gecko-dev/layout/generic/TextDrawTarget.h

611 lines
21 KiB
C
Raw Normal View History

Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
/* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */
/* 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/. */
#ifndef TextDrawTarget_h
#define TextDrawTarget_h
#include "mozilla/gfx/2D.h"
#include "mozilla/layers/WebRenderLayerManager.h"
#include "mozilla/layers/WebRenderBridgeChild.h"
#include "mozilla/webrender/WebRenderAPI.h"
#include "mozilla/layers/StackingContextHelper.h"
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
namespace mozilla {
namespace layout {
using namespace gfx;
// This is used by all Advanced Layers users, so we use plain gfx types
struct TextRunFragment {
ScaledFont* font;
Color color;
nsTArray<gfx::Glyph> glyphs;
};
// Only webrender handles this, so we use webrender types
struct SelectionFragment {
wr::ColorF color;
wr::LayoutRect rect;
};
// Selections are used in nsTextFrame to hack in sub-frame style changes.
// Most notably text-shadows can be changed by selections, and so we need to
// group all the glyphs and decorations attached to a shadow. We do this by
// having shadows apply to an entire SelectedTextRunFragment, and creating
// one for each "piece" of selection.
//
// For instance, this text:
//
// Hello [there] my name [is Mega]man
// ^ ^
// normal selection Ctrl+F highlight selection (yeah it's very overloaded)
//
// Would be broken up into 5 SelectedTextRunFragments
//
// ["Hello ", "there", " my name ", "is Mega", "man"]
//
// For almost all nsTextFrames, there will be only one SelectedTextRunFragment.
struct SelectedTextRunFragment {
Maybe<SelectionFragment> selection;
nsTArray<wr::TextShadow> shadows;
nsTArray<TextRunFragment> text;
nsTArray<wr::Line> beforeDecorations;
nsTArray<wr::Line> afterDecorations;
};
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
// This class is fake DrawTarget, used to intercept text draw calls, while
// also collecting up the other aspects of text natively.
//
// When using advanced-layers in nsDisplayText's constructor, we construct this
// and run the full painting algorithm with this as the DrawTarget. This is
// done to avoid having to massively refactor gecko's text painting code (which
// has lots of components shared between other rendering algorithms).
//
// In some phases of the painting algorithm, we can grab the relevant values
// and feed them directly into TextDrawTarget. For instance, selections,
// decorations, and shadows are handled in this manner. In those cases we can
// also short-circuit the painting algorithm to save work.
//
// In other phases, the computed values are sufficiently buried in complex
// code that it's best for us to just intercept the final draw calls. This
// is how we handle computing the glyphs of the main text and text-emphasis
// (see our overloaded FillGlyphs implementation).
//
// To be clear: this is a big hack. With time we hope to refactor the codebase
// so that all the elements of text are handled directly by TextDrawTarget,
// which is to say everything is done like we do selections and shadows now.
// This design is a good step for doing this work incrementally.
//
// This is also likely to be a bit buggy (missing or misinterpreted info)
// while we further develop the design.
//
// This does not currently support SVG text effects.
class TextDrawTarget : public DrawTarget
{
public:
// The different phases of drawing the text we're in
// Each should only happen once, and in the given order.
enum class Phase : uint8_t {
eSelection, eUnderline, eOverline, eGlyphs, eEmphasisMarks, eLineThrough
};
explicit TextDrawTarget()
: mCurrentlyDrawing(Phase::eSelection), mHasUnsupportedFeatures(false)
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
{
mCurrentTarget = gfx::Factory::CreateDrawTarget(gfx::BackendType::SKIA, IntSize(1, 1), gfx::SurfaceFormat::B8G8R8A8);
SetSelectionIndex(0);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
// Prevent this from being copied
TextDrawTarget(const TextDrawTarget& src) = delete;
TextDrawTarget& operator=(const TextDrawTarget&) = delete;
// Change the phase of text we're drawing.
void StartDrawing(Phase aPhase) { mCurrentlyDrawing = aPhase; }
void FoundUnsupportedFeature() { mHasUnsupportedFeatures = true; }
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
void SetSelectionIndex(size_t i) {
// i should only be accessed if i-1 has already been
MOZ_ASSERT(mParts.Length() <= i);
if (mParts.Length() == i){
mParts.AppendElement();
}
mCurrentPart = &mParts[i];
}
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
// This overload just stores the glyphs/font/color.
