Text manipulation: first and last unit in a paragraph should not contain only excluded tokens
https://bugs.webkit.org/show_bug.cgi?id=212759
Reviewed by Wenson Hsieh.
Source/WebCore:
In r262398, we literally made text of one Node as the minimum unit for text manipulation. This patches introduce
a struct ManipulationUnit for that. Now a paragraph can be represented as multiple ManipulationUnits. When all
tokens in a ManipulationUnit are excluded, it means the ManipulationUnit is excluded and should not be
manipulated. To record ManipulationUnits in a paragraph based on our current implementation, we need to keep the
excluded ManipulationUnits surrounded by non-excluded ManipulationUnits, but we can safely remove the leading
and trailing excluded ManipulationUnits. In this case, we can limit the range of paragraph further and thus less
text replacement work.
Covered by existing test.
* editing/TextManipulationController.cpp:
(WebCore::TextManipulationController::parse):
(WebCore::TextManipulationController::addItemIfPossible):
(WebCore::TextManipulationController::observeParagraphs):
* editing/TextManipulationController.h:
Tools:
Modify existing test for changed behavior that leading and trailing excluded units are not included in paragraph
now.
* TestWebKitAPI/Tests/WebKitCocoa/TextManipulation.mm:
(TestWebKitAPI::TEST):
git-svn-id: http://svn.webkit.org/repository/webkit/trunk@262601 268f45cc-cd09-0410-ab3c-d52691b4dbfc
diff --git a/Source/WebCore/editing/TextManipulationController.cpp b/Source/WebCore/editing/TextManipulationController.cpp
index b3de989..2b36397 100644
--- a/Source/WebCore/editing/TextManipulationController.cpp
+++ b/Source/WebCore/editing/TextManipulationController.cpp
@@ -294,12 +294,13 @@
return false;
}
-TextManipulationController::ManipulationTokens TextManipulationController::parse(StringView text, Node* textNode)
+TextManipulationController::ManipulationUnit TextManipulationController::parse(StringView text, Node* textNode)
{
Vector<ManipulationToken> tokens;
ExclusionRuleMatcher exclusionRuleMatcher(m_exclusionRules);
size_t positionOfLastNonHTMLSpace = WTF::notFound;
size_t startPositionOfCurrentToken = 0;
+ bool isNodeExcluded = exclusionRuleMatcher.isExcluded(textNode);
bool containsOnlyHTMLSpace = true;
bool containsLineBreak = false;
bool firstTokenContainsLineBreak = false;
@@ -315,7 +316,7 @@
containsLineBreak = true;
if (positionOfLastNonHTMLSpace != WTF::notFound && startPositionOfCurrentToken <= positionOfLastNonHTMLSpace) {
auto tokenString = text.substring(startPositionOfCurrentToken, positionOfLastNonHTMLSpace + 1 - startPositionOfCurrentToken).toString();
- tokens.append(ManipulationToken { m_tokenIdentifier.generate(), tokenString, tokenInfo(textNode), exclusionRuleMatcher.isExcluded(textNode) });
+ tokens.append(ManipulationToken { m_tokenIdentifier.generate(), tokenString, tokenInfo(textNode), isNodeExcluded });
startPositionOfCurrentToken = positionOfLastNonHTMLSpace + 1;
}
@@ -337,11 +338,36 @@
if (startPositionOfCurrentToken < text.length()) {
auto tokenString = text.substring(startPositionOfCurrentToken, index + 1 - startPositionOfCurrentToken).toString();
- tokens.append(ManipulationToken { m_tokenIdentifier.generate(), tokenString, tokenInfo(textNode), exclusionRuleMatcher.isExcluded(textNode) });
+ tokens.append(ManipulationToken { m_tokenIdentifier.generate(), tokenString, tokenInfo(textNode), isNodeExcluded });
lastTokenContainsLineBreak = false;
}
- return { WTFMove(tokens), containsOnlyHTMLSpace, containsLineBreak, firstTokenContainsLineBreak, lastTokenContainsLineBreak };
+ return { WTFMove(tokens), *textNode, containsOnlyHTMLSpace || isNodeExcluded, containsLineBreak, firstTokenContainsLineBreak, lastTokenContainsLineBreak };
+}
+
+void TextManipulationController::addItemIfPossible(Vector<ManipulationUnit>&& units)
+{
+ if (units.isEmpty())
+ return;
+
+ size_t index = 0;
+ size_t end = units.size();
+ while (index < units.size() && units[index].areAllTokensExcluded)
+ ++index;
+
+ while (end > 0 && units[end - 1].areAllTokensExcluded)
+ --end;
+
+ if (index == end)
+ return;
+
+ auto startPosition = firstPositionInOrBeforeNode(units.first().node.ptr());
+ auto endPosition = positionAfterNode(units.last().node.ptr());
+ Vector<ManipulationToken> tokens;
+ for (; index < end; ++index)
+ tokens.appendVector(WTFMove(units[index].tokens));
+
+ addItem(ManipulationItemData { startPosition, endPosition, nullptr, nullQName(), WTFMove(tokens) });
}
void TextManipulationController::observeParagraphs(const Position& start, const Position& end)
@@ -351,24 +377,20 @@
auto document = makeRefPtr(start.document());
ASSERT(document);
- ParagraphContentIterator iterator { start, end };
// TextIterator's constructor may have updated the layout and executed arbitrary scripts.
