| /* |
| * (C) 1999 Lars Knoll ([email protected]) |
| * (C) 2000 Dirk Mueller ([email protected]) |
| * Copyright (C) 2004-2023 Apple Inc. All rights reserved. |
| * Copyright (C) 2015 Google Inc. All rights reserved. |
| * Copyright (C) 2006 Andrew Wellington ([email protected]) |
| * Copyright (C) 2006 Graham Dennis ([email protected]) |
| * |
| * This library is free software; you can redistribute it and/or |
| * modify it under the terms of the GNU Library General Public |
| * License as published by the Free Software Foundation; either |
| * version 2 of the License, or (at your option) any later version. |
| * |
| * This library is distributed in the hope that it will be useful, |
| * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| * Library General Public License for more details. |
| * |
| * You should have received a copy of the GNU Library General Public License |
| * along with this library; see the file COPYING.LIB. If not, write to |
| * the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor, |
| * Boston, MA 02110-1301, USA. |
| * |
| */ |
| |
| #include "config.h" |
| #include "RenderText.h" |
| |
| #include "AXObjectCache.h" |
| #include "BreakablePositions.h" |
| #include "BreakingContext.h" |
| #include "DocumentMarkerController.h" |
| #include "DocumentView.h" |
| #include "FloatQuad.h" |
| #include "Hyphenation.h" |
| #include "InlineIteratorBoxInlines.h" |
| #include "InlineIteratorLineBoxInlines.h" |
| #include "InlineIteratorLogicalOrderTraversal.h" |
| #include "InlineIteratorSVGTextBox.h" |
| #include "InlineIteratorTextBoxInlines.h" |
| #include "InlineRunAndOffset.h" |
| #include "LayoutInlineTextBox.h" |
| #include "LayoutIntegrationLineLayout.h" |
| #include "LogicalSelectionOffsetCachesInlines.h" |
| #include "LineSelection.h" |
| #include "LocalFrame.h" |
| #include "LocalFrameView.h" |
| #include "MathVariant.h" |
| #include "Range.h" |
| #include "RenderBlock.h" |
| #include "RenderCombineText.h" |
| #include "RenderInline.h" |
| #include "RenderLayer.h" |
| #include "RenderObjectInlines.h" |
| #include "RenderTextInlines.h" |
| #include "RenderSVGInlineText.h" |
| #include "RenderView.h" |
| #include "RenderedDocumentMarker.h" |
| #include "SVGElementTypeHelpers.h" |
| #include "SVGInlineTextBox.h" |
| #include "SelectionGeometry.h" |
| #include "Settings.h" |
| #include "SurrogatePairAwareTextIterator.h" |
| #include "Text.h" |
| #include "TextResourceDecoder.h" |
| #include "TextTransform.h" |
| #include "TextUtil.h" |
| #include "VisiblePosition.h" |
| #include "WidthIterator.h" |
| #include <wtf/BitSet.h> |
| #include <wtf/NeverDestroyed.h> |
| #include <wtf/SortedArrayMap.h> |
| #include <wtf/TZoneMallocInlines.h> |
| #include <wtf/text/CharacterProperties.h> |
| #include <wtf/text/StringBuilder.h> |
| #include <wtf/text/TextBreakIterator.h> |
| #include <wtf/unicode/CharacterNames.h> |
| |
| #if PLATFORM(IOS_FAMILY) |
| #include "Document.h" |
| #include "EditorClient.h" |
| #include "Page.h" |
| #endif |
| |
| namespace WebCore { |
| |
| using namespace WTF::Unicode; |
| |
| WTF_MAKE_TZONE_ALLOCATED_IMPL(RenderText); |
| |
| struct SameSizeAsRenderText : public RenderObject { |
| #if ENABLE(TEXT_AUTOSIZING) |
| float candidateTextSize; |
| #endif |
| float widths[4]; |
| String text; |
| std::optional<bool> canUseSimplifiedTextMeasuring; |
| std::optional<bool> hasPositionDependentContentWidth; |
| std::optional<bool> m_hasStrongDirectionalityContent; |
| uint32_t bitfields : 14; |
| }; |
| |
| static_assert(sizeof(RenderText) == sizeof(SameSizeAsRenderText), "RenderText should stay small"); |
| |
| class SecureTextTimer final : private TimerBase { |
| WTF_MAKE_TZONE_ALLOCATED_INLINE(SecureTextTimer); |
| public: |
| explicit SecureTextTimer(RenderText&); |
| void restart(unsigned offsetAfterLastTypedCharacter); |
| |
| unsigned takeOffsetAfterLastTypedCharacter(); |
| |
| private: |
| void fired() override; |
| RenderText& m_renderer; |
| unsigned m_offsetAfterLastTypedCharacter { 0 }; |
| }; |
| |
| using SecureTextTimerMap = SingleThreadWeakHashMap<RenderText, std::unique_ptr<SecureTextTimer>>; |
| |
| static SecureTextTimerMap& secureTextTimers() |
| { |
| static NeverDestroyed<SecureTextTimerMap> map; |
| return map.get(); |
| } |
| |
| inline SecureTextTimer::SecureTextTimer(RenderText& renderer) |
| : m_renderer(renderer) |
| { |
| } |
| |
| inline void SecureTextTimer::restart(unsigned offsetAfterLastTypedCharacter) |
| { |
| m_offsetAfterLastTypedCharacter = offsetAfterLastTypedCharacter; |
| startOneShot(1_s * m_renderer.settings().passwordEchoDurationInSeconds()); |
| } |
| |
| inline unsigned SecureTextTimer::takeOffsetAfterLastTypedCharacter() |
| { |
| unsigned offset = m_offsetAfterLastTypedCharacter; |
| m_offsetAfterLastTypedCharacter = 0; |
| return offset; |
| } |
| |
| void SecureTextTimer::fired() |
| { |
| ASSERT(secureTextTimers().get(m_renderer) == this); |
| m_offsetAfterLastTypedCharacter = 0; |
| m_renderer.setText(m_renderer.text(), true /* forcing setting text as it may be masked later */); |
| } |
| |
| static HashMap<SingleThreadWeakRef<const RenderText>, String>& originalTextMap() |
| { |
| static NeverDestroyed<HashMap<SingleThreadWeakRef<const RenderText>, String>> map; |
| return map; |
| } |
| |
| static HashMap<SingleThreadWeakRef<const RenderText>, SingleThreadWeakPtr<RenderInline>>& inlineWrapperForDisplayContentsMap() |
| { |
| static NeverDestroyed<HashMap<SingleThreadWeakRef<const RenderText>, SingleThreadWeakPtr<RenderInline>>> map; |
| return map; |
| } |
| |
| static constexpr char16_t convertNoBreakSpaceToSpace(char16_t character) |
| { |
| return character == noBreakSpace ? ' ' : character; |
| } |
| |
| static inline size_t capitalizeCharacter(String textContent, unsigned startCharacterOffset, StringBuilder& output) |
| { |
| if (startCharacterOffset >= textContent.length()) { |
| ASSERT_NOT_REACHED(); |
| return 0; |
| } |
| |
| auto capitalize = [&](const char16_t* contentToCapitalize, size_t length) -> size_t { |
| if (length == 1) { |
| if ((*contentToCapitalize >= 'A' && *contentToCapitalize <= 'Z') || *contentToCapitalize == ' ') |
| return 0; |
| if (*contentToCapitalize <= 'z') { |
| char32_t lastCharacter = u_totitle(*contentToCapitalize); |
| output.append(lastCharacter); |
| return 1; |
| } |
| } |
| |
| char16_t capitalizedCharacter; |
| UErrorCode status = U_ZERO_ERROR; |
| auto realLength = u_strToTitle(&capitalizedCharacter, 1, contentToCapitalize, length, nullptr, "", &status); |
| if (U_SUCCESS(status) && realLength == 1) { |
| output.append(capitalizedCharacter); |
| return realLength; |
| } |
| |
| // Decomposed ligatures may need more space. |
| std::span<char16_t> capitalizedStringData; |
| auto capitalizedString = String::createUninitialized(realLength, capitalizedStringData); |
| status = U_ZERO_ERROR; |
| u_strToTitle(capitalizedStringData.data(), capitalizedStringData.size(), contentToCapitalize, length, nullptr, "", &status); |
| if (U_SUCCESS(status)) { |
| output.append(capitalizedString); |
| return length; |
| } |
| return 0; |
| }; |
| |
| if (textContent.is8Bit()) { |
| auto characterToCapitalize = textContent[startCharacterOffset]; |
| return capitalize(&characterToCapitalize, 1); |
| } |
| |
| auto content = textContent.span16().subspan(startCharacterOffset); |
| auto contentIterator = SurrogatePairAwareTextIterator { content, startCharacterOffset, textContent.length() }; |
| unsigned capitalizedContentLength = 0; |
| char32_t currentCharacter = 0; |
| contentIterator.consume(currentCharacter, capitalizedContentLength); |
| if (startCharacterOffset + capitalizedContentLength > textContent.length()) { |
| ASSERT_NOT_REACHED(); |
| return 0; |
| } |
| return capitalize(content.data(), capitalizedContentLength); |
| } |
| |
| String capitalize(const String& string) |
| { |
| Vector<char16_t> previousCharacter(1, ' '); |
| return capitalize(string, previousCharacter); |
| } |
| |
| String capitalize(const String& string, Vector<char16_t> previousCharacter) |
| { |
| int32_t length = string.length(); |
| int32_t previousCharacterLength = previousCharacter.size(); |
| auto& stringImpl = *string.impl(); |
| |
| static_assert(String::MaxLength < std::numeric_limits<unsigned>::max(), "Must be able to add one without overflowing unsigned"); |
| |
| // Replace NO BREAK SPACE with a normal spaces since ICU does not treat it as a word separator. |
| Vector<char16_t> stringWithPrevious(previousCharacterLength + length); |
| for (int32_t i = 0; i < previousCharacterLength; ++i) |
| stringWithPrevious[i] = convertNoBreakSpaceToSpace(previousCharacter[i]); |
| for (int32_t i = previousCharacterLength; i < length + previousCharacterLength; ++i) |
| stringWithPrevious[i] = convertNoBreakSpaceToSpace(stringImpl[i - previousCharacterLength]); |
| |
| auto* breakIterator = WTF::wordBreakIterator(stringWithPrevious.span()); |
| if (!breakIterator) |
| return string; |
| |
| StringBuilder result; |
| result.reserveCapacity(length); |
| |
| int32_t startOfWord = ubrk_first(breakIterator); |
| for (int32_t endOfWord = ubrk_next(breakIterator); endOfWord != UBRK_DONE; startOfWord = endOfWord, endOfWord = ubrk_next(breakIterator)) { |
| // Do not try to titlecase the previous content. |
| if (startOfWord >= previousCharacterLength) { |
| auto capitalizedContentLength = capitalizeCharacter(string, startOfWord - previousCharacterLength, result); |
| for (int32_t i = startOfWord + capitalizedContentLength; i < endOfWord; ++i) |
| result.append(stringImpl[i - previousCharacterLength]); |
| } else { |
| // This is previous and continous non-titlecased current content. Only append current content. |
| for (int32_t i = startOfWord; i < endOfWord; ++i) { |
| if (i < previousCharacterLength) |
| continue; |
| result.append(stringImpl[i - previousCharacterLength]); |
| } |
| } |
| } |
| return result == string ? string : result.toString(); |
| } |
| |
| static LayoutRect selectionRectForTextBox(const InlineIterator::TextBox& textBox, unsigned rangeStart, unsigned rangeEnd) |
| { |
| if (auto* svgTextBox = dynamicDowncast<InlineIterator::SVGTextBox>(textBox)) |
| return svgTextBox->localSelectionRect(rangeStart, rangeEnd); |
| |
| bool isCaretCase = rangeStart == rangeEnd; |
| |
| auto [clampedStart, clampedEnd] = textBox.selectableRange().clamp(rangeStart, rangeEnd); |
| |
| if (clampedStart >= clampedEnd) { |
| if (isCaretCase) { |
| // handle unitary range, e.g.: representing caret position |
| bool isCaretWithinTextBox = rangeStart >= textBox.start() && rangeStart < textBox.end(); |
| // For last text box in a InlineTextBox chain, we allow the caret to move to a position 'after' the end of the last text box. |