void
FillGlyphs(ScaledFont* aFont,
const GlyphBuffer& aBuffer,
const Pattern& aPattern,
const DrawOptions& aOptions,
const GlyphRenderingOptions* aRenderingOptions) override
{
// FIXME: figure out which of these asserts are real
MOZ_RELEASE_ASSERT(aOptions.mCompositionOp == CompositionOp::OP_OVER);
MOZ_RELEASE_ASSERT(aOptions.mAlpha == 1.0f);
// Make sure we're only given color patterns
MOZ_RELEASE_ASSERT(aPattern.GetType() == PatternType::COLOR);
const ColorPattern* colorPat = static_cast<const ColorPattern*>(&aPattern);
// Make sure the font exists
MOZ_RELEASE_ASSERT(aFont);
// FIXME(?): Deal with AA on the DrawOptions, and the GlyphRenderingOptions
if (mCurrentlyDrawing != Phase::eGlyphs &&
mCurrentlyDrawing != Phase::eEmphasisMarks) {
MOZ_CRASH("TextDrawTarget received glyphs in wrong phase");
}
// We need to push a new TextRunFragment whenever the font/color changes
// (usually this implies some font fallback from mixing languages/emoji)
TextRunFragment* fragment;
if (mCurrentPart->text.IsEmpty() ||
mCurrentPart->text.LastElement().font != aFont ||
mCurrentPart->text.LastElement().color != colorPat->mColor) {
fragment = mCurrentPart->text.AppendElement();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
fragment->font = aFont;
fragment->color = colorPat->mColor;
} else {
fragment = &mCurrentPart->text.LastElement();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
nsTArray<Glyph>& glyphs = fragment->glyphs;
size_t oldLength = glyphs.Length();
glyphs.SetLength(oldLength + aBuffer.mNumGlyphs);
PodCopy(glyphs.Elements() + oldLength, aBuffer.mGlyphs, aBuffer.mNumGlyphs);
// If there's a skew for synthetic italics we need to apply it, as the font
// code applies the inverse transformation to glyph positions in anticipation.
Matrix trans = GetTransform();
if (trans._21 != 0) {
Matrix skew = Matrix(1, trans._12,
trans._21, 1,
0, 0);
for (size_t i = oldLength; i < oldLength + aBuffer.mNumGlyphs; ++i) {
auto position = &glyphs[i].mPosition;
*position = skew.TransformPoint(*position);
}
}
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void
AppendShadow(const wr::TextShadow& aShadow) {
mCurrentPart->shadows.AppendElement(aShadow);
}
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
void
SetSelectionRect(const LayoutDeviceRect& aRect, const Color& aColor)
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
{
SelectionFragment frag;
frag.rect = wr::ToLayoutRect(aRect);
frag.color = wr::ToColorF(aColor);
mCurrentPart->selection = Some(frag);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void
AppendDecoration(const Point& aStart,
const Point& aEnd,
const float aThickness,
const bool aVertical,
const Color& aColor,
const uint8_t aStyle)
{
wr::Line* decoration;
switch (mCurrentlyDrawing) {
case Phase::eUnderline:
case Phase::eOverline:
decoration = mCurrentPart->beforeDecorations.AppendElement();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
break;
case Phase::eLineThrough:
decoration = mCurrentPart->afterDecorations.AppendElement();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
break;
default:
MOZ_CRASH("TextDrawTarget received Decoration in wrong phase");
}
// This function is basically designed to slide into the decoration drawing
// code of nsCSSRendering with minimum disruption, to minimize the
// chances of implementation drift. As such, it mostly looks like a call
// to a skia-style StrokeLine method: two end-points, with a thickness
// and style. Notably the end-points are *centered* in the block direction,
// even though webrender wants a rect-like representation, where the points
// are on corners.
//
// So we mangle the format here in a single centralized place, where neither
// webrender nor nsCSSRendering has to care about this mismatch.
