mirror of
https://gitlab.winehq.org/wine/wine-gecko.git
synced 2024-09-13 09:24:08 -07:00
1018 lines
36 KiB
C++
1018 lines
36 KiB
C++
/* -*- Mode: C++; tab-width: 20; indent-tabs-mode: nil; c-basic-offset: 4 -*-
|
|
* This Source Code Form is subject to the terms of the Mozilla Public
|
|
* License, v. 2.0. If a copy of the MPL was not distributed with this
|
|
* file, You can obtain one at http://mozilla.org/MPL/2.0/. */
|
|
|
|
#include "gfxUtils.h"
|
|
#include "gfxContext.h"
|
|
#include "gfxPlatform.h"
|
|
#include "gfxDrawable.h"
|
|
#include "mozilla/gfx/2D.h"
|
|
#include "mozilla/RefPtr.h"
|
|
#include "nsRegion.h"
|
|
#include "yuv_convert.h"
|
|
#include "ycbcr_to_rgb565.h"
|
|
#include "GeckoProfiler.h"
|
|
#include "ImageContainer.h"
|
|
#include "gfx2DGlue.h"
|
|
#include "gfxPrefs.h"
|
|
|
|
#ifdef XP_WIN
|
|
#include "gfxWindowsPlatform.h"
|
|
#endif
|
|
|
|
using namespace mozilla;
|
|
using namespace mozilla::layers;
|
|
using namespace mozilla::gfx;
|
|
|
|
#include "DeprecatedPremultiplyTables.h"
|
|
|
|
static const uint8_t PremultiplyValue(uint8_t a, uint8_t v) {
|
|
return gfxUtils::sPremultiplyTable[a*256+v];
|
|
}
|
|
|
|
static const uint8_t UnpremultiplyValue(uint8_t a, uint8_t v) {
|
|
return gfxUtils::sUnpremultiplyTable[a*256+v];
|
|
}
|
|
|
|
void
|
|
gfxUtils::PremultiplyDataSurface(DataSourceSurface *aSurface)
|
|
{
|
|
// Only premultiply ARGB32
|
|
if (aSurface->GetFormat() != SurfaceFormat::B8G8R8A8) {
|
|
return;
|
|
}
|
|
|
|
DataSourceSurface::MappedSurface map;
|
|
if (!aSurface->Map(DataSourceSurface::MapType::READ_WRITE, &map)) {
|
|
return;
|
|
}
|
|
MOZ_ASSERT(map.mStride == aSurface->GetSize().width * 4,
|
|
"Source surface stride isn't tightly packed");
|
|
|
|
uint8_t *src = map.mData;
|
|
uint8_t *dst = map.mData;
|
|
|
|
uint32_t dim = aSurface->GetSize().width * aSurface->GetSize().height;
|
|
for (uint32_t i = 0; i < dim; ++i) {
|
|
#ifdef IS_LITTLE_ENDIAN
|
|
uint8_t b = *src++;
|
|
uint8_t g = *src++;
|
|
uint8_t r = *src++;
|
|
uint8_t a = *src++;
|
|
|
|
*dst++ = PremultiplyValue(a, b);
|
|
*dst++ = PremultiplyValue(a, g);
|
|
*dst++ = PremultiplyValue(a, r);
|
|
*dst++ = a;
|
|
#else
|
|
uint8_t a = *src++;
|
|
uint8_t r = *src++;
|
|
uint8_t g = *src++;
|
|
uint8_t b = *src++;
|
|
|
|
*dst++ = a;
|
|
*dst++ = PremultiplyValue(a, r);
|
|
*dst++ = PremultiplyValue(a, g);
|
|
*dst++ = PremultiplyValue(a, b);
|
|
#endif
|
|
}
|
|
|
|
aSurface->Unmap();
|
|
}
|
|
|
|
TemporaryRef<DataSourceSurface>
|
|
gfxUtils::UnpremultiplyDataSurface(DataSourceSurface* aSurface)
|
|