| bool isCaretWithinLastTextBox = rangeStart >= textBox.start() && rangeStart <= textBox.end(); |
| |
| auto isLastTextBox = !textBox.nextTextBox(); |
| |
| if ((isLastTextBox && !isCaretWithinLastTextBox) || (!isLastTextBox && !isCaretWithinTextBox)) |
| return { }; |
| } else { |
| bool isRangeWithinTextBox = (rangeStart >= textBox.start() && rangeStart <= textBox.end()); |
| if (!isRangeWithinTextBox) |
| return { }; |
| } |
| } |
| |
| auto lineSelectionRect = LineSelection::logicalRect(*textBox.lineBox()); |
| auto selectionRect = LayoutRect { 0, lineSelectionRect.y(), textBox.logicalWidth(), lineSelectionRect.height() }; |
| |
| auto textRun = textBox.textRun(); |
| if (clampedStart || clampedEnd != textRun.length()) |
| textBox.fontCascade().adjustSelectionRectForText(textBox.renderer().canUseSimplifiedTextMeasuring().value_or(false), textRun, selectionRect, clampedStart, clampedEnd); |
| |
| if (!textBox.writingMode().isLogicalLeftLineLeft()) |
| selectionRect.setX(textBox.logicalWidth() - selectionRect.maxX()); |
| selectionRect.move(textBox.logicalLeftIgnoringInlineDirection(), 0); |
| |
| return snappedSelectionRect(selectionRect, textBox.logicalRightIgnoringInlineDirection(), textBox.writingMode()); |
| } |
| |
| static unsigned offsetForPositionInRun(const InlineIterator::TextBox& textBox, float x) |
| { |
| if (textBox.isLineBreak()) |
| return 0; |
| |
| auto runPosition = x - textBox.logicalLeftIgnoringInlineDirection(); |
| if (runPosition > textBox.logicalWidth()) |
| return textBox.isInlineFlipped() ? 0 : textBox.length(); |
| if (runPosition < 0) |
| return textBox.isInlineFlipped() ? textBox.length() : 0; |
| |
| if (!textBox.writingMode().isLogicalLeftLineLeft()) |
| runPosition = textBox.logicalWidth() - runPosition; |
| return textBox.fontCascade().offsetForPosition(textBox.textRun(InlineIterator::TextRunMode::Editing), runPosition, true); |
| } |
| |
| inline RenderText::RenderText(Type type, Node& node, const String& text) |
| : RenderObject(type, node, TypeFlag::IsText, { }) |
| , m_text(text) |
| , m_containsOnlyASCII(text.impl()->containsOnlyASCII()) |
| { |
| ASSERT(!m_text.isNull()); |
| computeFontCodePath(); |
| ASSERT(isRenderText()); |
| } |
| |
| RenderText::RenderText(Type type, Text& textNode, const String& text) |
| : RenderText(type, static_cast<Node&>(textNode), text) |
| { |
| } |
| |
| RenderText::RenderText(Type type, Document& document, const String& text) |
| : RenderText(type, static_cast<Node&>(document), text) |
| { |
| } |
| |
| RenderText::~RenderText() |
| { |
| // Do not add any code here. Add it to willBeDestroyed() instead. |
| ASSERT(!originalTextMap().contains(this)); |
| } |
| |
| Layout::InlineTextBox* RenderText::layoutBox() |
| { |
| return downcast<Layout::InlineTextBox>(RenderObject::layoutBox()); |
| } |
| |
| const Layout::InlineTextBox* RenderText::layoutBox() const |
| { |
| return downcast<Layout::InlineTextBox>(RenderObject::layoutBox()); |
| } |
| |
| ASCIILiteral RenderText::renderName() const |
| { |
| return "RenderText"_s; |
| } |
| |
| Text* RenderText::textNode() const |
| { |
| return downcast<Text>(RenderObject::node()); |
| } |
| |
| bool RenderText::computeUseBackslashAsYenSymbol() const |
| { |
| const RenderStyle& style = this->style(); |
| const auto& fontDescription = style.fontDescription(); |
| if (style.fontCascade().useBackslashAsYenSymbol()) |
| return true; |
| if (fontDescription.isSpecifiedFont()) |
| return false; |
| const PAL::TextEncoding* encoding = document().decoder() ? &document().decoder()->encoding() : 0; |
| if (encoding && encoding->backslashAsCurrencySymbol() != '\\') |
| return true; |
| return false; |
| } |
| |
| void RenderText::initiateFontLoadingByAccessingGlyphDataAndComputeCanUseSimplifiedTextMeasuring(const String& textContent) |
| { |
| auto& style = this->style(); |
| auto& fontCascade = style.fontCascade(); |
| // See webkit.org/b/252668 |
| auto fontVariant = AutoVariant; |
| m_canUseSimplifiedTextMeasuring = canUseSimpleFontCodePath(); |
| #if USE(FONT_VARIANT_VIA_FEATURES) |
| auto fontVariantCaps = fontCascade.fontDescription().variantCaps(); |
| if (fontVariantCaps == FontVariantCaps::Small || fontVariantCaps == FontVariantCaps::AllSmall || fontVariantCaps == FontVariantCaps::Petite || fontVariantCaps == FontVariantCaps::AllPetite) { |
| // This matches the behavior of ComplexTextController::collectComplexTextRuns(): that function doesn't perform font fallback |
| // on the capitalized characters when small caps is enabled, so we shouldn't here either. |
| fontVariant = NormalVariant; |
| m_canUseSimplifiedTextMeasuring = false; |
| } |
| #endif |
| auto whitespaceIsCollapsed = style.collapseWhiteSpace(); |
| Ref primaryFont = fontCascade.primaryFont(); |
| m_canUseSimplifiedTextMeasuring = *m_canUseSimplifiedTextMeasuring && !fontCascade.wordSpacing() && !fontCascade.letterSpacing() && !primaryFont->syntheticBoldOffset() && (&firstLineStyle() == &style || &fontCascade == &firstLineStyle().fontCascade()); |
| |
| if (*m_canUseSimplifiedTextMeasuring) { |
| // Additional check on the font codepath. |
| auto run = TextRun { textContent }; |
| run.setCharacterScanForCodePath(false); |
| m_canUseSimplifiedTextMeasuring = fontCascade.codePath(run) == FontCascade::CodePath::Simple; |
| } |
| |
| m_hasPositionDependentContentWidth = false; |
| m_hasStrongDirectionalityContent = false; |
| auto mayHaveStrongDirectionalityContent = !textContent.is8Bit(); |
| // FIXME: Pre-warm glyph loading in FontCascade with the most common range. |
| WTF::BitSet<256> hasSeen; |
| for (char32_t character : StringView(textContent).codePoints()) { |
| if (character < 256) { |
| if (hasSeen.testAndSet(character)) |
| continue; |
| } |
| m_canUseSimplifiedTextMeasuring = *m_canUseSimplifiedTextMeasuring && fontCascade.canUseSimplifiedTextMeasuring(character, fontVariant, whitespaceIsCollapsed, primaryFont); |
| m_hasPositionDependentContentWidth = *m_hasPositionDependentContentWidth || character == tabCharacter; |
| m_hasStrongDirectionalityContent = *m_hasStrongDirectionalityContent || (mayHaveStrongDirectionalityContent && Layout::TextUtil::isStrongDirectionalityCharacter(character)); |
| } |
| } |
| |
| void RenderText::styleDidChange(Style::Difference diff, const RenderStyle* oldStyle) |
| { |
| // There is no need to ever schedule repaints from a style change of a text run, since |
| // we already did this for the parent of the text run. |
| // We do have to schedule layouts, though, since a style change can force us to |
| // need to relayout. |
| if (diff == Style::DifferenceResult::Layout) { |
| setNeedsLayoutAndPreferredWidthsUpdate(); |
| m_knownToHaveNoOverflowAndNoFallbackFonts = false; |
| } |
| |
| const RenderStyle& newStyle = style(); |
| if (!oldStyle) |
| initiateFontLoadingByAccessingGlyphDataAndComputeCanUseSimplifiedTextMeasuring(m_text); |
| if (oldStyle && !oldStyle->fontCascadeEqual(newStyle)) |
| m_canUseSimplifiedTextMeasuring = { }; |
| |
| bool needsResetText = false; |
| if (!oldStyle) { |
| m_useBackslashAsYenSymbol = computeUseBackslashAsYenSymbol(); |
| needsResetText = m_useBackslashAsYenSymbol; |
| } else if (oldStyle->fontCascade().useBackslashAsYenSymbol() != newStyle.fontCascade().useBackslashAsYenSymbol()) { |
| m_useBackslashAsYenSymbol = computeUseBackslashAsYenSymbol(); |
| needsResetText = true; |
| } |
| |
| auto oldTransform = oldStyle ? oldStyle->textTransform() : Style::TextTransform { CSS::Keyword::None { } }; |
| TextSecurity oldSecurity = oldStyle ? oldStyle->textSecurity() : TextSecurity::None; |
| if (needsResetText || oldTransform != newStyle.textTransform() || oldSecurity != newStyle.textSecurity()) |
| RenderText::setText(originalText(), true); |
| |
| // FIXME: First line change on the block comes in as equal on text. |
| auto needsLayoutBoxStyleUpdate = layoutBox() && (diff >= Style::DifferenceResult::RecompositeLayer || (&style() != &firstLineStyle())); |
| if (needsLayoutBoxStyleUpdate) |
| LayoutIntegration::LineLayout::updateStyle(*this); |
| |
| if (CheckedPtr cache = document().existingAXObjectCache()) |
| cache->onStyleChange(*this, diff, oldStyle, newStyle); |
| |
| setHorizontalWritingMode(newStyle.writingMode().isHorizontal()); |
| } |
| |
| void RenderText::willBeDestroyed() |
| { |
| if (m_hasSecureTextTimer) |
| secureTextTimers().remove(*this); |
| |
| if (m_originalTextDiffersFromRendered) |
| originalTextMap().remove(this); |
| |
| setInlineWrapperForDisplayContents(nullptr); |
| |
| RenderObject::willBeDestroyed(); |
| } |
| |
| String RenderText::originalText() const |
| { |
| return m_originalTextDiffersFromRendered ? originalTextMap().get(this) : m_text; |
| } |
| |
| void RenderText::boundingRects(Vector<LayoutRect>& rects, const LayoutPoint& accumulatedOffset) const |
| { |
| for (auto& run : InlineIterator::textBoxesFor(*this)) { |
| auto rect = LayoutRect { run.visualRectIgnoringBlockDirection() }; |
| rect.moveBy(accumulatedOffset); |
| rects.append(rect); |
| } |
| } |
| |
| Vector<IntRect> RenderText::absoluteRectsForRange(unsigned start, unsigned end, bool useSelectionHeight, bool* wasFixed) const |
| { |
| OptionSet<RenderObject::BoundingRectBehavior> behavior; |
| if (useSelectionHeight) |
| behavior.add(RenderObject::BoundingRectBehavior::UseSelectionHeight); |
| return absoluteQuadsForRange(start, end, behavior, wasFixed).map([](auto& quad) { |
| return quad.enclosingBoundingBox(); |
| }); |
| } |
| |
| // This function is similar in spirit to addLineBoxRects, but returns rectangles |
| // which are annotated with additional state which helps the iPhone draw selections in its unique way. |
| // Full annotations are added in this class. |
| void RenderText::collectSelectionGeometries(Vector<SelectionGeometry>& rects, unsigned start, unsigned end) |
| { |
| for (auto textBox = InlineIterator::lineLeftmostTextBoxFor(*this); textBox; textBox = textBox.traverseNextTextBox()) { |
| LayoutRect rect; |
| if (start <= textBox->start() && textBox->end() <= end) |
| rect = selectionRectForTextBox(*textBox, start, end); |
| else { |
| unsigned realEnd = std::min(textBox->end(), end); |
| rect = selectionRectForTextBox(*textBox, start, realEnd); |
| if (rect.isEmpty()) |
| continue; |
| } |
| |
| bool separateLines = false; |
| if (textBox->lineBox()->isFirstAfterPageBreak()) { |
| // The top of the second line in a column aligns with the bottom of the first line in adjacent columns, |
| // which can cause the geometries united, extending the highlight box into unselected content. |
| // To prevent this, use setSeparateFromPreviousLine for each column to ensure the geometries won't be united. |
