/* -*- Mode: C++; tab-width: 8; indent-tabs-mode: nil; c-basic-offset: 2 -*- */ /* vim: set ts=8 sts=2 et sw=2 tw=80: */ /* This Source Code Form is subject to the terms of the Mozilla Public * License, v. 2.0. If a copy of the MPL was not distributed with this * file, You can obtain one at http://mozilla.org/MPL/2.0/. */ #include "nsMathMLmunderoverFrame.h" #include "nsLayoutUtils.h" #include "nsPresContext.h" #include "nsMathMLmmultiscriptsFrame.h" #include "mozilla/dom/Document.h" #include "mozilla/dom/MathMLElement.h" #include #include "gfxContext.h" #include "gfxMathTable.h" #include "gfxTextRun.h" #include "mozilla/PresShell.h" #include "mozilla/StaticPrefs_mathml.h" using namespace mozilla; // // -- attach an underscript-overscript pair to a base // implementation // -- attach an overscript to a base - implementation // -- attach an underscript to a base - implementation // nsIFrame* NS_NewMathMLmunderoverFrame(PresShell* aPresShell, ComputedStyle* aStyle) { return new (aPresShell) nsMathMLmunderoverFrame(aStyle, aPresShell->GetPresContext()); } NS_IMPL_FRAMEARENA_HELPERS(nsMathMLmunderoverFrame) nsMathMLmunderoverFrame::~nsMathMLmunderoverFrame() = default; nsresult nsMathMLmunderoverFrame::AttributeChanged(int32_t aNameSpaceID, nsAtom* aAttribute, int32_t aModType) { if (nsGkAtoms::accent_ == aAttribute || nsGkAtoms::accentunder_ == aAttribute) { // When we have automatic data to update within ourselves, we ask our // parent to re-layout its children return ReLayoutChildren(GetParent()); } return nsMathMLContainerFrame::AttributeChanged(aNameSpaceID, aAttribute, aModType); } NS_IMETHODIMP nsMathMLmunderoverFrame::UpdatePresentationData(uint32_t aFlagsValues, uint32_t aFlagsToUpdate) { nsMathMLContainerFrame::UpdatePresentationData(aFlagsValues, aFlagsToUpdate); // disable the stretch-all flag if we are going to act like a // subscript-superscript pair if (NS_MATHML_EMBELLISH_IS_MOVABLELIMITS(mEmbellishData.flags) && StyleFont()->mMathStyle == StyleMathStyle::Compact) { mPresentationData.flags &= ~NS_MATHML_STRETCH_ALL_CHILDREN_HORIZONTALLY; } else { mPresentationData.flags |= NS_MATHML_STRETCH_ALL_CHILDREN_HORIZONTALLY; } return NS_OK; } NS_IMETHODIMP nsMathMLmunderoverFrame::InheritAutomaticData(nsIFrame* aParent) { // let the base class get the default from our parent nsMathMLContainerFrame::InheritAutomaticData(aParent); mPresentationData.flags |= NS_MATHML_STRETCH_ALL_CHILDREN_HORIZONTALLY; return NS_OK; } void nsMathMLmunderoverFrame::Destroy(DestroyContext& aContext) { if (!mPostReflowIncrementScriptLevelCommands.IsEmpty()) { PresShell()->CancelReflowCallback(this); } nsMathMLContainerFrame::Destroy(aContext); } uint8_t nsMathMLmunderoverFrame::ScriptIncrement(nsIFrame* aFrame) { nsIFrame* child = mFrames.FirstChild(); if (!aFrame || aFrame == child) { return 0; } child = child->GetNextSibling(); if (aFrame == child) { if (mContent->IsMathMLElement(nsGkAtoms::mover_)) { return mIncrementOver ? 1 : 0; } return mIncrementUnder ? 1 : 0; } if (child && aFrame == child->GetNextSibling()) { // must be a over frame of munderover return mIncrementOver ? 