{
|
|
// Only premultiply ARGB32
|
|
if (aSurface->GetFormat() != SurfaceFormat::B8G8R8A8) {
|
|
return aSurface;
|
|
}
|
|
|
|
DataSourceSurface::MappedSurface map;
|
|
if (!aSurface->Map(DataSourceSurface::MapType::READ, &map)) {
|
|
return nullptr;
|
|
}
|
|
|
|
RefPtr<DataSourceSurface> dest = Factory::CreateDataSourceSurfaceWithStride(aSurface->GetSize(),
|
|
aSurface->GetFormat(),
|
|
map.mStride);
|
|
|
|
DataSourceSurface::MappedSurface destMap;
|
|
if (!dest->Map(DataSourceSurface::MapType::WRITE, &destMap)) {
|
|
aSurface->Unmap();
|
|
return nullptr;
|
|
}
|
|
|
|
uint8_t *src = map.mData;
|
|
uint8_t *dst = destMap.mData;
|
|
|
|
for (int32_t i = 0; i < aSurface->GetSize().height; ++i) {
|
|
uint8_t *srcRow = src + (i * map.mStride);
|
|
uint8_t *dstRow = dst + (i * destMap.mStride);
|
|
|
|
for (int32_t j = 0; j < aSurface->GetSize().width; ++j) {
|
|
#ifdef IS_LITTLE_ENDIAN
|
|
uint8_t b = *srcRow++;
|
|
uint8_t g = *srcRow++;
|
|
uint8_t r = *srcRow++;
|
|
uint8_t a = *srcRow++;
|
|
|
|
*dstRow++ = UnpremultiplyValue(a, b);
|
|
*dstRow++ = UnpremultiplyValue(a, g);
|
|
*dstRow++ = UnpremultiplyValue(a, r);
|
|
*dstRow++ = a;
|
|
#else
|
|
uint8_t a = *srcRow++;
|
|
uint8_t r = *srcRow++;
|
|
uint8_t g = *srcRow++;
|
|
uint8_t b = *srcRow++;
|
|
|
|
*dstRow++ = a;
|
|
*dstRow++ = UnpremultiplyValue(a, r);
|
|
*dstRow++ = UnpremultiplyValue(a, g);
|
|
*dstRow++ = UnpremultiplyValue(a, b);
|
|
#endif
|
|
}
|
|
}
|
|
|
|
aSurface->Unmap();
|
|
dest->Unmap();
|
|
return dest;
|
|
}
|
|
|
|
void
|
|
gfxUtils::ConvertBGRAtoRGBA(uint8_t* aData, uint32_t aLength)
|
|
{
|
|
MOZ_ASSERT((aLength % 4) == 0, "Loop below will pass srcEnd!");
|
|
|
|
uint8_t *src = aData;
|
|
uint8_t *srcEnd = src + aLength;
|
|
|
|
uint8_t buffer[4];
|
|
for (; src != srcEnd; src += 4) {
|
|
buffer[0] = src[2];
|
|
buffer[1] = src[1];
|
|
buffer[2] = src[0];
|
|
|
|
src[0] = buffer[0];
|
|
src[1] = buffer[1];
|
|
src[2] = buffer[2];
|
|
}
|
|
}
|
|
|
|
static bool
|
|
IsSafeImageTransformComponent(gfxFloat aValue)
|
|
{
|
|
return aValue >= -32768 && aValue <= 32767;
|
|
}
|
|
|
|
#ifndef MOZ_GFX_OPTIMIZE_MOBILE
|
|
/**
|
|
* This returns the fastest operator to use for solid surfaces which have no
|
|
* alpha channel or their alpha channel is uniformly opaque.
|
|
* This differs per render mode.
|
|
*/
|
|
static gfxContext::GraphicsOperator
|
|
OptimalFillOperator()
|
|
{
|
|
#ifdef XP_WIN
|
|
if (gfxWindowsPlatform::GetPlatform()->GetRenderMode() ==
|
|
gfxWindowsPlatform::RENDER_DIRECT2D) {
|
|
// D2D -really- hates operator source.