| separateLines = true; |
| if (textBox->isHorizontal()) |
| rect.shiftYEdgeTo(textBox->lineBox()->logicalTop()); |
| else |
| rect.shiftXEdgeTo(textBox->lineBox()->logicalTop()); |
| } |
| |
| RenderBlock* containingBlock = this->containingBlock(); |
| // Map rect, extended left to leftOffset, and right to rightOffset, through transforms to get minX and maxX. |
| LogicalSelectionOffsetCaches cache(*containingBlock); |
| LayoutUnit leftOffset = containingBlock->logicalLeftSelectionOffset(*containingBlock, LayoutUnit(textBox->logicalTop()), cache); |
| LayoutUnit rightOffset = containingBlock->logicalRightSelectionOffset(*containingBlock, LayoutUnit(textBox->logicalTop()), cache); |
| LayoutRect extentsRect = rect; |
| if (textBox->isHorizontal()) { |
| extentsRect.setX(leftOffset); |
| extentsRect.setWidth(rightOffset - leftOffset); |
| } else { |
| extentsRect.setY(leftOffset); |
| extentsRect.setHeight(rightOffset - leftOffset); |
| } |
| extentsRect = localToAbsoluteQuad(FloatRect(extentsRect)).enclosingBoundingBox(); |
| if (!textBox->isHorizontal()) |
| extentsRect = extentsRect.transposedRect(); |
| bool isFirstOnLine = !textBox->nextLineLeftwardOnLine(); |
| bool isLastOnLine = !textBox->nextLineRightwardOnLine(); |
| |
| bool containsStart = textBox->start() <= start && textBox->end() >= start; |
| bool containsEnd = textBox->start() <= end && textBox->end() >= end; |
| |
| bool isFixed = false; |
| auto absoluteQuad = localToAbsoluteQuad(FloatRect(rect), UseTransforms, &isFixed); |
| bool boxIsHorizontal = !is<InlineIterator::SVGTextBoxIterator>(textBox) ? textBox->isHorizontal() : !writingMode().isVertical(); |
| |
| auto selectionGeometry = SelectionGeometry(absoluteQuad, HTMLElement::selectionRenderingBehavior(textNode()), textBox->direction(), extentsRect.x(), extentsRect.maxX(), extentsRect.maxY(), 0, textBox->isLineBreak(), isFirstOnLine, isLastOnLine, containsStart, containsEnd, boxIsHorizontal, isFixed, view().pageNumberForBlockProgressionOffset(absoluteQuad.enclosingBoundingBox().x())); |
| selectionGeometry.setSeparateFromPreviousLine(separateLines); |
| rects.append(selectionGeometry); |
| } |
| } |
| |
| static std::optional<IntRect> ellipsisRectForTextBox(const InlineIterator::TextBox& textBox, unsigned start, unsigned end) |
| { |
| auto lineBox = textBox.lineBox(); |
| if (!lineBox->hasEllipsis()) |
| return { }; |
| |
| auto selectableRange = textBox.selectableRange(); |
| if (!selectableRange.truncation) |
| return { }; |
| |
| auto ellipsisStartPosition = std::max<unsigned>(start - selectableRange.start, 0); |
| auto ellipsisEndPosition = std::min<unsigned>(end - selectableRange.start, selectableRange.length); |
| // The ellipsis should be considered to be selected if the end of |
| // the selection is past the beginning of the truncation and the |
| // beginning of the selection is before or at the beginning of the truncation. |
| if (ellipsisEndPosition < *selectableRange.truncation && ellipsisStartPosition > *selectableRange.truncation) |
| return { }; |
| |
| return IntRect { lineBox->ellipsisVisualRect(InlineIterator::LineBox::AdjustedForSelection::Yes) }; |
| } |
| |
| enum class ClippingOption { NoClipping, ClipToEllipsis }; |
| |
| // FIXME: Unify with absoluteQuadsForRange. |
| static Vector<FloatQuad> collectAbsoluteQuads(const RenderText& textRenderer, bool* wasFixed, ClippingOption clipping) |
| { |
| Vector<FloatQuad> quads; |
| for (auto& textBox : InlineIterator::textBoxesFor(textRenderer)) { |
| auto* svgTextBox = dynamicDowncast<InlineIterator::SVGTextBox>(textBox); |
| auto boundaries = svgTextBox ? svgTextBox->calculateBoundariesIncludingSVGTransform() : textBox.visualRectIgnoringBlockDirection(); |
| |
| // Shorten the width of this text box if it ends in an ellipsis. |
| if (clipping == ClippingOption::ClipToEllipsis) { |
| if (auto ellipsisRect = ellipsisRectForTextBox(textBox, 0, textRenderer.text().length())) { |
| if (textRenderer.writingMode().isHorizontal()) |
| boundaries.setWidth(ellipsisRect->maxX() - boundaries.x()); |
| else |
| boundaries.setHeight(ellipsisRect->maxY() - boundaries.y()); |
| } |
| } |
| |
| quads.append(textRenderer.localToAbsoluteQuad(boundaries, UseTransforms, wasFixed)); |
| } |
| return quads; |
| } |
| |
| Vector<FloatQuad> RenderText::absoluteQuadsClippedToEllipsis() const |
| { |
| return collectAbsoluteQuads(*this, nullptr, ClippingOption::ClipToEllipsis); |
| } |
| |
| void RenderText::absoluteQuads(Vector<FloatQuad>& quads, bool* wasFixed) const |
| { |
| quads.appendVector(collectAbsoluteQuads(*this, wasFixed, ClippingOption::NoClipping)); |
| } |
| |
| static FloatRect localQuadForTextBox(const InlineIterator::TextBox& textBox, unsigned start, unsigned end, bool useSelectionHeight) |
| { |
| LayoutRect boxSelectionRect = selectionRectForTextBox(textBox, start, end); |
| if (!boxSelectionRect.height()) |
| return { }; |
| if (useSelectionHeight) |
| return boxSelectionRect; |
| |
| auto rect = textBox.visualRectIgnoringBlockDirection(); |
| if (textBox.isHorizontal()) { |
| boxSelectionRect.setHeight(rect.height()); |
| boxSelectionRect.setY(rect.y()); |
| } else { |
| boxSelectionRect.setWidth(rect.width()); |
| boxSelectionRect.setX(rect.x()); |
| } |
| return boxSelectionRect; |
| } |
| |
| static Vector<LayoutRect> characterRects(const InlineIterator::TextBox& run, unsigned rangeStart, unsigned rangeEnd) |
| { |
| auto [clampedStart, clampedEnd] = run.selectableRange().clamp(rangeStart, rangeEnd); |
| if (clampedStart >= clampedEnd) |
| return { }; |
| |
| if (auto* svgTextBox = dynamicDowncast<InlineIterator::SVGTextBox>(run)) { |
| return Vector<LayoutRect>(clampedEnd - clampedStart, [&, clampedStart = clampedStart](size_t i) { |
| size_t index = clampedStart + i; |
| return svgTextBox->localSelectionRect(index, index + 1); |
| }); |
| } |
| |
| auto lineSelectionRect = LineSelection::logicalRect(*run.lineBox()); |
| auto selectionRect = LayoutRect { run.logicalLeftIgnoringInlineDirection(), lineSelectionRect.y(), run.logicalWidth(), lineSelectionRect.height() }; |
| return run.fontCascade().characterSelectionRectsForText(run.textRun(), selectionRect, clampedStart, clampedEnd).map([&](auto& characterRect) { |
| return snappedSelectionRect(characterRect, run.logicalRightIgnoringInlineDirection(), run.writingMode()); |
| }); |
| } |
| |
| Vector<FloatQuad> RenderText::absoluteQuadsForRange(unsigned start, unsigned end, OptionSet<RenderObject::BoundingRectBehavior> behavior, bool* wasFixed) const |
| { |
| bool useSelectionHeight = behavior.contains(RenderObject::BoundingRectBehavior::UseSelectionHeight); |
| bool ignoreEmptyTextSelections = behavior.contains(RenderObject::BoundingRectBehavior::IgnoreEmptyTextSelections); |
| bool computeIndividualCharacterRects = behavior.contains(RenderObject::BoundingRectBehavior::ComputeIndividualCharacterRects); |
| |
| // Work around signed/unsigned issues. This function takes unsigneds, and is often passed UINT_MAX |
| // to mean "all the way to the end". LegacyInlineTextBox coordinates are unsigneds, so changing this |
| // function to take ints causes various internal mismatches. But selectionRect takes ints, and |
| // passing UINT_MAX to it causes trouble. Ideally we'd change selectionRect to take unsigneds, but |
| // that would cause many ripple effects, so for now we'll just clamp our unsigned parameters to INT_MAX. |
| ASSERT(end == UINT_MAX || end <= INT_MAX); |
| ASSERT(start <= INT_MAX); |
| start = std::min(start, static_cast<unsigned>(INT_MAX)); |
| end = std::min(end, static_cast<unsigned>(INT_MAX)); |
| |
| const unsigned caretMinOffset = static_cast<unsigned>(this->caretMinOffset()); |
| const unsigned caretMaxOffset = static_cast<unsigned>(this->caretMaxOffset()); |
| |
| // Narrows |start| and |end| into |caretMinOffset| and |caretMaxOffset| to ignore unrendered leading |
| // and trailing whitespaces. |
| start = std::min(std::max(caretMinOffset, start), caretMaxOffset); |
| end = std::min(std::max(caretMinOffset, end), caretMaxOffset); |
| |
| Vector<FloatQuad> quads; |
| for (auto& textBox : InlineIterator::textBoxesFor(*this)) { |
| if (ignoreEmptyTextSelections && !textBox.selectableRange().intersects(start, end)) |
| continue; |
| |
| if (computeIndividualCharacterRects) { |
| auto rects = characterRects(textBox, start, end); |
| if (!quads.tryReserveCapacity(quads.size() + rects.size())) |
| continue; |
| |
| if (!useSelectionHeight) { |
| for (auto& rect : rects) { |
| auto visualRect = textBox.visualRectIgnoringBlockDirection(); |
| if (textBox.isHorizontal()) { |
| rect.setHeight(visualRect.height()); |
| rect.setY(visualRect.y()); |
| } else { |
| rect.setWidth(visualRect.width()); |
| rect.setX(visualRect.x()); |
| } |
| } |
| } |
| |
| for (auto& rect : rects) { |
| if (FloatRect localRect { rect }; !localRect.isZero()) |
| quads.append(localToAbsoluteQuad(localRect, UseTransforms, wasFixed)); |
| } |
| continue; |
| } |
| |
| if (start <= textBox.start() && textBox.end() <= end) { |
| auto* svgTextBox = dynamicDowncast<InlineIterator::SVGTextBox>(textBox); |
| auto boundaries = svgTextBox ? svgTextBox->calculateBoundariesIncludingSVGTransform() : textBox.visualRectIgnoringBlockDirection(); |
| |
| if (useSelectionHeight) { |
| LayoutRect selectionRect = selectionRectForTextBox(textBox, start, end); |
| if (textBox.isHorizontal()) { |
| boundaries.setHeight(selectionRect.height()); |
| boundaries.setY(selectionRect.y()); |
| } else { |
| boundaries.setWidth(selectionRect.width()); |
| boundaries.setX(selectionRect.x()); |
| } |
| } |
| quads.append(localToAbsoluteQuad(boundaries, UseTransforms, wasFixed)); |
| continue; |
| } |
| FloatRect rect = localQuadForTextBox(textBox, start, end, useSelectionHeight); |
| if (!rect.isZero()) |
| quads.append(localToAbsoluteQuad(rect, UseTransforms, wasFixed)); |
| } |
| return quads; |
| } |
| |
| Position RenderText::positionForPoint(const LayoutPoint& point, HitTestSource source) |
| { |
| return positionForPoint(point, source, nullptr).position(); |
| } |
| |
| enum ShouldAffinityBeDownstream { AlwaysDownstream, AlwaysUpstream, UpstreamIfPositionIsNotAtStart }; |
| |
| static bool lineDirectionPointFitsInBox(int pointLineDirection, const InlineIterator::TextBoxIterator& textRun, ShouldAffinityBeDownstream& shouldAffinityBeDownstream) |
| { |
| shouldAffinityBeDownstream = AlwaysDownstream; |
| |
| // the x coordinate is equal to the left edge of this box |
| // the affinity must be downstream so the position doesn't jump back to the previous line |
| // except when box is the first box in the line |
| if (pointLineDirection <= textRun->logicalLeft()) { |
