItemsSource="{x:Bind ViewModel.CustomActionPasteFormats, Mode=OneWay}"
KeyDown="PasteOptionsListView_KeyDown"
ScrollViewer.VerticalScrollBarVisibility="Disabled"
ScrollViewer.VerticalScrollMode="Disabled"
SelectionMode="Single"
SingleSelectionFollowsFocus="True"
TabIndex="2"
XYFocusKeyboardNavigation="Enabled" />
SelectionMode="None"
TabIndex="2" />
</Grid>
</ScrollViewer>
</Grid>

View File

@@ -208,103 +208,5 @@ namespace AdvancedPaste.Pages
Clipboard.SetHistoryItemAsContent(item.Item);
}
}
/// <summary>
/// Sets initial focus to the paste options list when AI is disabled.
/// </summary>
public void SetInitialFocusToPasteOptions()
{
try
{
if (PasteOptionsListView.Items.Count > 0)
{
PasteOptionsListView.SelectedIndex = 0;
PasteOptionsListView.Focus(FocusState.Programmatic);
Logger.LogTrace("Focus set to PasteOptionsListView");
}
}
catch (Exception ex)
{
Logger.LogError("Failed to set focus to paste options", ex);
}
}
/// <summary>
/// Gets the appropriate arrow key for "forward" direction based on RTL settings.
/// </summary>
private VirtualKey GetForwardKey()
{
return FlowDirection == FlowDirection.RightToLeft ? VirtualKey.Left : VirtualKey.Right;
}
/// <summary>
/// Gets the appropriate arrow key for "backward" direction based on RTL settings.
/// </summary>
private VirtualKey GetBackwardKey()
{
return FlowDirection == FlowDirection.RightToLeft ? VirtualKey.Right : VirtualKey.Left;
}
/// <summary>
/// Handles keyboard navigation on the paste options ListViews.
/// Enter key invokes the selected item.
/// </summary>
private async void PasteOptionsListView_KeyDown(object sender, Microsoft.UI.Xaml.Input.KeyRoutedEventArgs e)
{
if (sender is ListView listView && e.Key == VirtualKey.Enter)
{
if (listView.SelectedItem is PasteFormat format)
{
e.Handled = true;
await ViewModel.ExecutePasteFormatAsync(format, PasteActionSource.InAppKeyboardShortcut);
}
}
}
/// <summary>
/// Handles keyboard navigation on the clipboard history button.
/// Right arrow (or Left in RTL) opens the flyout.
/// </summary>
private void ClipboardHistoryButton_KeyDown(object sender, Microsoft.UI.Xaml.Input.KeyRoutedEventArgs e)
{
if (e.Key == GetForwardKey())
{
e.Handled = true;
if (ClipboardHistoryButton.Flyout is Flyout flyout)
{
flyout.ShowAt(ClipboardHistoryButton);
Logger.LogTrace("Clipboard history flyout opened via keyboard");
}
}
}
/// <summary>
/// Handles keyboard navigation within the clipboard history flyout.
/// Escape or Left arrow (Right in RTL) closes the flyout and returns focus to the button.
/// </summary>
private void ClipboardHistoryItemsView_KeyDown(object sender, Microsoft.UI.Xaml.Input.KeyRoutedEventArgs e)
{
if (e.Key == VirtualKey.Escape || e.Key == GetBackwardKey())
{
e.Handled = true;
ClipboardHistoryFlyout.Hide();
}
}
/// <summary>
/// Handles the clipboard history flyout closing event.
/// Returns focus to the clipboard history button.
/// </summary>
private void ClipboardHistoryFlyout_Closed(object sender, object e)
{
try
{
ClipboardHistoryButton.Focus(FocusState.Programmatic);
}
catch (Exception ex)
{
Logger.LogError("Failed to return focus to clipboard history button", ex);
}
}
}
}

View File

@@ -57,6 +57,23 @@ namespace AdvancedPaste.Helpers
{
Logger.LogTrace();
// Unwrap Google Sheets wrapper element (preserve children, remove wrapper)
foreach (var googleSheetsWrapper in node.DescendantsAndSelf("google-sheets-html-origin").ToArray())
{
var parent = googleSheetsWrapper.ParentNode;
if (parent == null)
{
continue;
}
foreach (var child in googleSheetsWrapper.ChildNodes.ToArray())
{
parent.InsertBefore(child, googleSheetsWrapper);
}
googleSheetsWrapper.Remove();
}
// Remove specific elements by tag name, CSS class, or other attributes
// Example: Remove all <script> elements
foreach (var scriptNode in node.DescendantsAndSelf("script").ToArray())
@@ -64,6 +81,18 @@ namespace AdvancedPaste.Helpers
scriptNode.Remove();
}
// Remove style elements (CSS not relevant for Markdown)
foreach (var styleNode in node.DescendantsAndSelf("style").ToArray())
{
styleNode.Remove();
}
// Remove colgroup elements (column width info not needed for Markdown)
foreach (var colgroupNode in node.DescendantsAndSelf("colgroup").ToArray())
{
colgroupNode.Remove();
}
// Ignore specific elements like <sup> elements
foreach (var ignoredElement in node.DescendantsAndSelf("sup").ToArray())
{