|
|
return gfxContext::OPERATOR_OVER;
|
|
} else {
|
|
#endif
|
|
return gfxContext::OPERATOR_SOURCE;
|
|
#ifdef XP_WIN
|
|
}
|
|
#endif
|
|
}
|
|
|
|
// EXTEND_PAD won't help us here; we have to create a temporary surface to hold
|
|
// the subimage of pixels we're allowed to sample.
|
|
static already_AddRefed<gfxDrawable>
|
|
CreateSamplingRestrictedDrawable(gfxDrawable* aDrawable,
|
|
gfxContext* aContext,
|
|
const gfxMatrix& aUserSpaceToImageSpace,
|
|
const gfxRect& aSourceRect,
|
|
const gfxRect& aSubimage,
|
|
const SurfaceFormat aFormat)
|
|
{
|
|
PROFILER_LABEL("gfxUtils", "CreateSamplingRestricedDrawable",
|
|
js::ProfileEntry::Category::GRAPHICS);
|
|
|
|
gfxRect userSpaceClipExtents = aContext->GetClipExtents();
|
|
// This isn't optimal --- if aContext has a rotation then GetClipExtents
|
|
// will have to do a bounding-box computation, and TransformBounds might
|
|
// too, so we could get a better result if we computed image space clip
|
|
// extents in one go --- but it doesn't really matter and this is easier
|
|
// to understand.
|
|
gfxRect imageSpaceClipExtents =
|
|
aUserSpaceToImageSpace.TransformBounds(userSpaceClipExtents);
|
|
// Inflate by one pixel because bilinear filtering will sample at most
|
|
// one pixel beyond the computed image pixel coordinate.
|
|
imageSpaceClipExtents.Inflate(1.0);
|
|
|
|
gfxRect needed = imageSpaceClipExtents.Intersect(aSourceRect);
|
|
needed = needed.Intersect(aSubimage);
|
|
needed.RoundOut();
|
|
|
|
// if 'needed' is empty, nothing will be drawn since aFill
|
|
// must be entirely outside the clip region, so it doesn't
|
|
// matter what we do here, but we should avoid trying to
|
|
// create a zero-size surface.
|
|
if (needed.IsEmpty())
|
|
return nullptr;
|
|
|
|
nsRefPtr<gfxDrawable> drawable;
|
|
gfxIntSize size(int32_t(needed.Width()), int32_t(needed.Height()));
|
|
|
|
nsRefPtr<gfxImageSurface> image = aDrawable->GetAsImageSurface();
|
|
if (image && gfxRect(0, 0, image->GetSize().width, image->GetSize().height).Contains(needed)) {
|
|
nsRefPtr<gfxASurface> temp = image->GetSubimage(needed);
|
|
drawable = new gfxSurfaceDrawable(temp, size, gfxMatrix().Translate(-needed.TopLeft()));
|
|
} else {
|
|
RefPtr<DrawTarget> target =
|
|
gfxPlatform::GetPlatform()->CreateOffscreenContentDrawTarget(ToIntSize(size),
|
|
aFormat);
|
|
if (!target) {
|
|
return nullptr;
|
|
}
|
|
|
|
nsRefPtr<gfxContext> tmpCtx = new gfxContext(target);
|
|
tmpCtx->SetOperator(OptimalFillOperator());
|
|
aDrawable->Draw(tmpCtx, needed - needed.TopLeft(), true,
|
|
GraphicsFilter::FILTER_FAST, gfxMatrix().Translate(needed.TopLeft()));
|
|
drawable = new gfxSurfaceDrawable(target, size, gfxMatrix().Translate(-needed.TopLeft()));
|
|
}
|
|
|
|
return drawable.forget();
|
|
}
|
|
#endif // !MOZ_GFX_OPTIMIZE_MOBILE
|
|
|
|
// working around cairo/pixman bug (bug 364968)
|
|
// Our device-space-to-image-space transform may not be acceptable to pixman.