| shouldAffinityBeDownstream = !textRun->nextLogicalLeftwardOnLine() ? UpstreamIfPositionIsNotAtStart : AlwaysDownstream; |
| return true; |
| } |
| |
| #if !PLATFORM(IOS_FAMILY) |
| // and the x coordinate is to the left of the right edge of this box |
| // check to see if position goes in this box |
| if (pointLineDirection < textRun->logicalRight()) { |
| shouldAffinityBeDownstream = UpstreamIfPositionIsNotAtStart; |
| return true; |
| } |
| #endif |
| |
| // box is first on line |
| // and the x coordinate is to the left of the first text box left edge |
| if (!textRun->nextLogicalLeftwardOnLineIgnoringLineBreak() && pointLineDirection < textRun->logicalLeft()) |
| return true; |
| |
| if (!textRun->nextLogicalRightwardOnLineIgnoringLineBreak()) { |
| // box is last on line |
| // and the x coordinate is to the right of the last text box right edge |
| // generate VisiblePosition, use Affinity::Upstream affinity if possible |
| shouldAffinityBeDownstream = UpstreamIfPositionIsNotAtStart; |
| return true; |
| } |
| |
| return false; |
| } |
| |
| static PositionWithAffinity createPositionWithAffinityForBox(const InlineIterator::BoxIterator& run, unsigned offset, ShouldAffinityBeDownstream shouldAffinityBeDownstream) |
| { |
| auto affinity = VisiblePosition::defaultAffinity; |
| switch (shouldAffinityBeDownstream) { |
| case AlwaysDownstream: |
| affinity = Affinity::Downstream; |
| break; |
| case AlwaysUpstream: |
| affinity = Affinity::Upstream; |
| break; |
| case UpstreamIfPositionIsNotAtStart: |
| affinity = offset > run->minimumCaretOffset() ? Affinity::Upstream : Affinity::Downstream; |
| break; |
| } |
| return run->renderer().createPositionWithAffinity(offset, affinity); |
| } |
| |
| static PositionWithAffinity createPositionWithAffinityAfterAdjustingOffsetForBiDi(const InlineIterator::TextBoxIterator& run, unsigned offset, ShouldAffinityBeDownstream shouldAffinityBeDownstream) |
| { |
| if (offset && offset < run->length()) |
| return createPositionWithAffinityForBox(run, run->start() + offset, shouldAffinityBeDownstream); |
| |
| bool positionIsAtStartOfBox = !offset; |
| if (positionIsAtStartOfBox == run->isLeftToRightDirection()) { |
| // offset is on the left edge |
| |
| auto previousRun = run->nextLineLeftwardOnLineIgnoringLineBreak(); |
| if ((previousRun && previousRun->bidiLevel() == run->bidiLevel()) |
| || run->renderer().containingBlock()->writingMode().bidiDirection() == run->direction()) // FIXME: left on 12CBA |
| return createPositionWithAffinityForBox(run, run->leftmostCaretOffset(), shouldAffinityBeDownstream); |
| |
| if (previousRun && previousRun->bidiLevel() > run->bidiLevel()) { |
| // e.g. left of B in aDC12BAb |
| auto leftmostRun = previousRun; |
| for (; previousRun; previousRun.traverseLineLeftwardOnLineIgnoringLineBreak()) { |
| if (previousRun->bidiLevel() <= run->bidiLevel()) |
| break; |
| leftmostRun = previousRun; |
| } |
| return createPositionWithAffinityForBox(leftmostRun, leftmostRun->rightmostCaretOffset(), shouldAffinityBeDownstream); |
| } |
| |
| if (!previousRun || previousRun->bidiLevel() < run->bidiLevel()) { |
| // e.g. left of D in aDC12BAb |
| InlineIterator::BoxIterator rightmostRun = run; |
| for (auto nextRun = run->nextLineRightwardOnLineIgnoringLineBreak(); nextRun; nextRun.traverseLineRightwardOnLineIgnoringLineBreak()) { |
| if (nextRun->bidiLevel() < run->bidiLevel()) |
| break; |
| rightmostRun = nextRun; |
| } |
| return createPositionWithAffinityForBox(rightmostRun, |
| run->isLeftToRightDirection() ? rightmostRun->maximumCaretOffset() : rightmostRun->minimumCaretOffset(), shouldAffinityBeDownstream); |
| } |
| |
| return createPositionWithAffinityForBox(run, run->rightmostCaretOffset(), shouldAffinityBeDownstream); |
| } |
| |
| auto nextRun = run->nextLineRightwardOnLineIgnoringLineBreak(); |
| if ((nextRun && nextRun->bidiLevel() == run->bidiLevel()) |
| || run->renderer().containingBlock()->writingMode().bidiDirection() == run->direction()) |
| return createPositionWithAffinityForBox(run, run->rightmostCaretOffset(), shouldAffinityBeDownstream); |
| |
| // offset is on the right edge |
| if (nextRun && nextRun->bidiLevel() > run->bidiLevel()) { |
| // e.g. right of C in aDC12BAb |
| auto rightmostRun = nextRun; |
| for (; nextRun; nextRun.traverseLineRightwardOnLineIgnoringLineBreak()) { |
| if (nextRun->bidiLevel() <= run->bidiLevel()) |
| break; |
| rightmostRun = nextRun; |
| } |
| |
| return createPositionWithAffinityForBox(rightmostRun, rightmostRun->leftmostCaretOffset(), shouldAffinityBeDownstream); |
| } |
| |
| if (!nextRun || nextRun->bidiLevel() < run->bidiLevel()) { |
| // e.g. right of A in aDC12BAb |
| InlineIterator::BoxIterator leftmostRun = run; |
| for (auto previousRun = run->nextLineLeftwardOnLineIgnoringLineBreak(); previousRun; previousRun.traverseLineLeftwardOnLineIgnoringLineBreak()) { |
| if (previousRun->bidiLevel() < run->bidiLevel()) |
| break; |
| leftmostRun = previousRun; |
| } |
| |
| return createPositionWithAffinityForBox(leftmostRun, |
| run->isLeftToRightDirection() ? leftmostRun->minimumCaretOffset() : leftmostRun->maximumCaretOffset(), shouldAffinityBeDownstream); |
| } |
| |
| return createPositionWithAffinityForBox(run, run->leftmostCaretOffset(), shouldAffinityBeDownstream); |
| } |
| |
| |
| PositionWithAffinity RenderText::positionForPoint(const LayoutPoint& point, HitTestSource, const RenderFragmentContainer*) |
| { |
| auto firstRun = InlineIterator::lineLeftmostTextBoxFor(*this); |
| |
| if (!firstRun || !text().length()) |
| return createPositionWithAffinity(0, Affinity::Downstream); |
| |
| auto logicalPoint = firstRun->isHorizontal() |
| ? LayoutPoint { point.x(), point.y() } |
| : LayoutPoint { point.y(), point.x() }; |
| bool blocksAreFlipped = writingMode().isBlockFlipped(); |
| |
| InlineIterator::TextBoxIterator lastRun; |
| for (auto run = firstRun; run; run.traverseNextTextBox()) { |
| if (run->isLineBreak() && !run->nextLineLeftwardOnLine() && run->nextLineRightwardOnLine() && !run->nextLineRightwardOnLine()->isLineBreak()) |
| run.traverseNextTextBox(); |
| |
| auto lineBox = run->lineBox(); |
| auto top = LayoutUnit { std::min(previousLineBoxContentBottomOrBorderAndPadding(*lineBox), lineBox->contentLogicalTop()) }; |
| if (logicalPoint.y() > top || (!blocksAreFlipped && logicalPoint.y() == top)) { |
| auto bottom = LayoutUnit { LineSelection::logicalBottom(*lineBox) }; |
| if (auto nextLineBox = lineBox->next()) |
| bottom = std::min(bottom, LayoutUnit { nextLineBox->contentLogicalTop() }); |
| |
| if (logicalPoint.y() < bottom || (blocksAreFlipped && logicalPoint.y() == bottom)) { |
| ShouldAffinityBeDownstream shouldAffinityBeDownstream; |
| #if PLATFORM(IOS_FAMILY) |
| if (logicalPoint.x() != run->logicalLeft() && logicalPoint.x() < run->logicalLeft() + run->logicalWidth()) { |
| auto half = LayoutUnit { run->logicalLeft() + run->logicalWidth() / 2.f }; |
| shouldAffinityBeDownstream = logicalPoint.x() < half ? AlwaysDownstream : AlwaysUpstream; |
| return createPositionWithAffinityAfterAdjustingOffsetForBiDi(run, offsetForPositionInRun(*run, logicalPoint.x()), shouldAffinityBeDownstream); |
| } |
| #endif |
| if (lineDirectionPointFitsInBox(logicalPoint.x(), run, shouldAffinityBeDownstream)) |
| return createPositionWithAffinityAfterAdjustingOffsetForBiDi(run, offsetForPositionInRun(*run, logicalPoint.x()), shouldAffinityBeDownstream); |
| } |
| } |
| lastRun = run; |
| } |
| |
| if (lastRun) { |
| ShouldAffinityBeDownstream shouldAffinityBeDownstream; |
| lineDirectionPointFitsInBox(logicalPoint.x(), lastRun, shouldAffinityBeDownstream); |
| return createPositionWithAffinityAfterAdjustingOffsetForBiDi(lastRun, offsetForPositionInRun(*lastRun, logicalPoint.x()) + lastRun->start(), shouldAffinityBeDownstream); |
| } |
| return createPositionWithAffinity(0, Affinity::Downstream); |
| } |
| |
| static inline std::optional<float> combineTextWidth(const RenderText& renderer, const FontCascade& fontCascade, const RenderStyle& style) |
| { |
| if (!style.hasTextCombine()) |
| return { }; |
| auto* combineTextRenderer = dynamicDowncast<RenderCombineText>(renderer); |
| if (!combineTextRenderer) |
| return { }; |
| return combineTextRenderer->isCombined() ? std::make_optional(combineTextRenderer->combinedTextWidth(fontCascade)) : std::nullopt; |
| } |
| |
| ALWAYS_INLINE float RenderText::widthFromCache(const FontCascade& fontCascade, unsigned start, unsigned length, float xPos, SingleThreadWeakHashSet<const Font>* fallbackFonts, GlyphOverflow* glyphOverflow, const RenderStyle& style) const |
| { |
| if (auto width = combineTextWidth(*this, fontCascade, style)) |
| return *width; |
| |
| TextRun run = RenderBlock::constructTextRun(*this, start, length, style); |
| run.setCharacterScanForCodePath(!canUseSimpleFontCodePath()); |
| run.setTabSize(!style.collapseWhiteSpace(), Style::toPlatform(style.tabSize())); |
| run.setXPos(xPos); |
| return fontCascade.width(run, fallbackFonts, glyphOverflow); |
| } |
| |
| ALWAYS_INLINE float RenderText::widthFromCacheConsideringPossibleTrailingSpace(const RenderStyle& style, const FontCascade& font, unsigned startIndex, unsigned wordLen, float xPos, bool currentCharacterIsSpace, WordTrailingSpace& wordTrailingSpace, SingleThreadWeakHashSet<const Font>& fallbackFonts, GlyphOverflow& glyphOverflow) const |
| { |
| return measureTextConsideringPossibleTrailingSpace(currentCharacterIsSpace, startIndex, wordLen, wordTrailingSpace, fallbackFonts, [&] (unsigned from, unsigned len) { |
| return widthFromCache(font, from, len, xPos, &fallbackFonts, &glyphOverflow, style); |
| }); |
| } |
| |
| inline bool isHangablePunctuationAtLineStart(char16_t c) |
| { |
| return U_GET_GC_MASK(c) & (U_GC_PS_MASK | U_GC_PI_MASK | U_GC_PF_MASK); |
| } |
| |
| inline bool isHangablePunctuationAtLineEnd(char16_t c) |
| { |
| return U_GET_GC_MASK(c) & (U_GC_PE_MASK | U_GC_PI_MASK | U_GC_PF_MASK); |
| } |
| |
| float RenderText::hangablePunctuationStartWidth(unsigned index) const |
| { |
| unsigned length = text().length(); |
| if (index >= length) |
| return 0; |
| |
| if (!isHangablePunctuationAtLineStart(text()[index])) |
| return 0; |
| |
| auto& style = this->style(); |
| return widthFromCache(style.fontCascade(), index, 1, 0, 0, 0, style); |
| } |
| |
| float RenderText::hangablePunctuationEndWidth(unsigned index) const |
| { |
| unsigned length = text().length(); |
| if (index >= length) |
| return 0; |
| |
| if (!isHangablePunctuationAtLineEnd(text()[index])) |
| return 0; |
| |
| auto& style = this->style(); |
| return widthFromCache(style.fontCascade(), index, 1, 0, 0, 0, style); |
| } |
| |
| bool RenderText::isHangableStopOrComma(char16_t c) |