1 : 0; } return 0; // frame not found } void nsMathMLmunderoverFrame::SetIncrementScriptLevel(uint32_t aChildIndex, bool aIncrement) { nsIFrame* child = PrincipalChildList().FrameAt(aChildIndex); if (!child || !child->GetContent()->IsMathMLElement() || child->GetContent()->GetPrimaryFrame() != child) { return; } auto element = dom::MathMLElement::FromNode(child->GetContent()); if (element->GetIncrementScriptLevel() == aIncrement) { return; } if (mPostReflowIncrementScriptLevelCommands.IsEmpty()) { PresShell()->PostReflowCallback(this); } mPostReflowIncrementScriptLevelCommands.AppendElement( SetIncrementScriptLevelCommand{aChildIndex, aIncrement}); } bool nsMathMLmunderoverFrame::ReflowFinished() { SetPendingPostReflowIncrementScriptLevel(); return true; } void nsMathMLmunderoverFrame::ReflowCallbackCanceled() { // Do nothing, at this point our work will just be useless. mPostReflowIncrementScriptLevelCommands.Clear(); } void nsMathMLmunderoverFrame::SetPendingPostReflowIncrementScriptLevel() { MOZ_ASSERT(!mPostReflowIncrementScriptLevelCommands.IsEmpty()); nsTArray commands = std::move(mPostReflowIncrementScriptLevelCommands); for (const auto& command : commands) { nsIFrame* child = PrincipalChildList().FrameAt(command.mChildIndex); if (!child || !child->GetContent()->IsMathMLElement()) { continue; } auto element = dom::MathMLElement::FromNode(child->GetContent()); element->SetIncrementScriptLevel(command.mDoIncrement, true); } } NS_IMETHODIMP nsMathMLmunderoverFrame::TransmitAutomaticData() { // At this stage, all our children are in sync and we can fully // resolve our own mEmbellishData struct //--------------------------------------------------------------------- /* The REC says: As regards munder (respectively mover) : The default value of accentunder is false, unless underscript is an element or an embellished operator. If underscript is an element, the value of its accent attribute is used as the default value of accentunder. If underscript is an embellished operator, the accent attribute of the element at its core is used as the default value. As with all attributes, an explicitly given value overrides the default. XXX The winner is the outermost setting in conflicting settings like these: ... ... As regards munderover: The accent and accentunder attributes have the same effect as the attributes with the same names on and , respectively. Their default values are also computed in the same manner as described for those elements, with the default value of accent depending on overscript and the default value of accentunder depending on underscript. */ nsIFrame* overscriptFrame = nullptr; nsIFrame* underscriptFrame = nullptr; nsIFrame* baseFrame = mFrames.FirstChild(); if (baseFrame) { if (mContent->IsAnyOfMathMLElements(nsGkAtoms::munder_, nsGkAtoms::munderover_)) { underscriptFrame = baseFrame->GetNextSibling(); } else { NS_ASSERTION(mContent->IsMathMLElement(nsGkAtoms::mover_), "mContent->NodeInfo()->NameAtom() not recognized"); overscriptFrame = baseFrame->GetNextSibling(); } } if (underscriptFrame && mContent->IsMathMLElement(nsGkAtoms::munderover_)) { overscriptFrame = underscriptFrame->GetNextSibling(); } // if our base is an embellished operator, let its state bubble to us (in // particular, this is where we get the flag for // NS_MATHML_EMBELLISH_MOVABLELIMITS). Our flags are reset to the default // values of false if the base frame isn't embellished. mPresentationData.baseFrame = baseFrame; GetEmbellishDataFrom(baseFrame, mEmbellishData); // The default value of accentunder is false, unless the underscript is // embellished and its core is an accent nsEmbellishData embellishData; nsAutoString value; if (mContent->IsAnyOfMathMLElements(nsGkAtoms::munder_, nsGkAtoms::munderover_)) { GetEmbellishDataFrom(underscriptFrame, embellishData); if (NS_MATHML_EMBELLISH_IS_ACCENT(embellishData.flags)) { mEmbellishData.flags |= NS_MATHML_EMBELLISH_ACCENTUNDER; } else { mEmbellishData.flags &= ~NS_MATHML_EMBELLISH_ACCENTUNDER; } // if we have an accentunder attribute, it overrides what the underscript // said if (mContent->AsElement()->GetAttr(nsGkAtoms::accentunder_, value)) { if (value.EqualsLiteral("true")) { mEmbellishData.flags |= NS_MATHML_EMBELLISH_ACCENTUNDER; } else if (value.EqualsLiteral("false")) { mEmbellishData.flags &= ~NS_MATHML_EMBELLISH_ACCENTUNDER; } } } // The default value of accent is false, unless the overscript is embellished // and its core is an accent if (mContent->IsAnyOfMathMLElements(nsGkAtoms::mover_, nsGkAtoms::munderover_)) { GetEmbellishDataFrom(overscriptFrame, embellishData); if (NS_MATHML_EMBELLISH_IS_ACCENT(embellishData.flags)) { mEmbellishData.flags |= NS_MATHML_EMBELLISH_ACCENTOVER; } else { mEmbellishData.flags &= ~NS_MATHML_EMBELLISH_ACCENTOVER; } // if we have an accent attribute, it overrides what the overscript said if (mContent->AsElement()->GetAttr(nsGkAtoms::accent_, value)) { if (value.EqualsLiteral("true")) { mEmbellishData.flags |= NS_MATHML_EMBELLISH_ACCENTOVER; } else if (value.EqualsLiteral("false")) { mEmbellishData.flags &= ~NS_MATHML_EMBELLISH_ACCENTOVER; } } } bool subsupDisplay = NS_MATHML_EMBELLISH_IS_MOVABLELIMITS(mEmbellishData.flags) && StyleFont()->mMathStyle == StyleMathStyle::Compact; // disable the stretch-all flag if we are going to act like a superscript if (subsupDisplay) { mPresentationData.flags &= ~NS_MATHML_STRETCH_ALL_CHILDREN_HORIZONTALLY; } // Now transmit any change that we want to our children so that they // can update their mPresentationData structs //--------------------------------------------------------------------- /* The REC says: Within underscript, always sets displaystyle to "false", but increments scriptlevel by 1 only when accentunder is "false". Within overscript, always sets displaystyle to "false", but increments scriptlevel by 1 only when accent is "false". Within subscript and superscript it increments scriptlevel by 1, and sets displaystyle to "false", but leaves both attributes unchanged within base. The TeXBook treats 'over' like a superscript, so p.141 or Rule 13a say it shouldn't be compressed. However, The TeXBook says that math accents and \overline change uncramped styles to their cramped counterparts. */ if (mContent->IsAnyOfMathMLElements(nsGkAtoms::mover_, nsGkAtoms::munderover_)) { uint32_t compress = NS_MATHML_EMBELLISH_IS_ACCENTOVER(mEmbellishData.flags) ? NS_MATHML_COMPRESSED : 0; mIncrementOver = !NS_MATHML_EMBELLISH_IS_ACCENTOVER(mEmbellishData.flags) || subsupDisplay; SetIncrementScriptLevel( mContent->IsMathMLElement(nsGkAtoms::mover_) ? 1 : 2, mIncrementOver); if (mIncrementOver) { PropagateFrameFlagFor(overscriptFrame, NS_FRAME_MATHML_SCRIPT_DESCENDANT); } PropagatePresentationDataFor(overscriptFrame, compress, compress); } /* The TeXBook treats 'under' like a subscript, so p.141 or Rule 13a say it should be compressed */ if (mContent->IsAnyOfMathMLElements(nsGkAtoms::munder_, nsGkAtoms::munderover_)) { mIncrementUnder = !NS_MATHML_EMBELLISH_IS_ACCENTUNDER(mEmbellishData.flags) || subsupDisplay; SetIncrementScriptLevel(1, mIncrementUnder); if (mIncrementUnder) { PropagateFrameFlagFor(underscriptFrame, NS_FRAME_MATHML_SCRIPT_DESCENDANT); } PropagatePresentationDataFor(underscriptFrame, NS_MATHML_COMPRESSED, NS_MATHML_COMPRESSED); } /* Set flags for dtls font feature settings. dtls Dotless Forms This feature provides dotless forms for Math Alphanumeric characters, such as U+1D422 MATHEMATICAL BOLD SMALL I, U+1D423 MATHEMATICAL BOLD SMALL J, U+1D456 U+MATHEMATICAL ITALIC SMALL I, U+1D457 MATHEMATICAL ITALIC SMALL J, and so on. The dotless forms are to be used as base forms for placing mathematical accents over them. To opt out of this change, add the following to the stylesheet: "font-feature-settings: 'dtls' 0" */ if (overscriptFrame && NS_MATHML_EMBELLISH_IS_ACCENTOVER(mEmbellishData.flags) && !NS_MATHML_EMBELLISH_IS_MOVABLELIMITS(mEmbellishData.flags)) { PropagatePresentationDataFor(baseFrame, NS_MATHML_DTLS, NS_MATHML_DTLS); } return NS_OK; } /* The REC says: * If the base is an operator with movablelimits="true" (or an embellished operator whose element core has movablelimits="true"), and displaystyle="false", then underscript and overscript are drawn in a subscript and superscript position, respectively. In this case, the accent and accentunder attributes are ignored. This is often used for limits on symbols such as ∑. i.e.,: if (NS_MATHML_EMBELLISH_IS_MOVABLELIMITS(mEmbellishDataflags) && StyleFont()->mMathStyle == StyleMathStyle::Compact) { // place like subscript-superscript pair } else { // place like underscript-overscript pair } */ /* virtual */ nsresult nsMathMLmunderoverFrame::Place(DrawTarget* aDrawTarget, bool aPlaceOrigin, ReflowOutput& aDesiredSize) { float fontSizeInflation = nsLayoutUtils::FontSizeInflationFor(this); if (NS_MATHML_EMBELLISH_IS_MOVABLELIMITS(mEmbellishData.flags) && StyleFont()->mMathStyle == StyleMathStyle::Compact) { // place like sub sup or subsup if (mContent->IsMathMLElement(nsGkAtoms::munderover_)) { return nsMathMLmmultiscriptsFrame::PlaceMultiScript( PresContext(), aDrawTarget, aPlaceOrigin, aDesiredSize, this, 0, 0, fontSizeInflation); } else if (mContent->IsMathMLElement(nsGkAtoms::munder_)) { return nsMathMLmmultiscriptsFrame::PlaceMultiScript( PresContext(), aDrawTarget, aPlaceOrigin, aDesiredSize, this, 0, 0, fontSizeInflation); } else { NS_ASSERTION(mContent->IsMathMLElement(nsGkAtoms::mover_), "mContent->NodeInfo()->NameAtom() not recognized"); return nsMathMLmmultiscriptsFrame::PlaceMultiScript( PresContext(), aDrawTarget, aPlaceOrigin, aDesiredSize, this, 0, 0, fontSizeInflation); } } //////////////////////////////////// // Get the children's desired sizes nsBoundingMetrics bmBase, bmUnder, bmOver; ReflowOutput baseSize(aDesiredSize.GetWritingMode()); ReflowOutput underSize(aDesiredSize.GetWritingMode()); ReflowOutput overSize(aDesiredSize.GetWritingMode()); nsIFrame* overFrame = nullptr; nsIFrame* underFrame = nullptr; nsIFrame* baseFrame = mFrames.FirstChild(); underSize.SetBlockStartAscent(0); overSize.SetBlockStartAscent(0); bool haveError = false; if (baseFrame) { if (mContent->IsAnyOfMathMLElements(nsGkAtoms::munder_, nsGkAtoms::munderover_)) { underFrame = baseFrame->GetNextSibling(); } else if (mContent->IsMathMLElement(nsGkAtoms::mover_)) { overFrame = baseFrame->GetNextSibling(); } } if (underFrame && mContent->IsMathMLElement(nsGkAtoms::munderover_)) { overFrame = underFrame->GetNextSibling(); } if (mContent->IsMathMLElement(nsGkAtoms::munder_)) { if (!baseFrame || !underFrame || underFrame->GetNextSibling()) { // report an error, encourage