|
|
struct MOZ_STACK_CLASS AutoCairoPixmanBugWorkaround
|
|
{
|
|
AutoCairoPixmanBugWorkaround(gfxContext* aContext,
|
|
const gfxMatrix& aDeviceSpaceToImageSpace,
|
|
const gfxRect& aFill,
|
|
const gfxASurface* aSurface)
|
|
: mContext(aContext), mSucceeded(true), mPushedGroup(false)
|
|
{
|
|
// Quartz's limits for matrix are much larger than pixman
|
|
if (!aSurface || aSurface->GetType() == gfxSurfaceType::Quartz)
|
|
return;
|
|
|
|
if (!IsSafeImageTransformComponent(aDeviceSpaceToImageSpace.xx) ||
|
|
!IsSafeImageTransformComponent(aDeviceSpaceToImageSpace.xy) ||
|
|
!IsSafeImageTransformComponent(aDeviceSpaceToImageSpace.yx) ||
|
|
!IsSafeImageTransformComponent(aDeviceSpaceToImageSpace.yy)) {
|
|
NS_WARNING("Scaling up too much, bailing out");
|
|
mSucceeded = false;
|
|
return;
|
|
}
|
|
|
|
if (IsSafeImageTransformComponent(aDeviceSpaceToImageSpace.x0) &&
|
|
IsSafeImageTransformComponent(aDeviceSpaceToImageSpace.y0))
|
|
return;
|
|
|
|
// We'll push a group, which will hopefully reduce our transform's
|
|
// translation so it's in bounds.
|
|
gfxMatrix currentMatrix = mContext->CurrentMatrix();
|
|
mContext->Save();
|
|
|
|
// Clip the rounded-out-to-device-pixels bounds of the
|
|
// transformed fill area. This is the area for the group we
|
|
// want to push.
|
|
mContext->IdentityMatrix();
|
|
gfxRect bounds = currentMatrix.TransformBounds(aFill);
|
|
bounds.RoundOut();
|
|
mContext->Clip(bounds);
|
|
mContext->SetMatrix(currentMatrix);
|
|
mContext->PushGroup(gfxContentType::COLOR_ALPHA);
|
|
mContext->SetOperator(gfxContext::OPERATOR_OVER);
|
|
|
|
mPushedGroup = true;
|
|
}
|
|
|
|
~AutoCairoPixmanBugWorkaround()
|
|
{
|
|
if (mPushedGroup) {
|
|
mContext->PopGroupToSource();
|
|
mContext->Paint();
|
|
mContext->Restore();
|
|
}
|
|
}
|
|
|
|
bool PushedGroup() { return mPushedGroup; }
|
|
bool Succeeded() { return mSucceeded; }
|
|
|
|
private:
|
|
gfxContext* mContext;
|
|
bool mSucceeded;
|
|
bool mPushedGroup;
|
|
};
|
|
|
|
static gfxMatrix
|
|
DeviceToImageTransform(gfxContext* aContext,
|
|
const gfxMatrix& aUserSpaceToImageSpace)
|
|
{
|
|
gfxFloat deviceX, deviceY;
|
|
nsRefPtr<gfxASurface> currentTarget =
|
|
aContext->CurrentSurface(&deviceX, &deviceY);
|
|
gfxMatrix currentMatrix = aContext->CurrentMatrix();
|
|
gfxMatrix deviceToUser = gfxMatrix(currentMatrix).Invert();
|
|
deviceToUser.Translate(-gfxPoint(-deviceX, -deviceY));
|
|
return gfxMatrix(deviceToUser).Multiply(aUserSpaceToImageSpace);
|
|
}
|
|
|
|
/* These heuristics are based on Source/WebCore/platform/graphics/skia/ImageSkia.cpp:computeResamplingMode() */
|
|
#ifdef MOZ_GFX_OPTIMIZE_MOBILE
|
|
static GraphicsFilter ReduceResamplingFilter(GraphicsFilter aFilter,
|
|
int aImgWidth, int aImgHeight,
|
|
float aSourceWidth, float aSourceHeight)
|
|
{
|
|
// Images smaller than this in either direction are considered "small" and
|
|
// are not resampled ever (see below).
|
|
const int kSmallImageSizeThreshold = 8;
|
|
|
|
// The amount an image can be stretched in a single direction before we
|
|
// say that it is being stretched so much that it must be a line or
|
|
// background that doesn't need resampling.