| { |
| return c == 0x002C || c == 0x002E || c == 0x060C || c == 0x06D4 || c == 0x3001 |
| || c == 0x3002 || c == 0xFF0C || c == 0xFF0E || c == 0xFE50 || c == 0xFE51 |
| || c == 0xFE52 || c == 0xFF61 || c == 0xFF64; |
| } |
| |
| unsigned RenderText::firstCharacterIndexStrippingSpaces() const |
| { |
| if (!style().collapseWhiteSpace()) |
| return 0; |
| |
| unsigned i = 0; |
| for (unsigned length = text().length() ; i < length; ++i) { |
| if (text()[i] != ' ' && (text()[i] != '\n' || style().preserveNewline()) && text()[i] != '\t') |
| break; |
| } |
| return i; |
| } |
| |
| unsigned RenderText::lastCharacterIndexStrippingSpaces() const |
| { |
| if (!text().length()) |
| return 0; |
| |
| if (!style().collapseWhiteSpace()) |
| return text().length() - 1; |
| |
| int i = text().length() - 1; |
| for ( ; i >= 0; --i) { |
| if (text()[i] != ' ' && (text()[i] != '\n' || style().preserveNewline()) && text()[i] != '\t') |
| break; |
| } |
| return i; |
| } |
| |
| RenderText::Widths RenderText::trimmedPreferredWidths(float leadWidth, bool& stripFrontSpaces) |
| { |
| auto& style = this->style(); |
| bool collapseWhiteSpace = style.collapseWhiteSpace(); |
| |
| if (!collapseWhiteSpace) |
| stripFrontSpaces = false; |
| |
| if (m_hasTab || needsPreferredLogicalWidthsUpdate() || !m_minWidth || !m_maxWidth) |
| computePreferredLogicalWidths(leadWidth, !m_minWidth || !m_maxWidth); |
| |
| Widths widths; |
| |
| widths.beginWS = !stripFrontSpaces && m_hasBeginWS; |
| widths.endWS = m_hasEndWS; |
| |
| unsigned length = this->length(); |
| |
| if (!length || (stripFrontSpaces && text().containsOnly<isASCIIWhitespace>())) |
| return widths; |
| |
| widths.endZeroSpace = text()[length - 1] == zeroWidthSpace; |
| |
| widths.min = m_minWidth.value_or(-1); |
| widths.max = m_maxWidth.value_or(-1); |
| |
| widths.beginMin = m_beginMinWidth; |
| widths.endMin = m_endMinWidth; |
| |
| widths.hasBreakableChar = m_hasBreakableChar; |
| widths.hasBreak = m_hasBreak; |
| widths.endsWithBreak = m_hasBreak && text()[length - 1] == '\n'; |
| |
| if (text()[0] == ' ' || (text()[0] == '\n' && !style.preserveNewline()) || text()[0] == '\t') { |
| auto& font = style.fontCascade(); // FIXME: This ignores first-line. |
| if (stripFrontSpaces) |
| widths.max -= font.width(RenderBlock::constructTextRun(span(space), style)); |
| else |
| widths.max += font.wordSpacing(); |
| } |
| |
| stripFrontSpaces = collapseWhiteSpace && m_hasEndWS; |
| |
| if (!style.autoWrap() || widths.min > widths.max) |
| widths.min = widths.max; |
| |
| // Compute our max widths by scanning the string for newlines. |
| if (widths.hasBreak) { |
| auto& font = style.fontCascade(); // FIXME: This ignores first-line. |
| bool firstLine = true; |
| widths.beginMax = widths.max; |
| widths.endMax = widths.max; |
| for (unsigned i = 0; i < length; i++) { |
| unsigned lineLength = 0; |
| while (i + lineLength < length && text()[i + lineLength] != '\n') |
| lineLength++; |
| |
| if (lineLength) { |
| widths.endMax = widthFromCache(font, i, lineLength, leadWidth + widths.endMax, 0, 0, style); |
| if (firstLine) { |
| firstLine = false; |
| leadWidth = 0; |
| widths.beginMax = widths.endMax; |
| } |
| i += lineLength; |
| } else if (firstLine) { |
| widths.beginMax = 0; |
| firstLine = false; |
| leadWidth = 0; |
| } |
| |
| if (i == length - 1) { |
| // A <pre> run that ends with a newline, as in, e.g., |
| // <pre>Some text\n\n<span>More text</pre> |
| widths.endMax = 0; |
| } |
| } |
| } |
| |
| return widths; |
| } |
| |
| static inline bool isSpaceAccordingToStyle(char16_t c, const RenderStyle& style) |
| { |
| return c == ' ' || (c == noBreakSpace && style.nbspMode() == NBSPMode::Space); |
| } |
| |
| float RenderText::minLogicalWidth() const |
| { |
| if (needsPreferredLogicalWidthsUpdate() || !m_minWidth) |
| const_cast<RenderText*>(this)->computePreferredLogicalWidths(0, !needsPreferredLogicalWidthsUpdate()); |
| |
| return *m_minWidth; |
| } |
| |
| float RenderText::maxLogicalWidth() const |
| { |
| if (needsPreferredLogicalWidthsUpdate() || !m_maxWidth) |
| const_cast<RenderText*>(this)->computePreferredLogicalWidths(0, !needsPreferredLogicalWidthsUpdate()); |
| |
| return *m_maxWidth; |
| } |
| |
| TextBreakIterator::LineMode::Behavior mapLineBreakToIteratorMode(LineBreak lineBreak) |
| { |
| switch (lineBreak) { |
| case LineBreak::Auto: |
| case LineBreak::AfterWhiteSpace: |
| case LineBreak::Anywhere: |
| return TextBreakIterator::LineMode::Behavior::Default; |
| case LineBreak::Loose: |
| return TextBreakIterator::LineMode::Behavior::Loose; |
| case LineBreak::Normal: |
| return TextBreakIterator::LineMode::Behavior::Normal; |
| case LineBreak::Strict: |
| return TextBreakIterator::LineMode::Behavior::Strict; |
| } |
| ASSERT_NOT_REACHED(); |
| return TextBreakIterator::LineMode::Behavior::Default; |
| } |
| |
| TextBreakIterator::ContentAnalysis mapWordBreakToContentAnalysis(WordBreak wordBreak) |
| { |
| switch (wordBreak) { |
| case WordBreak::Normal: |
| case WordBreak::BreakAll: |
| case WordBreak::KeepAll: |
| case WordBreak::BreakWord: |
| return TextBreakIterator::ContentAnalysis::Mechanical; |
| case WordBreak::AutoPhrase: |
| return TextBreakIterator::ContentAnalysis::Linguistic; |
| } |
| return TextBreakIterator::ContentAnalysis::Mechanical; |
| } |
| |
| void RenderText::computePreferredLogicalWidths(float leadWidth, bool forcedMinMaxWidthComputation) |
| { |
| SingleThreadWeakHashSet<const Font> fallbackFonts; |
| GlyphOverflow glyphOverflow; |
| computePreferredLogicalWidths(leadWidth, fallbackFonts, glyphOverflow, forcedMinMaxWidthComputation); |
| if (fallbackFonts.isEmptyIgnoringNullReferences() && !glyphOverflow.left && !glyphOverflow.right && !glyphOverflow.top && !glyphOverflow.bottom) |
| m_knownToHaveNoOverflowAndNoFallbackFonts = true; |
| } |
| |
| static inline float hyphenWidth(RenderText& renderer, const FontCascade& font) |
| { |
| const RenderStyle& style = renderer.style(); |
| auto textRun = RenderBlock::constructTextRun(style.hyphenString().string(), style); |
| return font.width(textRun); |
| } |
| |
| float RenderText::maxWordFragmentWidth(const RenderStyle& style, const FontCascade& font, StringView word, unsigned minimumPrefixLength, unsigned minimumSuffixLength, bool currentCharacterIsSpace, unsigned characterIndex, float xPos, float entireWordWidth, WordTrailingSpace& wordTrailingSpace, SingleThreadWeakHashSet<const Font>& fallbackFonts, GlyphOverflow& glyphOverflow) |
| { |
| unsigned suffixStart = 0; |
| if (word.length() <= minimumSuffixLength) |
| return entireWordWidth; |
| |
| Vector<int, 8> hyphenLocations; |
| ASSERT(word.length() >= minimumSuffixLength); |
| unsigned hyphenLocation = word.length() - minimumSuffixLength; |
| while ((hyphenLocation = lastHyphenLocation(word, hyphenLocation, Style::toPlatform(style.computedLocale()))) >= std::max(minimumPrefixLength, 1U)) |
| hyphenLocations.append(hyphenLocation); |
| |
| if (hyphenLocations.isEmpty()) |
| return entireWordWidth; |
| |
| hyphenLocations.reverse(); |
| |
| // Consider the word "ABC-DEF-GHI" (where the '-' characters are hyphenation opportunities). We want to measure the width |
| // of "ABC-" and "DEF-", but not "GHI-". Instead, we should measure "GHI" the same way we measure regular unhyphenated |
| // words (by using wordTrailingSpace). Therefore, this function is split up into two parts - one that measures each prefix, |
| // and one that measures the single last suffix. |
| |
| // FIXME: Breaking the string at these places in the middle of words doesn't work with complex text. |
| float minimumFragmentWidthToConsider = font.size() * 5 / 4 + hyphenWidth(*this, font); |
| float maxFragmentWidth = 0; |
| for (size_t k = 0; k < hyphenLocations.size(); ++k) { |
| int fragmentLength = hyphenLocations[k] - suffixStart; |
| StringBuilder fragmentWithHyphen; |
| fragmentWithHyphen.append(word.substring(suffixStart, fragmentLength)); |
| fragmentWithHyphen.append(style.hyphenString()); |
| |
| TextRun run = RenderBlock::constructTextRun(fragmentWithHyphen.toString(), style); |
| run.setCharacterScanForCodePath(!canUseSimpleFontCodePath()); |
| float fragmentWidth = font.width(run, &fallbackFonts, &glyphOverflow); |
| |
| // Narrow prefixes are ignored. See tryHyphenating in RenderBlockLineLayout.cpp. |
| if (fragmentWidth <= minimumFragmentWidthToConsider) |
| continue; |
| |
| suffixStart += fragmentLength; |
| maxFragmentWidth = std::max(maxFragmentWidth, fragmentWidth); |
| } |
| |
| if (!suffixStart) { |
| // We didn't find any hyphenation opportunities that we're willing to actually use. |
| // Therefore, the width of the maximum fragment is just ... the width of the entire word. |
| return entireWordWidth; |
| } |
| |
| auto suffixWidth = widthFromCacheConsideringPossibleTrailingSpace(style, font, characterIndex + suffixStart, word.length() - suffixStart, xPos, currentCharacterIsSpace, wordTrailingSpace, fallbackFonts, glyphOverflow); |
| return std::max(maxFragmentWidth, suffixWidth); |
| } |
| |
| void RenderText::computePreferredLogicalWidths(float leadWidth, SingleThreadWeakHashSet<const Font>& fallbackFonts, GlyphOverflow& glyphOverflow, bool forcedMinMaxWidthComputation) |
| { |
| ASSERT_UNUSED(forcedMinMaxWidthComputation, m_hasTab || needsPreferredLogicalWidthsUpdate() || forcedMinMaxWidthComputation || !m_knownToHaveNoOverflowAndNoFallbackFonts); |
| |
| m_minWidth = 0; |
| m_beginMinWidth = 0; |
| m_endMinWidth = 0; |
| m_maxWidth = 0; |
| |
| float currMaxWidth = 0; |
| m_hasBreakableChar = false; |
| m_hasBreak = false; |
| m_hasTab = false; |
| m_hasBeginWS = false; |
| m_hasEndWS = false; |
| |
| auto& style = this->style(); |
| auto& font = style.fontCascade(); // FIXME: This ignores first-line. |
| float wordSpacing = font.wordSpacing(); |
| auto& string = text(); |
| unsigned length = string.length(); |
| auto iteratorMode = mapLineBreakToIteratorMode(style.lineBreak()); |
| auto contentAnalysis = mapWordBreakToContentAnalysis(style.wordBreak()); |
| CachedLineBreakIteratorFactory lineBreakIteratorFactory(string, Style::toPlatform(style.computedLocale()), iteratorMode, contentAnalysis); |
| bool needsWordSpacing = false; |
| bool ignoringSpaces = false; |
| bool isSpace = false; |
| bool firstWord = true; |
| bool firstLine = true; |
| std::optional<unsigned> nextBreakable; |
| unsigned lastWordBoundary = 0; |
| |
| WordTrailingSpace wordTrailingSpace(style); |
| // If automatic hyphenation is allowed, we keep track of the width of the widest word (or word |
| // fragment) encountered so far, and only try hyphenating words that are wider. |
| float maxWordWidth = std::numeric_limits<float>::max(); |