people to get their markups in order haveError = true; } } if (mContent->IsMathMLElement(nsGkAtoms::mover_)) { if (!baseFrame || !overFrame || overFrame->GetNextSibling()) { // report an error, encourage people to get their markups in order haveError = true; } } if (mContent->IsMathMLElement(nsGkAtoms::munderover_)) { if (!baseFrame || !underFrame || !overFrame || overFrame->GetNextSibling()) { // report an error, encourage people to get their markups in order haveError = true; } } if (haveError) { if (aPlaceOrigin) { ReportChildCountError(); } return PlaceAsMrow(aDrawTarget, aPlaceOrigin, aDesiredSize); } GetReflowAndBoundingMetricsFor(baseFrame, baseSize, bmBase); if (underFrame) { GetReflowAndBoundingMetricsFor(underFrame, underSize, bmUnder); } if (overFrame) { GetReflowAndBoundingMetricsFor(overFrame, overSize, bmOver); } nscoord onePixel = nsPresContext::CSSPixelsToAppUnits(1); //////////////////// // Place Children RefPtr fm = nsLayoutUtils::GetFontMetricsForFrame(this, fontSizeInflation); nscoord xHeight = fm->XHeight(); nscoord oneDevPixel = fm->AppUnitsPerDevPixel(); RefPtr mathFont = fm->GetThebesFontGroup()->GetFirstMathFont(); nscoord ruleThickness; GetRuleThickness(aDrawTarget, fm, ruleThickness); nscoord correction = 0; GetItalicCorrection(bmBase, correction); // there are 2 different types of placement depending on // whether we want an accented under or not nscoord underDelta1 = 0; // gap between base and underscript nscoord underDelta2 = 0; // extra space beneath underscript if (!NS_MATHML_EMBELLISH_IS_ACCENTUNDER(mEmbellishData.flags)) { // Rule 13a, App. G, TeXbook nscoord bigOpSpacing2, bigOpSpacing4, bigOpSpacing5, dummy; GetBigOpSpacings(fm, dummy, bigOpSpacing2, dummy, bigOpSpacing4, bigOpSpacing5); if (mathFont) { // XXXfredw The Open Type MATH table has some StretchStack* parameters // that we may use when the base is a stretchy horizontal operator. See // bug 963131. bigOpSpacing2 = mathFont->MathTable()->Constant( gfxMathTable::LowerLimitGapMin, oneDevPixel); bigOpSpacing4 = mathFont->MathTable()->Constant( gfxMathTable::LowerLimitBaselineDropMin, oneDevPixel); bigOpSpacing5 = 0; } underDelta1 = std::max(bigOpSpacing2, (bigOpSpacing4 - bmUnder.ascent)); underDelta2 = bigOpSpacing5; } else { // No corresponding rule in TeXbook - we are on our own here // XXX tune the gap delta between base and underscript // XXX Should we use Rule 10 like \underline does? // XXXfredw Perhaps use the Underbar* parameters of the MATH table. See // bug 963125. underDelta1 = ruleThickness + onePixel / 2; underDelta2 = ruleThickness; } // empty under? if (!(bmUnder.ascent + bmUnder.descent)) { underDelta1 = 0; underDelta2 = 0; } nscoord overDelta1 = 0; // gap between base and overscript nscoord overDelta2 = 0; // extra space above overscript if (!NS_MATHML_EMBELLISH_IS_ACCENTOVER(mEmbellishData.flags)) { // Rule 13a, App. G, TeXbook // XXXfredw The Open Type MATH table has some StretchStack* parameters // that we may use when the base is a stretchy horizontal operator. See // bug 963131. nscoord bigOpSpacing1, bigOpSpacing3, bigOpSpacing5, dummy; GetBigOpSpacings(fm, bigOpSpacing1, dummy, bigOpSpacing3, dummy, bigOpSpacing5); if (mathFont) { // XXXfredw The Open Type MATH table has some StretchStack* parameters // that we may use when the base is a stretchy horizontal operator. See // bug 963131. bigOpSpacing1 = mathFont->MathTable()->Constant( gfxMathTable::UpperLimitGapMin, oneDevPixel); bigOpSpacing3 = mathFont->MathTable()->Constant( gfxMathTable::UpperLimitBaselineRiseMin, oneDevPixel); bigOpSpacing5 = 0; } overDelta1 = std::max(bigOpSpacing1, (bigOpSpacing3 - bmOver.descent)); overDelta2 = bigOpSpacing5; // XXX This is not a TeX rule... // delta1 (as computed abvove) can become really big when bmOver.descent is // negative, e.g., if the content is &OverBar. In such case, we use the // height if (bmOver.descent < 0) overDelta1 = std::max(bigOpSpacing1, (bigOpSpacing3 - (bmOver.ascent + bmOver.descent))); } else { // Rule 12, App. G, TeXbook // We are going to modify this rule to make it more general. // The idea behind Rule 12 in the TeXBook is to keep the accent // as close to the base as possible, while ensuring that the // distance between the *baseline* of the accent char and // the *baseline* of the base is atleast x-height. // The idea is that for normal use, we would like all the accents // on a line to line up atleast x-height above the baseline // if possible. // When the ascent of the base is >= x-height, // the baseline of the accent char is placed just above the base // (specifically, the baseline of the accent char is placed // above the baseline of the base by the ascent of the base). // For ease of implementation, // this assumes that the font-designer designs accents // in such a way that the bottom of the accent is atleast x-height // above its baseline, otherwise there will be collisions // with the base. Also there should be proper padding between // the bottom of the accent char and its baseline. // The above rule may not be obvious from a first // reading of rule 12 in the TeXBook !!! // The mathml tag can use accent chars that // do not follow this convention. So we modify TeX's rule // so that TeX's rule gets subsumed for accents that follow // TeX's convention, // while also allowing accents that do not follow the convention : // we try to keep the *bottom* of the accent char atleast x-height // from the baseline of the base char. we also slap on an extra // padding between the accent and base chars. overDelta1 = ruleThickness + onePixel / 2; nscoord accentBaseHeight = xHeight; if (mathFont) { accentBaseHeight = mathFont->MathTable()->Constant( gfxMathTable::AccentBaseHeight, oneDevPixel); } if (bmBase.ascent < accentBaseHeight) { // also ensure at least accentBaseHeight above the baseline of the base overDelta1 += accentBaseHeight - bmBase.ascent; } overDelta2 = ruleThickness; } // empty over? if (!(bmOver.ascent + bmOver.descent)) { overDelta1 = 0; overDelta2 = 0; } nscoord dxBase = 0, dxOver = 0, dxUnder = 0; nsAutoString valueAlign; ////////// // pass 1, do what does: attach the overscript on the base // Ad-hoc - This is to override fonts which have ready-made _accent_ // glyphs with negative lbearing and rbearing. We want to position // the overscript ourselves nscoord overWidth = bmOver.width; if (!overWidth && (bmOver.rightBearing - bmOver.leftBearing > 0)) { overWidth = bmOver.rightBearing - bmOver.leftBearing; dxOver = -bmOver.leftBearing; } if (NS_MATHML_EMBELLISH_IS_ACCENTOVER(mEmbellishData.flags)) { mBoundingMetrics.width = bmBase.width; dxOver += correction; } else { mBoundingMetrics.width = std::max(bmBase.width, overWidth); dxOver += correction / 2; } dxOver += (mBoundingMetrics.width - overWidth) / 2; dxBase = (mBoundingMetrics.width - bmBase.width) / 2; mBoundingMetrics.ascent = bmBase.ascent + overDelta1 + bmOver.ascent + bmOver.descent; mBoundingMetrics.descent = bmBase.descent; mBoundingMetrics.leftBearing = std::min(dxBase + bmBase.leftBearing, dxOver + bmOver.leftBearing); mBoundingMetrics.rightBearing = std::max(dxBase + bmBase.rightBearing, dxOver + bmOver.rightBearing); ////////// // pass 2, do what does: attach the underscript on the previous // result. We conceptually view the previous result as an "anynomous base" // from where to attach the underscript. Hence if the underscript is empty, // we should end up like . If the overscript is empty, we should // end up like . nsBoundingMetrics bmAnonymousBase = mBoundingMetrics; nscoord ascentAnonymousBase = std::max(mBoundingMetrics.ascent + overDelta2, overSize.BlockStartAscent() + bmOver.descent + overDelta1 + bmBase.ascent); ascentAnonymousBase = std::max(ascentAnonymousBase, baseSize.BlockStartAscent()); // Width of non-spacing marks is zero so use left and right bearing. nscoord underWidth = bmUnder.width; if (!underWidth) { underWidth = bmUnder.rightBearing - bmUnder.leftBearing; dxUnder = -bmUnder.leftBearing; } nscoord maxWidth = std::max(bmAnonymousBase.width, underWidth); if (!NS_MATHML_EMBELLISH_IS_ACCENTUNDER(mEmbellishData.flags)) { GetItalicCorrection(bmAnonymousBase, correction); dxUnder += -correction / 2; } nscoord dxAnonymousBase = 0; dxUnder += (maxWidth - underWidth) / 2; dxAnonymousBase = (maxWidth - bmAnonymousBase.width) / 2; // adjust the offsets of the real base and overscript since their // final offsets should be relative to us... dxOver += dxAnonymousBase; dxBase += dxAnonymousBase; mBoundingMetrics.width = std::max(dxAnonymousBase + bmAnonymousBase.width, dxUnder + bmUnder.width); // At this point, mBoundingMetrics.ascent = bmAnonymousBase.ascent mBoundingMetrics.descent = bmAnonymousBase.descent + underDelta1 + bmUnder.ascent + bmUnder.descent; mBoundingMetrics.leftBearing = std::min(dxAnonymousBase + bmAnonymousBase.leftBearing, dxUnder + bmUnder.leftBearing); mBoundingMetrics.rightBearing = std::max(dxAnonymousBase + bmAnonymousBase.rightBearing, dxUnder + bmUnder.rightBearing); aDesiredSize.SetBlockStartAscent(ascentAnonymousBase); aDesiredSize.Height() = aDesiredSize.BlockStartAscent() + std::max(mBoundingMetrics.descent + underDelta2, bmAnonymousBase.descent + underDelta1 + bmUnder.ascent + underSize.Height() - underSize.BlockStartAscent()); aDesiredSize.Height() = std::max(aDesiredSize.Height(), aDesiredSize.BlockStartAscent() + baseSize.Height() - baseSize.BlockStartAscent()); aDesiredSize.Width() = mBoundingMetrics.width; aDesiredSize.mBoundingMetrics = mBoundingMetrics; mReference.x = 0; mReference.y = aDesiredSize.BlockStartAscent(); if (aPlaceOrigin) { nscoord dy; // place overscript if (overFrame) { dy = aDesiredSize.BlockStartAscent() - mBoundingMetrics.ascent + bmOver.ascent - overSize.BlockStartAscent(); FinishReflowChild(overFrame, PresContext(), overSize, nullptr, dxOver, dy, ReflowChildFlags::Default); } // place base dy = aDesiredSize.BlockStartAscent() - baseSize.BlockStartAscent(); FinishReflowChild(baseFrame, PresContext(), baseSize, nullptr, dxBase, dy, ReflowChildFlags::Default); // place underscript if (underFrame) { dy = aDesiredSize.BlockStartAscent() + mBoundingMetrics.descent - bmUnder.descent - underSize.BlockStartAscent(); FinishReflowChild(underFrame, PresContext(), underSize, nullptr, dxUnder, dy, ReflowChildFlags::Default); } } return NS_OK; }