|
|
const float kLargeStretch = 3.0f;
|
|
|
|
if (aImgWidth <= kSmallImageSizeThreshold
|
|
|| aImgHeight <= kSmallImageSizeThreshold) {
|
|
// Never resample small images. These are often used for borders and
|
|
// rules (think 1x1 images used to make lines).
|
|
return GraphicsFilter::FILTER_NEAREST;
|
|
}
|
|
|
|
if (aImgHeight * kLargeStretch <= aSourceHeight || aImgWidth * kLargeStretch <= aSourceWidth) {
|
|
// Large image tiling detected.
|
|
|
|
// Don't resample if it is being tiled a lot in only one direction.
|
|
// This is trying to catch cases where somebody has created a border
|
|
// (which might be large) and then is stretching it to fill some part
|
|
// of the page.
|
|
if (fabs(aSourceWidth - aImgWidth)/aImgWidth < 0.5 || fabs(aSourceHeight - aImgHeight)/aImgHeight < 0.5)
|
|
return GraphicsFilter::FILTER_NEAREST;
|
|
|
|
// The image is growing a lot and in more than one direction. Resampling
|
|
// is slow and doesn't give us very much when growing a lot.
|
|
return aFilter;
|
|
}
|
|
|
|
/* Some notes on other heuristics:
|
|
The Skia backend also uses nearest for backgrounds that are stretched by
|
|
a large amount. I'm not sure this is common enough for us to worry about
|
|
now. It also uses nearest for backgrounds/avoids high quality for images
|
|
that are very slightly scaled. I'm also not sure that very slightly
|
|
scaled backgrounds are common enough us to worry about.
|
|
|
|
We don't currently have much support for doing high quality interpolation.
|
|
The only place this currently happens is on Quartz and we don't have as
|
|
much control over it as would be needed. Webkit avoids using high quality
|
|
resampling during load. It also avoids high quality if the transformation
|
|
is not just a scale and translation
|
|
|
|
WebKit bug #40045 added code to avoid resampling different parts
|
|
of an image with different methods by using a resampling hint size.
|
|
It currently looks unused in WebKit but it's something to watch out for.
|
|
*/
|
|
|
|
return aFilter;
|
|
}
|
|
#else
|
|
static GraphicsFilter ReduceResamplingFilter(GraphicsFilter aFilter,
|
|
int aImgWidth, int aImgHeight,
|
|
int aSourceWidth, int aSourceHeight)
|
|
{
|
|
// Just pass the filter through unchanged
|
|
return aFilter;
|
|
}
|
|
#endif
|
|
|
|
/* static */ void
|
|
gfxUtils::DrawPixelSnapped(gfxContext* aContext,
|
|
gfxDrawable* aDrawable,
|
|
const gfxMatrix& aUserSpaceToImageSpace,
|
|
const gfxRect& aSubimage,
|
|
const gfxRect& aSourceRect,
|
|
const gfxRect& aImageRect,
|
|
const gfxRect& aFill,
|
|
const SurfaceFormat aFormat,
|
|
GraphicsFilter aFilter,
|
|
uint32_t aImageFlags)
|
|
{
|
|
PROFILER_LABEL("gfxUtils", "DrawPixelSnapped",
|
|
js::ProfileEntry::Category::GRAPHICS);
|
|
|
|
bool doTile = !aImageRect.Contains(aSourceRect) &&
|
|
!(aImageFlags & imgIContainer::FLAG_CLAMP);
|
|
|
|
nsRefPtr<gfxASurface> currentTarget = aContext->CurrentSurface();
|
|
gfxMatrix deviceSpaceToImageSpace =
|
|
DeviceToImageTransform(aContext, aUserSpaceToImageSpace);
|
|
|
|
AutoCairoPixmanBugWorkaround workaround(aContext, deviceSpaceToImageSpace,
|
|
aFill, currentTarget);
|
|
if (!workaround.Succeeded())
|
|
return;
|
|
|
|
nsRefPtr<gfxDrawable> drawable = aDrawable;
|
|
|
|
aFilter = ReduceResamplingFilter(aFilter, aImageRect.Width(), aImageRect.Height(), aSourceRect.Width(), aSourceRect.Height());
|
|
|
|
gfxMatrix userSpaceToImageSpace = aUserSpaceToImageSpace;
|
|
|
|
// On Mobile, we don't ever want to do this; it has the potential for
|
|
// allocating very large temporary surfaces, especially since we'll
|
|
// do full-page snapshots often (see bug 749426).