| unsigned minimumPrefixLength = 0; |
| unsigned minimumSuffixLength = 0; |
| if (style.hyphens() == Hyphens::Auto && canHyphenate(Style::toPlatform(style.computedLocale()))) { |
| maxWordWidth = 0; |
| |
| // Map 'hyphenate-limit-{before,after}: auto;' to 2. |
| minimumPrefixLength = style.hyphenateLimitBefore().tryValue().value_or(2).value; |
| minimumSuffixLength = style.hyphenateLimitAfter().tryValue().value_or(2).value; |
| } |
| |
| std::optional<LayoutUnit> firstGlyphLeftOverflow; |
| |
| bool breakNBSP = style.autoWrap() && style.nbspMode() == NBSPMode::Space; |
| |
| bool breakAnywhere = style.lineBreak() == LineBreak::Anywhere && style.autoWrap(); |
| // Note the deliberate omission of word-wrap/overflow-wrap's break-word value from this breakAll check. |
| // Those do not affect minimum preferred sizes. Note that break-word is a non-standard value for |
| // word-break, but we support it as though it means break-all. |
| bool breakAll = (style.wordBreak() == WordBreak::BreakAll || style.wordBreak() == WordBreak::BreakWord || style.overflowWrap() == OverflowWrap::Anywhere) && style.autoWrap(); |
| bool keepAllWords = style.wordBreak() == WordBreak::KeepAll; |
| bool canUseLineBreakShortcut = iteratorMode == TextBreakIterator::LineMode::Behavior::Default |
| && contentAnalysis == TextBreakIterator::ContentAnalysis::Mechanical; |
| |
| for (unsigned i = 0; i < length; i++) { |
| char16_t c = string[i]; |
| |
| bool previousCharacterIsSpace = isSpace; |
| |
| bool isNewline = false; |
| if (c == '\n') { |
| if (style.preserveNewline()) { |
| m_hasBreak = true; |
| isNewline = true; |
| isSpace = false; |
| } else |
| isSpace = true; |
| } else if (c == '\t') { |
| if (!style.collapseWhiteSpace()) { |
| m_hasTab = true; |
| isSpace = false; |
| } else |
| isSpace = true; |
| } else |
| isSpace = c == ' '; |
| |
| if ((isSpace || isNewline) && !i) |
| m_hasBeginWS = true; |
| if ((isSpace || isNewline) && i == length - 1) |
| m_hasEndWS = true; |
| |
| ignoringSpaces |= style.collapseWhiteSpace() && previousCharacterIsSpace && isSpace; |
| ignoringSpaces &= isSpace; |
| |
| // Ignore spaces and soft hyphens |
| if (ignoringSpaces) { |
| ASSERT(lastWordBoundary == i); |
| lastWordBoundary++; |
| continue; |
| } else if (c == softHyphen && style.hyphens() != Hyphens::None) { |
| ASSERT(i >= lastWordBoundary); |
| currMaxWidth += widthFromCache(font, lastWordBoundary, i - lastWordBoundary, leadWidth + currMaxWidth, &fallbackFonts, &glyphOverflow, style); |
| if (!firstGlyphLeftOverflow) |
| firstGlyphLeftOverflow = glyphOverflow.left; |
| lastWordBoundary = i + 1; |
| continue; |
| } |
| |
| bool hasBreak = breakAll || BreakablePositions::isBreakable(lineBreakIteratorFactory, i, nextBreakable, breakNBSP, canUseLineBreakShortcut, keepAllWords, breakAnywhere); |
| bool betweenWords = true; |
| unsigned j = i; |
| while (c != '\n' && !isSpaceAccordingToStyle(c, style) && c != '\t' && c != zeroWidthSpace && (c != softHyphen || style.hyphens() == Hyphens::None)) { |
| char16_t previousCharacter = c; |
| j++; |
| if (j == length) |
| break; |
| c = string[j]; |
| if (U_IS_LEAD(previousCharacter) && U_IS_TRAIL(c)) |
| continue; |
| if (BreakablePositions::isBreakable(lineBreakIteratorFactory, j, nextBreakable, breakNBSP, canUseLineBreakShortcut, keepAllWords, breakAnywhere) && characterAt(j - 1) != softHyphen) |
| break; |
| if (breakAll) { |
| // FIXME: This code is ultra wrong. |
| // The spec says "word-break: break-all: Any typographic letter units are treated as ID(“ideographic characters”) for the purpose of line-breaking." |
| // The spec describes how a "typographic letter unit" is a cluster, not a code point: https://drafts.csswg.org/css-text-3/#typographic-character-unit |
| betweenWords = false; |
| break; |
| } |
| } |
| |
| unsigned wordLen = j - i; |
| if (wordLen) { |
| float currMinWidth = 0; |
| bool isSpace = (j < length) && isSpaceAccordingToStyle(c, style); |
| float w = widthFromCacheConsideringPossibleTrailingSpace(style, font, i, wordLen, leadWidth + currMaxWidth, isSpace, wordTrailingSpace, fallbackFonts, glyphOverflow); |
| if (c == softHyphen && style.hyphens() != Hyphens::None) |
| currMinWidth = hyphenWidth(*this, font); |
| |
| if (w > maxWordWidth) { |
| auto maxFragmentWidth = maxWordFragmentWidth(style, font, StringView(string).substring(i, wordLen), minimumPrefixLength, minimumSuffixLength, isSpace, i, leadWidth + currMaxWidth, w, wordTrailingSpace, fallbackFonts, glyphOverflow); |
| currMinWidth += maxFragmentWidth - w; // This, when combined with "currMinWidth += w" below, has the effect of executing "currMinWidth += maxFragmentWidth" instead. |
| maxWordWidth = std::max(maxWordWidth, maxFragmentWidth); |
| } |
| |
| if (!firstGlyphLeftOverflow) |
| firstGlyphLeftOverflow = glyphOverflow.left; |
| currMinWidth += w; |
| if (betweenWords) { |
| if (lastWordBoundary == i) |
| currMaxWidth += w; |
| else { |
| ASSERT(j >= lastWordBoundary); |
| currMaxWidth += widthFromCache(font, lastWordBoundary, j - lastWordBoundary, leadWidth + currMaxWidth, &fallbackFonts, &glyphOverflow, style); |
| } |
| lastWordBoundary = j; |
| } |
| |
| bool isCollapsibleWhiteSpace = (j < length) && style.isCollapsibleWhiteSpace(c); |
| if (j < length && style.autoWrap()) |
| m_hasBreakableChar = true; |
| |
| // Add in wordSpacing to our currMaxWidth, but not if this is the last word on a line or the |
| // last word in the run. |
| if ((isSpace || isCollapsibleWhiteSpace) && !containsOnlyCSSWhitespace(j, length - j)) |
| currMaxWidth += wordSpacing; |
| |
| if (firstWord) { |
| firstWord = false; |
| // If the first character in the run is breakable, then we consider ourselves to have a beginning |
| // minimum width of 0, since a break could occur right before our run starts, preventing us from ever |
| // being appended to a previous text run when considering the total minimum width of the containing block. |
| if (hasBreak) |
| m_hasBreakableChar = true; |
| m_beginMinWidth = hasBreak ? 0 : currMinWidth; |
| } |
| m_endMinWidth = currMinWidth; |
| |
| m_minWidth = std::max(currMinWidth, *m_minWidth); |
| |
| i += wordLen - 1; |
| } else { |
| // Nowrap can never be broken, so don't bother setting the |
| // breakable character boolean. Pre can only be broken if we encounter a newline. |
| if (style.autoWrap() || isNewline) |
| m_hasBreakableChar = true; |
| |
| if (isNewline) { // Only set if preserveNewline was true and we saw a newline. |
| if (firstLine) { |
| firstLine = false; |
| leadWidth = 0; |
| if (!style.autoWrap()) |
| m_beginMinWidth = currMaxWidth; |
| } |
| |
| if (currMaxWidth > *m_maxWidth) |
| m_maxWidth = currMaxWidth; |
| currMaxWidth = 0; |
| } else { |
| TextRun run = RenderBlock::constructTextRun(*this, i, 1, style); |
| run.setTabSize(!style.collapseWhiteSpace(), Style::toPlatform(style.tabSize())); |
| run.setXPos(leadWidth + currMaxWidth); |
| |
| currMaxWidth += font.width(run, &fallbackFonts); |
| glyphOverflow.right = 0; |
| needsWordSpacing = isSpace && !previousCharacterIsSpace && i == length - 1; |
| } |
| ASSERT(lastWordBoundary == i); |
| lastWordBoundary++; |
| } |
| } |
| |
| glyphOverflow.left = firstGlyphLeftOverflow.value_or(glyphOverflow.left); |
| |
| if ((needsWordSpacing && length > 1) || (ignoringSpaces && !firstWord)) |
| currMaxWidth += wordSpacing; |
| |
| m_maxWidth = std::max(currMaxWidth, *m_maxWidth); |
| |
| if (!style.autoWrap()) |
| m_minWidth = m_maxWidth; |
| |
| if (style.whiteSpaceCollapse() == WhiteSpaceCollapse::Preserve && style.textWrapMode() == TextWrapMode::NoWrap) { |
| if (firstLine) |
| m_beginMinWidth = *m_maxWidth; |
| m_endMinWidth = currMaxWidth; |
| } |
| |
| clearNeedsPreferredWidthsUpdate(); |
| } |
| |
| template<typename CharacterType> static inline bool containsOnlyCollapsibleWhitespace(std::span<const CharacterType> characters, const RenderStyle& style) |
| { |
| for (auto character : characters) { |
| if (!style.isCollapsibleWhiteSpace(character)) |
| return false; |
| } |
| return true; |
| } |
| |
| bool RenderText::containsOnlyCollapsibleWhitespace() const |
| { |
| if (text().is8Bit()) |
| return WebCore::containsOnlyCollapsibleWhitespace(text().span8(), style()); |
| return WebCore::containsOnlyCollapsibleWhitespace(text().span16(), style()); |
| } |
| |
| // FIXME: merge this with isCSSSpace somehow |
| template<typename CharacterType> static inline bool containsOnlyPossiblyCollapsibleWhitespace(std::span<const CharacterType> characters) |
| { |
| for (auto character : characters) { |
| if (!(character == '\n' || character == ' ' || character == '\t')) |
| return false; |
| } |
| return true; |
| } |
| |
| bool RenderText::containsOnlyCSSWhitespace(unsigned from, unsigned length) const |
| { |
| ASSERT(from <= text().length()); |
| ASSERT(length <= text().length()); |
| ASSERT(from + length <= text().length()); |
| if (text().is8Bit()) |
| return containsOnlyPossiblyCollapsibleWhitespace(text().span8().subspan(from, length)); |
| return containsOnlyPossiblyCollapsibleWhitespace(text().span16().subspan(from, length)); |
| } |
| |
| Vector<std::pair<unsigned, unsigned>> RenderText::contentRangesBetweenOffsetsForType(DocumentMarkerType type, unsigned startOffset, unsigned endOffset) const |
| { |
| if (!textNode()) |
| return { }; |
| |
| CheckedPtr markerController = document().markersIfExists(); |
| if (!markerController) |
| return { }; |
| |
| auto markers = markerController->markersFor(*textNode(), type); |
| if (markers.isEmpty()) |
| return { }; |
| |
| Vector<std::pair<unsigned, unsigned>> contentRanges; |
| for (auto& marker : markers) { |
| unsigned markerStart = std::max(marker->startOffset(), startOffset); |
| unsigned markerEnd = std::min(marker->endOffset(), endOffset); |
| if (markerStart >= markerEnd || markerStart > endOffset || markerEnd < startOffset) |
| continue; |
| |
| contentRanges.append({ markerStart, markerEnd }); |
| } |
| return contentRanges; |
| } |
| |
| IntPoint RenderText::firstRunLocation() const |
| { |
| auto first = InlineIterator::lineLeftmostTextBoxFor(*this); |
| if (!first) |
| return { }; |
| return IntPoint(first->visualRectIgnoringBlockDirection().location()); |
| } |
| |
| void RenderText::setSelectionState(HighlightState state) |
| { |
| RenderObject::setSelectionState(state); |
| |
| // The containing block can be null in case of an orphaned tree. |
| RenderBlock* containingBlock = this->containingBlock(); |
| if (containingBlock && !containingBlock->isRenderView()) |
| containingBlock->setSelectionState(state); |
| } |
| |
| static inline bool isInlineFlowOrEmptyText(const RenderObject& renderer) |
| { |
| if (is<RenderInline>(renderer)) |
| return true; |