if (document != start.document() || document != end.document())
return;
- Vector<ManipulationToken> tokensInCurrentParagraph;
- Position startOfCurrentParagraph;
- Position endOfCurrentParagraph;
+ Vector<ManipulationUnit> unitsInCurrentParagraph;
RefPtr<Element> enclosingItemBoundaryElement;
-
+ ParagraphContentIterator iterator { start, end };
for (; !iterator.atEnd(); iterator.advance()) {
auto content = iterator.currentContent();
auto* contentNode = content.node.get();
ASSERT(contentNode);
if (enclosingItemBoundaryElement && !enclosingItemBoundaryElement->contains(contentNode)) {
- if (!tokensInCurrentParagraph.isEmpty())
- addItem(ManipulationItemData { startOfCurrentParagraph, endOfCurrentParagraph, nullptr, nullQName(), std::exchange(tokensInCurrentParagraph, { }) });
+ addItemIfPossible(std::exchange(unitsInCurrentParagraph, { }));
enclosingItemBoundaryElement = nullptr;
}
@@ -401,37 +423,28 @@
}
if (content.isReplacedContent) {
- if (!tokensInCurrentParagraph.isEmpty()) {
- tokensInCurrentParagraph.append(ManipulationToken { m_tokenIdentifier.generate(), "[]", tokenInfo(content.node.get()), true });
- endOfCurrentParagraph = positionAfterNode(contentNode);
- }
+ if (!unitsInCurrentParagraph.isEmpty())
+ unitsInCurrentParagraph.append(ManipulationUnit { { ManipulationToken { m_tokenIdentifier.generate(), "[]", tokenInfo(content.node.get()), true } }, *contentNode });
continue;
}
if (!content.isTextContent)
continue;
- auto tokensInCurrentNode = parse(content.text, contentNode);
- if (!tokensInCurrentParagraph.isEmpty() && tokensInCurrentNode.firstTokenContainsLineBreak)
- addItem(ManipulationItemData { startOfCurrentParagraph, endOfCurrentParagraph, nullptr, nullQName(), std::exchange(tokensInCurrentParagraph, { }) });
+ auto unitsInCurrentNode = parse(content.text, contentNode);
+ if (unitsInCurrentNode.firstTokenContainsLineBreak)
+ addItemIfPossible(std::exchange(unitsInCurrentParagraph, { }));
- if (tokensInCurrentParagraph.isEmpty()) {
- if (tokensInCurrentNode.containsOnlyHTMLSpace)
+ if (unitsInCurrentParagraph.isEmpty() && unitsInCurrentNode.areAllTokensExcluded)
continue;
- startOfCurrentParagraph = firstPositionInOrBeforeNode(contentNode);
- }
- tokensInCurrentParagraph.appendVector(tokensInCurrentNode.tokens);
- endOfCurrentParagraph = positionAfterNode(contentNode);
+ unitsInCurrentParagraph.append(WTFMove(unitsInCurrentNode));
- if (!tokensInCurrentParagraph.isEmpty() && tokensInCurrentNode.lastTokenContainsLineBreak) {
- ASSERT(!tokensInCurrentParagraph.isEmpty());
- addItem(ManipulationItemData { startOfCurrentParagraph, endOfCurrentParagraph, nullptr, nullQName(), std::exchange(tokensInCurrentParagraph, { }) });
- }
+ if (unitsInCurrentNode.lastTokenContainsLineBreak)
+ addItemIfPossible(std::exchange(unitsInCurrentParagraph, { }));
}
- if (!tokensInCurrentParagraph.isEmpty())
- addItem(ManipulationItemData { startOfCurrentParagraph, endOfCurrentParagraph, nullptr, nullQName(), WTFMove(tokensInCurrentParagraph) });
+ addItemIfPossible(std::exchange(unitsInCurrentParagraph, { }));
}
void TextManipulationController::didCreateRendererForElement(Element& element)