|
|
#ifdef MOZ_GFX_OPTIMIZE_MOBILE
|
|
// If the pattern translation is large we can get into trouble with pixman's
|
|
// 16 bit coordinate limits. For now, we only do this on platforms where
|
|
// we know we have the pixman limits. 16384.0 is a somewhat arbitrary
|
|
// large number to make sure we avoid the expensive fmod when we can, but
|
|
// still maintain a safe margin from the actual limit
|
|
if (doTile && (userSpaceToImageSpace.y0 > 16384.0 || userSpaceToImageSpace.x0 > 16384.0)) {
|
|
userSpaceToImageSpace.x0 = fmod(userSpaceToImageSpace.x0, aImageRect.width);
|
|
userSpaceToImageSpace.y0 = fmod(userSpaceToImageSpace.y0, aImageRect.height);
|
|
}
|
|
#else
|
|
// OK now, the hard part left is to account for the subimage sampling
|
|
// restriction. If all the transforms involved are just integer
|
|
// translations, then we assume no resampling will occur so there's
|
|
// nothing to do.
|
|
// XXX if only we had source-clipping in cairo!
|
|
if (aContext->CurrentMatrix().HasNonIntegerTranslation() ||
|
|
aUserSpaceToImageSpace.HasNonIntegerTranslation()) {
|
|
if (doTile || !aSubimage.Contains(aImageRect)) {
|
|
nsRefPtr<gfxDrawable> restrictedDrawable =
|
|
CreateSamplingRestrictedDrawable(aDrawable, aContext,
|
|
aUserSpaceToImageSpace, aSourceRect,
|
|
aSubimage, aFormat);
|
|
if (restrictedDrawable) {
|
|
drawable.swap(restrictedDrawable);
|
|
}
|
|
}
|
|
// We no longer need to tile: Either we never needed to, or we already
|
|
// filled a surface with the tiled pattern; this surface can now be
|
|
// drawn without tiling.
|
|
doTile = false;
|
|
}
|
|
#endif
|
|
|
|
drawable->Draw(aContext, aFill, doTile, aFilter, userSpaceToImageSpace);
|
|
}
|
|
|
|
/* static */ int
|
|
gfxUtils::ImageFormatToDepth(gfxImageFormat aFormat)
|
|
{
|
|
switch (aFormat) {
|
|
case gfxImageFormat::ARGB32:
|
|
return 32;
|
|
case gfxImageFormat::RGB24:
|
|
return 24;
|
|
case gfxImageFormat::RGB16_565:
|
|
return 16;
|
|
default:
|
|
break;
|
|
}
|
|
return 0;
|
|
}
|
|
|
|
static void
|
|
PathFromRegionInternal(gfxContext* aContext, const nsIntRegion& aRegion,
|
|
bool aSnap)
|
|
{
|
|
aContext->NewPath();
|
|
nsIntRegionRectIterator iter(aRegion);
|
|
const nsIntRect* r;
|
|
while ((r = iter.Next()) != nullptr) {
|
|
aContext->Rectangle(gfxRect(r->x, r->y, r->width, r->height), aSnap);
|
|
}
|
|
}
|
|
|
|
static void
|
|
ClipToRegionInternal(gfxContext* aContext, const nsIntRegion& aRegion,
|
|
bool aSnap)
|
|
{
|
|
PathFromRegionInternal(aContext, aRegion, aSnap);
|
|
aContext->Clip();
|
|
}
|
|
|
|
static TemporaryRef<Path>
|
|
PathFromRegionInternal(DrawTarget* aTarget, const nsIntRegion& aRegion,
|
|
bool aSnap)
|
|
{
|
|
Matrix mat = aTarget->GetTransform();
|
|
const gfxFloat epsilon = 0.000001;
|
|
#define WITHIN_E(a,b) (fabs((a)-(b)) < epsilon)
|
|
// We're essentially duplicating the logic in UserToDevicePixelSnapped here.