| auto* textRenderer = dynamicDowncast<RenderText>(renderer); |
| return textRenderer && textRenderer->text().isEmpty(); |
| } |
| |
| Vector<char16_t> RenderText::previousCharacter() const |
| { |
| const RenderObject* previousText = this; |
| while ((previousText = previousText->previousInPreOrder())) { |
| if (!previousText->isInFlow()) |
| continue; |
| if (!isInlineFlowOrEmptyText(*previousText)) |
| break; |
| } |
| auto* renderText = dynamicDowncast<RenderText>(previousText); |
| Vector<char16_t> previous; |
| if (!renderText) |
| previous.append(' '); |
| else { |
| auto& previousString = renderText->text(); |
| if (previousString.is8Bit()) |
| previous.append(previousString[previousString.length() - 1]); |
| else { |
| auto previousCharacterLength = [&] { |
| auto contentIterator = SurrogatePairAwareTextIterator { previousString.span16(), 0, previousString.length() }; |
| unsigned characterLength = 0; |
| char32_t currentCharacter = 0; |
| while (contentIterator.consume(currentCharacter, characterLength)) |
| contentIterator.advance(characterLength); |
| return characterLength; |
| }(); |
| |
| if (previousCharacterLength > previousString.length()) { |
| ASSERT_NOT_REACHED(); |
| return previous; |
| } |
| for (size_t i = previousString.length() - previousCharacterLength; i < previousString.length(); ++i) |
| previous.append(previousString[i]); |
| } |
| } |
| return previous; |
| } |
| |
| static String convertToFullSizeKana(const String& string) |
| { |
| // https://www.w3.org/TR/css-text-3/#small-kana |
| static constexpr SortedArrayMap sortedMap { std::to_array<std::pair<char32_t, char16_t>>({ |
| { 0x3041, 0x3042 }, |
| { 0x3043, 0x3044 }, |
| { 0x3045, 0x3046 }, |
| { 0x3047, 0x3048 }, |
| { 0x3049, 0x304A }, |
| { 0x3063, 0x3064 }, |
| { 0x3083, 0x3084 }, |
| { 0x3085, 0x3086 }, |
| { 0x3087, 0x3088 }, |
| { 0x308E, 0x308F }, |
| { 0x3095, 0x304B }, |
| { 0x3096, 0x3051 }, |
| { 0x30A1, 0x30A2 }, |
| { 0x30A3, 0x30A4 }, |
| { 0x30A5, 0x30A6 }, |
| { 0x30A7, 0x30A8 }, |
| { 0x30A9, 0x30AA }, |
| { 0x30C3, 0x30C4 }, |
| { 0x30E3, 0x30E4 }, |
| { 0x30E5, 0x30E6 }, |
| { 0x30E7, 0x30E8 }, |
| { 0x30EE, 0x30EF }, |
| { 0x30F5, 0x30AB }, |
| { 0x30F6, 0x30B1 }, |
| { 0x31F0, 0x30AF }, |
| { 0x31F1, 0x30B7 }, |
| { 0x31F2, 0x30B9 }, |
| { 0x31F3, 0x30C8 }, |
| { 0x31F4, 0x30CC }, |
| { 0x31F5, 0x30CF }, |
| { 0x31F6, 0x30D2 }, |
| { 0x31F7, 0x30D5 }, |
| { 0x31F8, 0x30D8 }, |
| { 0x31F9, 0x30DB }, |
| { 0x31FA, 0x30E0 }, |
| { 0x31FB, 0x30E9 }, |
| { 0x31FC, 0x30EA }, |
| { 0x31FD, 0x30EB }, |
| { 0x31FE, 0x30EC }, |
| { 0x31FF, 0x30ED }, |
| { 0xFF67, 0xFF71 }, |
| { 0xFF68, 0xFF72 }, |
| { 0xFF69, 0xFF73 }, |
| { 0xFF6A, 0xFF74 }, |
| { 0xFF6B, 0xFF75 }, |
| { 0xFF6C, 0xFF94 }, |
| { 0xFF6D, 0xFF95 }, |
| { 0xFF6E, 0xFF96 }, |
| { 0xFF6F, 0xFF82 }, |
| { 0x1B132, 0x3053 }, |
| { 0x1B150, 0x3090 }, |
| { 0x1B151, 0x3091 }, |
| { 0x1B152, 0x3092 }, |
| { 0x1B155, 0x30B3 }, |
| { 0x1B164, 0x30F0 }, |
| { 0x1B165, 0x30F1 }, |
| { 0x1B166, 0x30F2 }, |
| { 0x1B167, 0x30F3 } |
| }) }; |
| |
| auto codePoints = StringView { string }.codePoints(); |
| |
| bool needsConversion = false; |
| for (auto character : codePoints) { |
| if (sortedMap.contains(character)) { |
| needsConversion = true; |
| break; |
| } |
| } |
| if (!needsConversion) |
| return string; |
| |
| StringBuilder result; |
| for (auto character : codePoints) { |
| if (auto found = sortedMap.tryGet(character)) |
| result.append(*found); |
| else |
| result.append(character); |
| } |
| return result.toString(); |
| } |
| |
| // https://w3c.github.io/mathml-core/#math-auto-transform |
| static String convertToMathAuto(const String& string) |
| { |
| #if ENABLE(MATHML) |
| StringView view = string; |
| if (auto codePoint = view.convertToSingleCodePoint()) { |
| char32_t transformedCodePoint = mathVariantMapCodePoint(codePoint.value(), MathVariant::Italic); |
| if (transformedCodePoint != codePoint.value()) |
| return String::fromCodePoint(transformedCodePoint); |
| } |
| #endif |
| return string; |
| } |
| |
| String applyTextTransform(const RenderStyle& style, const String& text) |
| { |
| Vector<char16_t> previousCharacter(1, ' '); |
| return applyTextTransform(style, text, previousCharacter); |
| } |
| |
| String applyTextTransform(const RenderStyle& style, const String& text, Vector<char16_t> previousCharacter) |
| { |
| auto transform = style.textTransform(); |
| |
| if (transform.isNone()) |
| return text; |
| |
| // https://w3c.github.io/csswg-drafts/css-text/#text-transform-order |
| auto modified = text; |
| if (transform.contains(Style::TextTransformValue::Capitalize)) |
| modified = capitalize(modified, previousCharacter); // FIXME: Need to take locale into account. |
| else if (transform.contains(Style::TextTransformValue::Uppercase)) |
| modified = modified.convertToUppercaseWithLocale(Style::toPlatform(style.computedLocale())); |
| else if (transform.contains(Style::TextTransformValue::Lowercase)) |
| modified = modified.convertToLowercaseWithLocale(Style::toPlatform(style.computedLocale())); |
| |
| if (transform.contains(Style::TextTransformValue::FullWidth)) |
| modified = transformToFullWidth(modified); |
| |
| if (transform.contains(Style::TextTransformValue::FullSizeKana)) |
| modified = convertToFullSizeKana(modified); |
| |
| if (transform.contains(Style::TextTransformValue::MathAuto)) |
| modified = convertToMathAuto(modified); |
| |
| return modified; |
| } |
| |
| void RenderText::setRenderedText(const String& newText) |
| { |
| ASSERT(!newText.isNull()); |
| |
| String originalText = this->originalText(); |
| |
| m_text = newText; |
| |
| if (m_useBackslashAsYenSymbol) |
| m_text = makeStringByReplacingAll(m_text, '\\', yenSign); |
| |
| const auto& style = this->style(); |
| if (!style.textTransform().isNone()) |
| m_text = applyTextTransform(style, m_text, previousCharacter()); |
| |
| // At rendering time, if certain fonts are used, these characters get swapped out with higher-quality PUA characters. |
| // See RenderBlock::updateSecurityDiscCharacters(). |
| switch (style.textSecurity()) { |
| case TextSecurity::None: |
| break; |
| #if !PLATFORM(IOS_FAMILY) |
| // We use the same characters here as for list markers. |
| // See the listMarkerText function in RenderListMarker.cpp. |
| case TextSecurity::Circle: |
| secureText(whiteBullet); |
| break; |
| case TextSecurity::Disc: |
| secureText(bullet); |
| break; |
| case TextSecurity::Square: |
| secureText(blackSquare); |
| break; |
| #else |
| // FIXME: Why this quirk on iOS? |
| case TextSecurity::Circle: |
| case TextSecurity::Disc: |
| case TextSecurity::Square: |
| secureText(blackCircle); |
| break; |
| #endif |
| } |
| |
| m_containsOnlyASCII = text().containsOnlyASCII(); |
| computeFontCodePath(); |
| m_canUseSimplifiedTextMeasuring = { }; |
| m_hasPositionDependentContentWidth = { }; |
| m_hasStrongDirectionalityContent = { }; |
| |
| if (m_text != originalText) { |
| originalTextMap().set(*this, originalText); |
| m_originalTextDiffersFromRendered = true; |
| } else if (m_originalTextDiffersFromRendered) { |
| originalTextMap().remove(*this); |
| m_originalTextDiffersFromRendered = false; |
| } |
| } |
| |
| void RenderText::secureText(char16_t maskingCharacter) |
| { |
| // This hides the text by replacing all the characters with the masking character. |
| // Offsets within the hidden text have to match offsets within the original text |
| // to handle things like carets and selection, so this won't work right if any |
| // of the characters are surrogate pairs or combining marks. Thus, this function |
| // does not attempt to handle either of those. |
| |
| unsigned length = text().length(); |
| if (!length) |
| return; |
| |
| char16_t characterToReveal = 0; |
| unsigned revealedCharactersOffset = 0; |
| |
| if (m_hasSecureTextTimer) { |
| if (SecureTextTimer* timer = secureTextTimers().get(*this)) { |
| // We take the offset out of the timer to make this one-shot. We count on this being called only once. |
| // If it's called a second time we assume the text is different and a character should not be revealed. |
| revealedCharactersOffset = timer->takeOffsetAfterLastTypedCharacter(); |
| if (revealedCharactersOffset && revealedCharactersOffset <= length) |
| characterToReveal = text()[--revealedCharactersOffset]; |
| } |
| } |
| |
| std::span<char16_t> characters; |
| m_text = String::createUninitialized(length, characters); |
| |
| std::ranges::fill(characters, maskingCharacter); |
| if (characterToReveal) |
| characters[revealedCharactersOffset] = characterToReveal; |
| } |
| |
| static void invalidateLineLayoutPathOnContentChangeIfNeeded(RenderText& renderer, std::optional<size_t> offset, size_t oldLength) |
| { |
| auto* container = LayoutIntegration::LineLayout::blockContainer(renderer); |
| if (!container) |
| return; |
| |
| auto* inlineLayout = container->inlineLayout(); |
| if (!inlineLayout) |
| return; |
| |
| if (LayoutIntegration::LineLayout::shouldInvalidateLineLayoutAfterContentChange(*container, renderer, *inlineLayout)) { |
| container->invalidateLineLayout(RenderBlockFlow::InvalidationReason::ContentChange); |
| return; |
| } |
| if (!inlineLayout->updateTextContent(renderer, offset, oldLength)) |
| container->invalidateLineLayout(RenderBlockFlow::InvalidationReason::ContentChange); |
| } |
| |
| void RenderText::setTextInternal(const String& text, bool force) |
| { |
| ASSERT(!text.isNull()); |
| |
| if (!force && text == originalText()) |
| return; |
| |
| m_text = text; |
| if (m_originalTextDiffersFromRendered) { |
| originalTextMap().remove(this); |
| m_originalTextDiffersFromRendered = false; |
| } |
| |
| setRenderedText(text); |
| |
| setNeedsLayoutAndPreferredWidthsUpdate(); |
| m_knownToHaveNoOverflowAndNoFallbackFonts = false; |
| |
| if (AXObjectCache* cache = document().existingAXObjectCache()) |
| cache->deferTextChangedIfNeeded(textNode()); |
| } |
| |
| void RenderText::setText(const String& newContent, bool force) |
| { |
| auto isDifferent = newContent != text(); |
| setTextInternal(newContent, force); |
| if (isDifferent || force) |
| invalidateLineLayoutPathOnContentChangeIfNeeded(*this, { }, { }); |
| } |
| |
| void RenderText::setTextWithOffset(const String& newText, unsigned offset) |
| { |
| if (text() == newText) |
| return; |
| |
| size_t oldLength = text().length(); |
| setTextInternal(newText, false); |
| invalidateLineLayoutPathOnContentChangeIfNeeded(*this, offset, oldLength); |
| } |
| |
| String RenderText::textWithoutConvertingBackslashToYenSymbol() const |
| { |
| if (!m_useBackslashAsYenSymbol || style().textSecurity() != TextSecurity::None) |
| return text(); |