decoration->baseline = (aVertical ? aStart.x : aStart.y) - aThickness / 2;
decoration->start = aVertical ? aStart.y : aStart.x;
decoration->end = aVertical ? aEnd.y : aEnd.x;
decoration->width = aThickness;
decoration->color = wr::ToColorF(aColor);
decoration->orientation = aVertical
? wr::LineOrientation::Vertical
: wr::LineOrientation::Horizontal;
switch (aStyle) {
case NS_STYLE_TEXT_DECORATION_STYLE_SOLID:
decoration->style = wr::LineStyle::Solid;
break;
case NS_STYLE_TEXT_DECORATION_STYLE_DOTTED:
decoration->style = wr::LineStyle::Dotted;
break;
case NS_STYLE_TEXT_DECORATION_STYLE_DASHED:
decoration->style = wr::LineStyle::Dashed;
break;
case NS_STYLE_TEXT_DECORATION_STYLE_WAVY:
decoration->style = wr::LineStyle::Wavy;
break;
// Double lines should be lowered to two solid lines
case NS_STYLE_TEXT_DECORATION_STYLE_DOUBLE:
default:
MOZ_CRASH("TextDrawTarget received unsupported line style");
}
}
const nsTArray<SelectedTextRunFragment>& GetParts() { return mParts; }
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
bool
CanSerializeFonts()
{
if (mHasUnsupportedFeatures) {
return false;
}
for (const SelectedTextRunFragment& part : GetParts()) {
for (const TextRunFragment& frag : part.text) {
if (!frag.font->CanSerialize()) {
return false;
}
}
}
return true;
}
// TextLayers don't support very complicated text right now. This checks
// if any of the problem cases exist.
bool
ContentsAreSimple()
{
ScaledFont* font = nullptr;
for (const SelectedTextRunFragment& part : GetParts()) {
// Can't handle shadows, selections, or decorations
if (part.shadows.Length() > 0 ||
part.beforeDecorations.Length() > 0 ||
part.afterDecorations.Length() > 0 ||
part.selection.isSome()) {
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
return false;
}
// Must only have one font (multiple colors is fine)
for (const mozilla::layout::TextRunFragment& text : part.text) {
if (!font) {
font = text.font;
}
if (font != text.font) {
return false;
}
}
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
// Must have an actual font (i.e. actual text)
if (!font) {
return false;
}
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
return true;
}
void
CreateWebRenderCommands(mozilla::wr::DisplayListBuilder& aBuilder,
const layers::StackingContextHelper& aSc,
layers::WebRenderLayerManager* aManager,
nsDisplayItem* aItem,
nsRect& aBounds) {
// Drawing order: selections,
// shadows,
// underline, overline, [grouped in one array]
// text, emphasisText, [grouped in one array]
// lineThrough
// Compute clip/bounds
auto appUnitsPerDevPixel = aItem->Frame()->PresContext()->AppUnitsPerDevPixel();
LayoutDeviceRect layoutBoundsRect = LayoutDeviceRect::FromAppUnits(
aBounds, appUnitsPerDevPixel);
LayoutDeviceRect layoutClipRect = layoutBoundsRect;
auto clip = aItem->GetClip();
if (clip.HasClip()) {
layoutClipRect = LayoutDeviceRect::FromAppUnits(
clip.GetClipRect(), appUnitsPerDevPixel);
}
LayerRect boundsRect = LayerRect::FromUnknownRect(layoutBoundsRect.ToUnknownRect());
LayerRect clipRect = LayerRect::FromUnknownRect(layoutClipRect.ToUnknownRect());
bool backfaceVisible = !aItem->BackfaceIsHidden();
wr::LayoutRect wrBoundsRect = aSc.ToRelativeLayoutRect(boundsRect);
wr::LayoutRect wrClipRect = aSc.ToRelativeLayoutRect(clipRect);
// Create commands
for (auto& part : GetParts()) {
if (part.selection) {
auto selection = part.selection.value();
aBuilder.PushRect(selection.rect, wrClipRect, backfaceVisible, selection.color);
}
}
for (auto& part : GetParts()) {
// WR takes the shadows in CSS-order (reverse of rendering order),
// because the drawing of a shadow actually occurs when it's popped.
for (const wr::TextShadow& shadow : part.shadows) {
aBuilder.PushTextShadow(wrBoundsRect, wrClipRect, backfaceVisible, shadow);
}
for (const wr::Line& decoration : part.beforeDecorations) {
aBuilder.PushLine(wrClipRect, backfaceVisible, decoration);
}
for (const mozilla::layout::TextRunFragment& text : part.text) {
aManager->WrBridge()->PushGlyphs(aBuilder, text.glyphs, text.font,
text.color, aSc, boundsRect, clipRect,
backfaceVisible);
}
for (const wr::Line& decoration : part.afterDecorations) {
aBuilder.PushLine(wrClipRect, backfaceVisible, decoration);
}
for (size_t i = 0; i < part.shadows.Length(); ++i) {
aBuilder.PopTextShadow();
}
}
}
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
private:
// The part of the text we're currently drawing (glyphs, underlines, etc.)