|
|
bool shouldNotSnap = !aSnap || (WITHIN_E(mat._11,1.0) &&
|
|
WITHIN_E(mat._22,1.0) &&
|
|
WITHIN_E(mat._12,0.0) &&
|
|
WITHIN_E(mat._21,0.0));
|
|
#undef WITHIN_E
|
|
|
|
RefPtr<PathBuilder> pb = aTarget->CreatePathBuilder();
|
|
nsIntRegionRectIterator iter(aRegion);
|
|
|
|
const nsIntRect* r;
|
|
if (shouldNotSnap) {
|
|
while ((r = iter.Next()) != nullptr) {
|
|
pb->MoveTo(Point(r->x, r->y));
|
|
pb->LineTo(Point(r->XMost(), r->y));
|
|
pb->LineTo(Point(r->XMost(), r->YMost()));
|
|
pb->LineTo(Point(r->x, r->YMost()));
|
|
pb->Close();
|
|
}
|
|
} else {
|
|
while ((r = iter.Next()) != nullptr) {
|
|
Rect rect(r->x, r->y, r->width, r->height);
|
|
|
|
rect.Round();
|
|
pb->MoveTo(rect.TopLeft());
|
|
pb->LineTo(rect.TopRight());
|
|
pb->LineTo(rect.BottomRight());
|
|
pb->LineTo(rect.BottomLeft());
|
|
pb->Close();
|
|
}
|
|
}
|
|
RefPtr<Path> path = pb->Finish();
|
|
return path;
|
|
}
|
|
|
|
static void
|
|
ClipToRegionInternal(DrawTarget* aTarget, const nsIntRegion& aRegion,
|
|
bool aSnap)
|
|
{
|
|
RefPtr<Path> path = PathFromRegionInternal(aTarget, aRegion, aSnap);
|
|
aTarget->PushClip(path);
|
|
}
|
|
|
|
/*static*/ void
|
|
gfxUtils::ClipToRegion(gfxContext* aContext, const nsIntRegion& aRegion)
|
|
{
|
|
ClipToRegionInternal(aContext, aRegion, false);
|
|
}
|
|
|
|
/*static*/ void
|
|
gfxUtils::ClipToRegion(DrawTarget* aTarget, const nsIntRegion& aRegion)
|
|
{
|
|
ClipToRegionInternal(aTarget, aRegion, false);
|
|
}
|
|
|
|
/*static*/ void
|
|
gfxUtils::ClipToRegionSnapped(gfxContext* aContext, const nsIntRegion& aRegion)
|
|
{
|
|
ClipToRegionInternal(aContext, aRegion, true);
|
|
}
|
|
|
|
/*static*/ void
|
|
gfxUtils::ClipToRegionSnapped(DrawTarget* aTarget, const nsIntRegion& aRegion)
|
|
{
|
|
ClipToRegionInternal(aTarget, aRegion, true);
|
|
}
|
|
|
|
/*static*/ gfxFloat
|
|
gfxUtils::ClampToScaleFactor(gfxFloat aVal)
|
|
{
|
|
// Arbitary scale factor limitation. We can increase this
|
|
// for better scaling performance at the cost of worse
|
|
// quality.
|
|
static const gfxFloat kScaleResolution = 2;
|
|
|
|
// Negative scaling is just a flip and irrelevant to
|
|
// our resolution calculation.
|
|
if (aVal < 0.0) {
|
|
aVal = -aVal;
|
|
}
|
|
|
|
bool inverse = false;
|
|
if (aVal < 1.0) {
|
|
inverse = true;
|
|
aVal = 1 / aVal;
|
|
}
|
|
|
|
gfxFloat power = log(aVal)/log(kScaleResolution);
|
|
|
|
// If power is within 1e-6 of an integer, round to nearest to
|
|
// prevent floating point errors, otherwise round up to the
|
|
// next integer value.