| |
| if (style().textTransform().isNone()) |
| return originalText(); |
| |
| return applyTextTransform(style(), originalText(), previousCharacter()); |
| } |
| |
| float RenderText::width(unsigned from, unsigned len, float xPos, bool firstLine, SingleThreadWeakHashSet<const Font>* fallbackFonts, GlyphOverflow* glyphOverflow) const |
| { |
| if (from >= text().length()) |
| return 0; |
| |
| if (from + len > text().length()) |
| len = text().length() - from; |
| |
| const RenderStyle& lineStyle = firstLine ? firstLineStyle() : style(); |
| return width(from, len, lineStyle.fontCascade(), xPos, fallbackFonts, glyphOverflow); |
| } |
| |
| float RenderText::width(unsigned from, unsigned length, const FontCascade& fontCascade, float xPos, SingleThreadWeakHashSet<const Font>* fallbackFonts, GlyphOverflow* glyphOverflow) const |
| { |
| ASSERT(from + length <= text().length()); |
| if (!text().length() || !length) |
| return 0.f; |
| |
| auto& style = this->style(); |
| if (auto width = combineTextWidth(*this, fontCascade, style)) |
| return *width; |
| |
| if (length == 1 && (characterAt(from) == space)) |
| return fontCascade.widthOfSpaceString(); |
| |
| float width = 0.f; |
| if (&fontCascade == &style.fontCascade()) { |
| if (!style.preserveNewline() && !from && length == text().length() && (!glyphOverflow || !glyphOverflow->computeBounds)) { |
| if (fallbackFonts) { |
| ASSERT(glyphOverflow); |
| if (needsPreferredLogicalWidthsUpdate() || !m_knownToHaveNoOverflowAndNoFallbackFonts) { |
| const_cast<RenderText*>(this)->computePreferredLogicalWidths(0, *fallbackFonts, *glyphOverflow); |
| if (fallbackFonts->isEmptyIgnoringNullReferences() && !glyphOverflow->left && !glyphOverflow->right && !glyphOverflow->top && !glyphOverflow->bottom) |
| m_knownToHaveNoOverflowAndNoFallbackFonts = true; |
| } |
| // The rare case of when we switch between IFC and legacy preferred width computation. |
| if (!m_maxWidth) |
| width = maxLogicalWidth(); |
| else |
| width = *m_maxWidth; |
| } else |
| width = maxLogicalWidth(); |
| } else |
| width = widthFromCache(fontCascade, from, length, xPos, fallbackFonts, glyphOverflow, style); |
| } else { |
| TextRun run = RenderBlock::constructTextRun(*this, from, length, style); |
| run.setCharacterScanForCodePath(!canUseSimpleFontCodePath()); |
| run.setTabSize(!style.collapseWhiteSpace(), Style::toPlatform(style.tabSize())); |
| run.setXPos(xPos); |
| |
| width = fontCascade.width(run, fallbackFonts, glyphOverflow); |
| } |
| |
| return clampTo(width, 0.f); |
| } |
| |
| IntRect RenderText::linesBoundingBox() const |
| { |
| auto firstTextBox = InlineIterator::lineLeftmostTextBoxFor(*this); |
| if (!firstTextBox) |
| return { }; |
| |
| auto boundingBox = firstTextBox->visualRectIgnoringBlockDirection(); |
| for (auto textBox = firstTextBox; ++textBox;) |
| boundingBox.uniteEvenIfEmpty(textBox->visualRectIgnoringBlockDirection()); |
| |
| return enclosingIntRect(boundingBox); |
| } |
| |
| auto RenderText::localRectsForRepaint(RepaintOutlineBounds) const -> RepaintRects |
| { |
| LayoutRect overflowRect = linesBoundingBox(); |
| // FIXME: layoutDelta needs to be applied in parts before/after transforms and |
| // repaint containers. https://bugs.webkit.org/show_bug.cgi?id=23308 |
| overflowRect.move(view().frameView().layoutContext().layoutDelta()); |
| |
| return RepaintRects { overflowRect }; |
| } |
| |
| LayoutRect RenderText::collectSelectionGeometriesForLineBoxes(const RenderLayerModelObject* repaintContainer, bool clipToVisibleContent, Vector<FloatQuad>* quads) |
| { |
| ASSERT(!needsLayout()); |
| |
| if (selectionState() == HighlightState::None) |
| return LayoutRect(); |
| if (!containingBlock()) |
| return LayoutRect(); |
| |
| // Now calculate startPos and endPos for painting selection. |
| // We include a selection while endPos > 0 |
| unsigned startOffset; |
| unsigned endOffset; |
| if (selectionState() == HighlightState::Inside) { |
| // We are fully selected. |
| startOffset = 0; |
| endOffset = text().length(); |
| } else { |
| startOffset = view().selection().startOffset(); |
| endOffset = view().selection().endOffset(); |
| if (selectionState() == HighlightState::Start) |
| endOffset = text().length(); |
| else if (selectionState() == HighlightState::End) |
| startOffset = 0; |
| } |
| |
| if (startOffset == endOffset) |
| return IntRect(); |
| |
| LayoutRect resultRect; |
| |
| for (auto& textBox : InlineIterator::textBoxesFor(*this)) { |
| LayoutRect rect; |
| rect.unite(selectionRectForTextBox(textBox, startOffset, endOffset)); |
| rect.unite(ellipsisRectForTextBox(textBox, startOffset, endOffset).value_or(IntRect { })); |
| if (rect.isEmpty()) |
| continue; |
| |
| resultRect.unite(rect); |
| |
| if (quads) |
| quads->append(localToContainerQuad(FloatRect(rect), repaintContainer)); |
| } |
| |
| if (clipToVisibleContent) |
| return computeRectForRepaint(resultRect, repaintContainer); |
| return localToContainerQuad(FloatRect(resultRect), repaintContainer).enclosingBoundingBox(); |
| } |
| |
| LayoutRect RenderText::collectSelectionGeometriesForLineBoxes(const RenderLayerModelObject* repaintContainer, bool clipToVisibleContent, Vector<FloatQuad>& quads) |
| { |
| return collectSelectionGeometriesForLineBoxes(repaintContainer, clipToVisibleContent, &quads); |
| } |
| |
| LayoutRect RenderText::selectionRectForRepaint(const RenderLayerModelObject* repaintContainer, bool clipToVisibleContent) |
| { |
| return collectSelectionGeometriesForLineBoxes(repaintContainer, clipToVisibleContent, nullptr); |
| } |
| |
| int RenderText::caretMinOffset() const |
| { |
| auto first = InlineIterator::lineLeftmostTextBoxFor(*this); |
| if (!first) |
| return 0; |
| |
| int minOffset = first->start(); |
| for (auto box = first; ++box;) |
| minOffset = std::min<int>(minOffset, box->start()); |
| |
| return minOffset; |
| } |
| |
| int RenderText::caretMaxOffset() const |
| { |
| auto first = InlineIterator::lineLeftmostTextBoxFor(*this); |
| if (!first) |
| return text().length(); |
| |
| int maxOffset = first->end(); |
| for (auto box = first; ++box;) |
| maxOffset = std::max<int>(maxOffset, box->end()); |
| |
| return maxOffset; |
| } |
| |
| unsigned RenderText::countRenderedCharacterOffsetsUntil(unsigned offset) const |
| { |
| unsigned result = 0; |
| for (auto& run : InlineIterator::textBoxesFor(*this)) { |
| auto start = run.start(); |
| auto length = run.length(); |
| if (offset < start) |
| return result; |
| if (offset <= start + length) { |
| result += offset - start; |
| return result; |
| } |
| result += length; |
| } |
| return result; |
| } |
| |
| enum class OffsetType { Character, Caret }; |
| static bool containsOffset(const RenderText& text, unsigned offset, OffsetType type) |
| { |
| for (auto [box, orderCache] = InlineIterator::firstTextBoxInLogicalOrderFor(text); box; box = InlineIterator::nextTextBoxInLogicalOrder(box, orderCache)) { |
| auto start = box->start(); |
| if (offset < start) |
| return false; |
| unsigned end = box->end(); |
| if (offset >= start && offset <= end) { |
| if (offset == end && (type == OffsetType::Character || box->isLineBreak())) |
| continue; |
| if (type == OffsetType::Character) |
| return true; |
| // Return false for offsets inside composed characters. |
| return !offset || offset == static_cast<unsigned>(text.nextOffset(text.previousOffset(offset))); |
| } |
| } |
| return false; |
| } |
| |
| bool RenderText::containsRenderedCharacterOffset(unsigned offset) const |
| { |
| return containsOffset(*this, offset, OffsetType::Character); |
| } |
| |
| bool RenderText::containsCaretOffset(unsigned offset) const |
| { |
| return containsOffset(*this, offset, OffsetType::Caret); |
| } |
| |
| bool RenderText::hasRenderedText() const |
| { |
| for (auto& box : InlineIterator::textBoxesFor(*this)) { |
| if (box.length()) |
| return true; |
| } |
| return false; |
| } |
| |
| int RenderText::previousOffset(int current) const |
| { |
| if (m_containsOnlyASCII || text().is8Bit()) |
| return current - 1; |
| |
| CachedTextBreakIterator iterator(text(), { }, TextBreakIterator::CaretMode { }, nullAtom()); |
| return iterator.preceding(current).value_or(current - 1); |
| } |
| |
| int RenderText::previousOffsetForBackwardDeletion(int current) const |
| { |
| CachedTextBreakIterator iterator(text(), { }, TextBreakIterator::DeleteMode { }, nullAtom()); |
| return iterator.preceding(current).value_or(0); |
| } |
| |
| int RenderText::nextOffset(int current) const |
| { |
| if (m_containsOnlyASCII || text().is8Bit()) |
| return current + 1; |
| |
| CachedTextBreakIterator iterator(text(), { }, TextBreakIterator::CaretMode { }, nullAtom()); |
| return iterator.following(current).value_or(current + 1); |
| } |
| |
| void RenderText::computeFontCodePath() |
| { |
| if (m_containsOnlyASCII || text().is8Bit()) { |
| m_fontCodePath = static_cast<unsigned>(FontCascade::CodePath::Simple); |
| return; |
| } |
| m_fontCodePath = static_cast<unsigned>(FontCascade::characterRangeCodePath(text().span16())); |
| } |
| |
| void RenderText::momentarilyRevealLastTypedCharacter(unsigned offsetAfterLastTypedCharacter) |
| { |
| if (style().textSecurity() == TextSecurity::None) |
| return; |
| m_hasSecureTextTimer = true; |
| auto& secureTextTimer = secureTextTimers().add(*this, nullptr).iterator->value; |
| if (!secureTextTimer) |
| secureTextTimer = makeUnique<SecureTextTimer>(*this); |
| secureTextTimer->restart(offsetAfterLastTypedCharacter); |
| } |
| |
| StringView RenderText::stringView(unsigned start, std::optional<unsigned> stop) const |
| { |
| unsigned destination = stop.value_or(text().length()); |
| ASSERT(start <= length()); |
| ASSERT(destination <= length()); |
| ASSERT(start <= destination); |
| return StringView { text() }.substring(start, destination - start); |
| } |
| |
| RenderInline* RenderText::inlineWrapperForDisplayContents() |
| { |
| ASSERT(m_hasInlineWrapperForDisplayContents == inlineWrapperForDisplayContentsMap().contains(this)); |
| |
| if (!m_hasInlineWrapperForDisplayContents) |
| return nullptr; |
| return inlineWrapperForDisplayContentsMap().get(this); |
| } |
| |
| void RenderText::setInlineWrapperForDisplayContents(RenderInline* wrapper) |
| { |
| ASSERT(m_hasInlineWrapperForDisplayContents == inlineWrapperForDisplayContentsMap().contains(this)); |
| |
| if (!wrapper) { |
| if (!m_hasInlineWrapperForDisplayContents) |
| return; |
| inlineWrapperForDisplayContentsMap().remove(*this); |
| m_hasInlineWrapperForDisplayContents = false; |
| return; |
| } |
| inlineWrapperForDisplayContentsMap().add(*this, wrapper); |
| m_hasInlineWrapperForDisplayContents = true; |
| } |
| |
| } // namespace WebCore |