LibWeb: Make Node.normalize() ignore CDATASection nodes

We hadn't modeled the "exclusive text node" concept correctly.
This commit is contained in:
Andreas Kling 2024-11-20 11:31:59 +01:00 committed by Andreas Kling
parent 6fc06f45c2
commit ab0dc83d28
Notes: github-actions[bot] 2024-11-20 15:12:01 +00:00
4 changed files with 101 additions and 9 deletions

View file

@ -235,10 +235,7 @@ WebIDL::ExceptionOr<void> Node::normalize()
Vector<Text*> nodes; Vector<Text*> nodes;
auto* current_node = node.previous_sibling(); auto* current_node = node.previous_sibling();
while (current_node) { while (current_node && current_node->is_exclusive_text()) {
if (!current_node->is_text())
break;
nodes.append(static_cast<Text*>(current_node)); nodes.append(static_cast<Text*>(current_node));
current_node = current_node->previous_sibling(); current_node = current_node->previous_sibling();
} }
@ -247,10 +244,7 @@ WebIDL::ExceptionOr<void> Node::normalize()
nodes.reverse(); nodes.reverse();
current_node = node.next_sibling(); current_node = node.next_sibling();
while (current_node) { while (current_node && current_node->is_exclusive_text()) {
if (!current_node->is_text())
break;
nodes.append(static_cast<Text*>(current_node)); nodes.append(static_cast<Text*>(current_node));
current_node = current_node->next_sibling(); current_node = current_node->next_sibling();
} }
@ -291,7 +285,7 @@ WebIDL::ExceptionOr<void> Node::normalize()
auto* current_node = node.next_sibling(); auto* current_node = node.next_sibling();
// 6. While currentNode is an exclusive Text node: // 6. While currentNode is an exclusive Text node:
while (current_node && is<Text>(*current_node)) { while (current_node && current_node->is_exclusive_text()) {
// 1. For each live range whose start node is currentNode, add length to its start offset and set its start node to node. // 1. For each live range whose start node is currentNode, add length to its start offset and set its start node to node.
for (auto& range : Range::live_ranges()) { for (auto& range : Range::live_ranges()) {
if (range->start_container() == current_node) if (range->start_container() == current_node)

View file

@ -111,6 +111,7 @@ public:
NodeType type() const { return m_type; } NodeType type() const { return m_type; }
bool is_element() const { return type() == NodeType::ELEMENT_NODE; } bool is_element() const { return type() == NodeType::ELEMENT_NODE; }
bool is_text() const { return type() == NodeType::TEXT_NODE || type() == NodeType::CDATA_SECTION_NODE; } bool is_text() const { return type() == NodeType::TEXT_NODE || type() == NodeType::CDATA_SECTION_NODE; }
bool is_exclusive_text() const { return type() == NodeType::TEXT_NODE; }
bool is_document() const { return type() == NodeType::DOCUMENT_NODE; } bool is_document() const { return type() == NodeType::DOCUMENT_NODE; }
bool is_document_type() const { return type() == NodeType::DOCUMENT_TYPE_NODE; } bool is_document_type() const { return type() == NodeType::DOCUMENT_TYPE_NODE; }
bool is_comment() const { return type() == NodeType::COMMENT_NODE; } bool is_comment() const { return type() == NodeType::COMMENT_NODE; }

View file

@ -0,0 +1,14 @@
Summary
Harness status: OK
Rerun
Found 4 tests
4 Pass
Details
Result Test Name MessagePass Node.normalize()
Pass Empty text nodes separated by a non-empty text node
Pass Empty text nodes
Pass Non-text nodes with empty textContent values.

View file

@ -0,0 +1,83 @@
<!DOCTYPE html>
<title>Node.normalize()</title>
<script src="../../resources/testharness.js"></script>
<script src="../../resources/testharnessreport.js"></script>
<div id=log></div>
<script>
test(function() {
var df = document.createDocumentFragment(),
t1 = document.createTextNode("1"),
t2 = document.createTextNode("2"),
t3 = document.createTextNode("3"),
t4 = document.createTextNode("4")
df.appendChild(t1)
df.appendChild(t2)
assert_equals(df.childNodes.length, 2)
assert_equals(df.textContent, "12")
var el = document.createElement('x')
df.appendChild(el)
el.appendChild(t3)
el.appendChild(t4)
document.normalize()
assert_equals(el.childNodes.length, 2)
assert_equals(el.textContent, "34")
assert_equals(df.childNodes.length, 3)
assert_equals(t1.data, "1")
df.normalize()
assert_equals(df.childNodes.length, 2)
assert_equals(df.firstChild, t1)
assert_equals(t1.data, "12")
assert_equals(t2.data, "2")
assert_equals(el.firstChild, t3)
assert_equals(t3.data, "34")
assert_equals(t4.data, "4")
})
// https://www.w3.org/Bugs/Public/show_bug.cgi?id=19837
test(function() {
var div = document.createElement("div")
var t1 = div.appendChild(document.createTextNode(""))
var t2 = div.appendChild(document.createTextNode("a"))
var t3 = div.appendChild(document.createTextNode(""))
assert_array_equals(div.childNodes, [t1, t2, t3])
div.normalize();
assert_array_equals(div.childNodes, [t2])
}, "Empty text nodes separated by a non-empty text node")
test(function() {
var div = document.createElement("div")
var t1 = div.appendChild(document.createTextNode(""))
var t2 = div.appendChild(document.createTextNode(""))
assert_array_equals(div.childNodes, [t1, t2])
div.normalize();
assert_array_equals(div.childNodes, [])
}, "Empty text nodes")
// The specification for normalize is clear that only "exclusive Text
// nodes" are to be modified. This excludes CDATASection nodes, which
// derive from Text. Naïve implementations may fail to skip
// CDATASection nodes, or even worse, try to test textContent or
// nodeValue without taking care to check the node type. They will
// fail this test.
test(function() {
// We create an XML document so that we can create CDATASection.
// Except for the CDATASection the result should be the same for
// an HTML document. (No non-Text node should be touched.)
var doc = new DOMParser().parseFromString("<div/>", "text/xml")
var div = doc.documentElement
var t1 = div.appendChild(doc.createTextNode("a"))
// The first parameter is the "target" of the processing
// instruction, and the 2nd is the text content.
var t2 = div.appendChild(doc.createProcessingInstruction("pi", ""))
var t3 = div.appendChild(doc.createTextNode("b"))
var t4 = div.appendChild(doc.createCDATASection(""))
var t5 = div.appendChild(doc.createTextNode("c"))
var t6 = div.appendChild(doc.createComment(""))
var t7 = div.appendChild(doc.createTextNode("d"))
var t8 = div.appendChild(doc.createElement("el"))
var t9 = div.appendChild(doc.createTextNode("e"))
var expected = [t1, t2, t3, t4, t5, t6, t7, t8, t9]
assert_array_equals(div.childNodes, expected)
div.normalize();
assert_array_equals(div.childNodes, expected)
}, "Non-text nodes with empty textContent values.")
</script>