Phase mCurrentlyDrawing;
// Which chunk of mParts is actively being populated
SelectedTextRunFragment* mCurrentPart;
// Chunks of the text, grouped by selection
nsTArray<SelectedTextRunFragment> mParts;
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
// A dummy to handle parts of the DrawTarget impl we don't care for
RefPtr<DrawTarget> mCurrentTarget;
// Whether Tofu or SVG fonts were encountered
bool mHasUnsupportedFeatures;
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
// The rest of this is dummy implementations of DrawTarget's API
public:
DrawTargetType GetType() const override {
return DrawTargetType::SOFTWARE_RASTER;
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
BackendType GetBackendType() const override {
return BackendType::WEBRENDER_TEXT;
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
bool IsRecording() const override { return true; }
bool IsCaptureDT() const override { return false; }
already_AddRefed<SourceSurface> Snapshot() override {
return mCurrentTarget->Snapshot();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<SourceSurface> IntoLuminanceSource(LuminanceType aLuminanceType,
float aOpacity) override {
return mCurrentTarget->IntoLuminanceSource(aLuminanceType, aOpacity);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
IntSize GetSize() override {
return mCurrentTarget->GetSize();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void Flush() override {
mCurrentTarget->Flush();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void DrawCapturedDT(DrawTargetCapture *aCaptureDT,
const Matrix& aTransform) override {
mCurrentTarget->DrawCapturedDT(aCaptureDT, aTransform);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void DrawSurface(SourceSurface *aSurface,
const Rect &aDest,
const Rect &aSource,
const DrawSurfaceOptions &aSurfOptions,
const DrawOptions &aOptions) override {
mCurrentTarget->DrawSurface(aSurface, aDest, aSource, aSurfOptions, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void DrawFilter(FilterNode *aNode,
const Rect &aSourceRect,
const Point &aDestPoint,
const DrawOptions &aOptions) override {
mCurrentTarget->DrawFilter(aNode, aSourceRect, aDestPoint, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void DrawSurfaceWithShadow(SourceSurface *aSurface,
const Point &aDest,
const Color &aColor,
const Point &aOffset,
Float aSigma,
CompositionOp aOperator) override {
mCurrentTarget->DrawSurfaceWithShadow(aSurface, aDest, aColor, aOffset, aSigma, aOperator);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void ClearRect(const Rect &aRect) override {
mCurrentTarget->ClearRect(aRect);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void CopySurface(SourceSurface *aSurface,
const IntRect &aSourceRect,
const IntPoint &aDestination) override {
mCurrentTarget->CopySurface(aSurface, aSourceRect, aDestination);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void FillRect(const Rect &aRect,
const Pattern &aPattern,
const DrawOptions &aOptions = DrawOptions()) override {
mCurrentTarget->FillRect(aRect, aPattern, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void StrokeRect(const Rect &aRect,
const Pattern &aPattern,
const StrokeOptions &aStrokeOptions,
const DrawOptions &aOptions) override {
mCurrentTarget->StrokeRect(aRect, aPattern, aStrokeOptions, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void StrokeLine(const Point &aStart,
const Point &aEnd,
const Pattern &aPattern,
const StrokeOptions &aStrokeOptions,
const DrawOptions &aOptions) override {
mCurrentTarget->StrokeLine(aStart, aEnd, aPattern, aStrokeOptions, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void Stroke(const Path *aPath,
const Pattern &aPattern,
const StrokeOptions &aStrokeOptions,
const DrawOptions &aOptions) override {
mCurrentTarget->Stroke(aPath, aPattern, aStrokeOptions, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void Fill(const Path *aPath,
const Pattern &aPattern,
const DrawOptions &aOptions) override {
mCurrentTarget->Fill(aPath, aPattern, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void StrokeGlyphs(ScaledFont* aFont,
const GlyphBuffer& aBuffer,
const Pattern& aPattern,
const StrokeOptions& aStrokeOptions,
const DrawOptions& aOptions,
const GlyphRenderingOptions* aRenderingOptions) override {
MOZ_ASSERT(mCurrentlyDrawing == Phase::eGlyphs);
mCurrentTarget->StrokeGlyphs(aFont, aBuffer, aPattern,
aStrokeOptions, aOptions, aRenderingOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void Mask(const Pattern &aSource,
const Pattern &aMask,
const DrawOptions &aOptions) override {
return mCurrentTarget->Mask(aSource, aMask, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void MaskSurface(const Pattern &aSource,
SourceSurface *aMask,
Point aOffset,
const DrawOptions &aOptions) override {
return mCurrentTarget->MaskSurface(aSource, aMask, aOffset, aOptions);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
bool Draw3DTransformedSurface(SourceSurface* aSurface,
const Matrix4x4& aMatrix) override {
