Bug 1951720: Don't exclude 0 width whitespace text nodes from the accessibility tree if they're at the end of a wrapped line. r=morgan

In bug 1951067, I pruned all 0 width whitespace text nodes from the accessibility tree.
However, it turns out that whitespace text nodes at the end of wrapped lines can also be 0 width.
These have semantic importance, since otherwise, words in separate inline nodes can be merged together without a space.
To fix this, explicitly check that the node is before a hard line break, not a soft (wrapped) line break.

Differential Revision: https://phabricator.services.mozilla.com/D242290

accessible/base/NotificationController.cpp

@@ -823,11 +823,10 @@ void NotificationController::WillRefresh(mozilla::TimeStamp aTime) {
     if (textAcc) {
       // Remove the TextLeafAccessible if:
       // 1. The rendered text is empty; or
-      // 2. The text is just a space, but its layout frame has a width of 0,
+      // 2. The text is invisible, semantically irrelevant whitespace before a
-      // so it isn't visible. This can happen if there is whitespace before an
+      // hard line break.
-      // invisible element at the end of a block.
       if (text.mString.IsEmpty() ||
-          (text.mString.EqualsLiteral(" ") && textFrame->GetRect().IsEmpty())) {
+          nsCoreUtils::IsTrimmedWhitespaceBeforeHardLineBreak(textFrame)) {
 #ifdef A11Y_LOG
         if (logging::IsEnabled(logging::eTree | logging::eText)) {
           logging::MsgBegin("TREE", "text node lost its content; doc: %p",

accessible/base/nsAccessibilityService.cpp

@@ -1306,7 +1306,7 @@ LocalAccessible* nsAccessibilityService::CreateAccessible(
     // Ignore not rendered text nodes and whitespace text nodes between table
     // cells.
     if (text.mString.IsEmpty() ||
-        (text.mString.EqualsLiteral(" ") && frame->GetRect().IsEmpty()) ||
+        nsCoreUtils::IsTrimmedWhitespaceBeforeHardLineBreak(frame) ||
         (aContext->IsTableRow() &&
          nsCoreUtils::IsWhitespaceString(text.mString))) {
       if (aIsSubtreeHidden) *aIsSubtreeHidden = true;

accessible/base/nsCoreUtils.cpp

@@ -661,3 +661,19 @@ Element* nsCoreUtils::GetAriaActiveDescendantElement(Element* aElement) {
 
   return nullptr;
 }
+
+bool nsCoreUtils::IsTrimmedWhitespaceBeforeHardLineBreak(nsIFrame* aFrame) {
+  if (!aFrame->GetRect().IsEmpty() ||
+      !aFrame->HasAnyStateBits(TEXT_END_OF_LINE)) {
+    return false;
+  }
+  // Normally, accessibility calls nsIFrame::GetRenderedText with
+  // TrailingWhitespace::NoTrim. Using TrailingWhitespace::Trim instead trims 0
+  // width whitespace before a hard line break, resulting in an empty string if
+  // that is all the frame contains. Note that TrailingWhitespace::Trim does
+  // *not* trim whitespace before a soft line break (wrapped line).
+  nsIFrame::RenderedText text = aFrame->GetRenderedText(
+      0, UINT32_MAX, nsIFrame::TextOffsetType::OffsetsInContentText,
+      nsIFrame::TrailingWhitespace::Trim);
+  return text.mString.IsEmpty();
+}

accessible/base/nsCoreUtils.h

@@ -336,6 +336,17 @@ class nsCoreUtils {
                                                        nsINode* aStartAncestor);
 
   static Element* GetAriaActiveDescendantElement(Element* aElement);
+
+  /**
+   * Return true if the given text frame is 0 width whitespace before a hard
+   * line break.  This is not visible and is semantically irrelevant. This can
+   * happen if there is whitespace before an invisible element at the end of a
+   * block. For example:
+   * <div><span>a</span> <span hidden>b</span></div>
+   * This results in a text node for "a" and a text node for " ". This function
+   * will return true for the latter node.
+   */
+  static bool IsTrimmedWhitespaceBeforeHardLineBreak(nsIFrame* aFrame);
 };
 
 #endif

accessible/tests/browser/e10s/browser_treeupdate_whitespace.js

@@ -69,24 +69,29 @@ addAccessibleTask(
 );
 
 /**
- * Test whitespace before a hidden element at the end of a block.
+ * Test whitespace before hard and soft line breaks.
  */
 addAccessibleTask(
-  `<div id="container"><span>a</span> <span id="b" hidden>b</span></div>`,
+  `
-  async function testBeforeHiddenElementAtEnd(browser, docAcc) {
+<div id="hardContainer"><span>a</span> <span id="b" hidden>b</span></div>
-    const container = findAccessibleChildByID(docAcc, "container");
+<div id="softContainer" style="width: 1ch; font-family: monospace;">
-    testAccessibleTree(container, {
+  <span>c</span> <span>d</span>
+</div>
+  `,
+  async function testBeforeLineBreaks(browser, docAcc) {
+    const hardContainer = findAccessibleChildByID(docAcc, "hardContainer");
+    testAccessibleTree(hardContainer, {
       role: ROLE_SECTION,
       children: [{ role: ROLE_TEXT_LEAF, name: "a" }],
     });
 
     info("Showing b");
-    let reordered = waitForEvent(EVENT_REORDER, container);
+    let reordered = waitForEvent(EVENT_REORDER, hardContainer);
     await invokeContentTask(browser, [], () => {
       content.document.getElementById("b").hidden = false;
     });
     await reordered;
-    testAccessibleTree(container, {
+    testAccessibleTree(hardContainer, {
       role: ROLE_SECTION,
       children: [
         { role: ROLE_TEXT_LEAF, name: "a" },
@@ -96,15 +101,25 @@ addAccessibleTask(
     });
 
     info("Hiding b");
-    reordered = waitForEvent(EVENT_REORDER, container);
+    reordered = waitForEvent(EVENT_REORDER, hardContainer);
     await invokeContentTask(browser, [], () => {
       content.document.getElementById("b").hidden = true;
     });
     await reordered;
-    testAccessibleTree(container, {
+    testAccessibleTree(hardContainer, {
       role: ROLE_SECTION,
       children: [{ role: ROLE_TEXT_LEAF, name: "a" }],
     });
+
+    const softContainer = findAccessibleChildByID(docAcc, "softContainer");
+    testAccessibleTree(softContainer, {
+      role: ROLE_SECTION,
+      children: [
+        { role: ROLE_TEXT_LEAF, name: "c" },
+        { role: ROLE_TEXT_LEAF, name: " " },
+        { role: ROLE_TEXT_LEAF, name: "d" },
+      ],
+    });
   },
   { chrome: true, topLevel: true }
 );