return mCurrentTarget->Draw3DTransformedSurface(aSurface, aMatrix);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void PushClip(const Path *aPath) override {
mCurrentTarget->PushClip(aPath);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void PushClipRect(const Rect &aRect) override {
mCurrentTarget->PushClipRect(aRect);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void PushDeviceSpaceClipRects(const IntRect* aRects, uint32_t aCount) override {
mCurrentTarget->PushDeviceSpaceClipRects(aRects, aCount);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void PopClip() override {
mCurrentTarget->PopClip();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void PushLayer(bool aOpaque, Float aOpacity,
SourceSurface* aMask,
const Matrix& aMaskTransform,
const IntRect& aBounds,
bool aCopyBackground) override {
mCurrentTarget->PushLayer(aOpaque, aOpacity, aMask, aMaskTransform, aBounds, aCopyBackground);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void PopLayer() override {
mCurrentTarget->PopLayer();
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<SourceSurface> CreateSourceSurfaceFromData(unsigned char *aData,
const IntSize &aSize,
int32_t aStride,
SurfaceFormat aFormat) const override {
return mCurrentTarget->CreateSourceSurfaceFromData(aData, aSize, aStride, aFormat);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<SourceSurface> OptimizeSourceSurface(SourceSurface *aSurface) const override {
return mCurrentTarget->OptimizeSourceSurface(aSurface);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<SourceSurface>
CreateSourceSurfaceFromNativeSurface(const NativeSurface &aSurface) const override {
return mCurrentTarget->CreateSourceSurfaceFromNativeSurface(aSurface);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<DrawTarget>
CreateSimilarDrawTarget(const IntSize &aSize, SurfaceFormat aFormat) const override {
return mCurrentTarget->CreateSimilarDrawTarget(aSize, aFormat);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<PathBuilder> CreatePathBuilder(FillRule aFillRule) const override {
return mCurrentTarget->CreatePathBuilder(aFillRule);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<FilterNode> CreateFilter(FilterType aType) override {
return mCurrentTarget->CreateFilter(aType);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
already_AddRefed<GradientStops>
CreateGradientStops(GradientStop *aStops,
uint32_t aNumStops,
ExtendMode aExtendMode) const override {
return mCurrentTarget->CreateGradientStops(aStops, aNumStops, aExtendMode);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void SetTransform(const Matrix &aTransform) override {
mCurrentTarget->SetTransform(aTransform);
// Need to do this to make inherited GetTransform to work
DrawTarget::SetTransform(aTransform);
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
}
void* GetNativeSurface(NativeSurfaceType aType) override {
return mCurrentTarget->GetNativeSurface(aType);
}
void DetachAllSnapshots() override { mCurrentTarget->DetachAllSnapshots(); }
Bug 1357545 - handle text-shadows/decorations with webrender (layers-free) r=jrmuizel This replaces our DrawTargetCapture hack with a similar but more powerful TextDrawTarget hack. The old design had several limitations: * It couldn't handle shadows * It couldn't handle selections * It couldn't handle font/color changes in a single text-run * It couldn't handle decorations (underline, overline, line-through) Mostly this was a consequence of the fact that it only modified the start and end of the rendering algorithm, and therefore couldn't distinguish draw calls for different parts of the text. This new design is based on a similar principle as DrawTargetCapture, but also passes down the TextDrawTarget in the drawing arguments, so that the drawing algorithm can notify us of changes in phase (e.g. "now we're doing underlines"). This also lets us directly pass data to TextDrawTarget when possible (as is done for shadows and selections). In doing this, I also improved the logic copied from ContainsOnlyColoredGlyphs to handle changes in font/color mid-text-run (which can happen because of font fallback). The end result is: * We handle all shadows natively * We handle all selections natively * We handle all decorations natively * We handle font/color changes in a single text-run * Although we still hackily intercept draw calls * But we don't need to buffer commands, reducing total memcopies In addition, this change integrates webrender's PushTextShadow and PushLine APIs, which were designed for this use case. This is only done in the layerless path; WebrenderTextLayer continues to be semantically limited, as we aren't actively maintaining non-layers-free webrender anymore. This also doesn't modify TextLayers, to minimize churn. In theory they can be augmented to support the richer semantics that TextDrawTarget has, but there's little motivation since the API is largely unused with this change. MozReview-Commit-ID: 4IjTsSW335h --HG-- extra : rebase_source : d69f69648ade5c7a8e6bb756f4b8ab9e2543e576
2017-06-19 14:58:28 +00:00
};
}
}
#endif