|
|
if (fabs(power - NS_round(power)) < 1e-6) {
|
|
power = NS_round(power);
|
|
} else if (inverse) {
|
|
power = floor(power);
|
|
} else {
|
|
power = ceil(power);
|
|
}
|
|
|
|
gfxFloat scale = pow(kScaleResolution, power);
|
|
|
|
if (inverse) {
|
|
scale = 1 / scale;
|
|
}
|
|
|
|
return scale;
|
|
}
|
|
|
|
|
|
/*static*/ void
|
|
gfxUtils::PathFromRegion(gfxContext* aContext, const nsIntRegion& aRegion)
|
|
{
|
|
PathFromRegionInternal(aContext, aRegion, false);
|
|
}
|
|
|
|
/*static*/ void
|
|
gfxUtils::PathFromRegionSnapped(gfxContext* aContext, const nsIntRegion& aRegion)
|
|
{
|
|
PathFromRegionInternal(aContext, aRegion, true);
|
|
}
|
|
|
|
gfxMatrix
|
|
gfxUtils::TransformRectToRect(const gfxRect& aFrom, const gfxPoint& aToTopLeft,
|
|
const gfxPoint& aToTopRight, const gfxPoint& aToBottomRight)
|
|
{
|
|
gfxMatrix m;
|
|
if (aToTopRight.y == aToTopLeft.y && aToTopRight.x == aToBottomRight.x) {
|
|
// Not a rotation, so xy and yx are zero
|
|
m.xy = m.yx = 0.0;
|
|
m.xx = (aToBottomRight.x - aToTopLeft.x)/aFrom.width;
|
|
m.yy = (aToBottomRight.y - aToTopLeft.y)/aFrom.height;
|
|
m.x0 = aToTopLeft.x - m.xx*aFrom.x;
|
|
m.y0 = aToTopLeft.y - m.yy*aFrom.y;
|
|
} else {
|
|
NS_ASSERTION(aToTopRight.y == aToBottomRight.y && aToTopRight.x == aToTopLeft.x,
|
|
"Destination rectangle not axis-aligned");
|
|
m.xx = m.yy = 0.0;
|
|
m.xy = (aToBottomRight.x - aToTopLeft.x)/aFrom.height;
|
|
m.yx = (aToBottomRight.y - aToTopLeft.y)/aFrom.width;
|
|
m.x0 = aToTopLeft.x - m.xy*aFrom.y;
|
|
m.y0 = aToTopLeft.y - m.yx*aFrom.x;
|
|
}
|
|
return m;
|
|
}
|
|
|
|
Matrix
|
|
gfxUtils::TransformRectToRect(const gfxRect& aFrom, const IntPoint& aToTopLeft,
|
|
const IntPoint& aToTopRight, const IntPoint& aToBottomRight)
|
|
{
|
|
Matrix m;
|
|
if (aToTopRight.y == aToTopLeft.y && aToTopRight.x == aToBottomRight.x) {
|
|
// Not a rotation, so xy and yx are zero
|
|
m._12 = m._21 = 0.0;
|
|
m._11 = (aToBottomRight.x - aToTopLeft.x)/aFrom.width;
|
|
m._22 = (aToBottomRight.y - aToTopLeft.y)/aFrom.height;
|
|
m._31 = aToTopLeft.x - m._11*aFrom.x;
|
|
m._32 = aToTopLeft.y - m._22*aFrom.y;
|
|
} else {
|
|
NS_ASSERTION(aToTopRight.y == aToBottomRight.y && aToTopRight.x == aToTopLeft.x,
|
|
"Destination rectangle not axis-aligned");
|
|
m._11 = m._22 = 0.0;
|
|
m._21 = (aToBottomRight.x - aToTopLeft.x)/aFrom.height;
|
|
m._12 = (aToBottomRight.y - aToTopLeft.y)/aFrom.width;
|
|
m._31 = aToTopLeft.x - m._21*aFrom.y;
|
|
m._32 = aToTopLeft.y - m._12*aFrom.x;
|
|
}
|
|
return m;
|
|
}
|
|
|
|
/* This function is sort of shitty. We truncate doubles
|
|
* to ints then convert those ints back to doubles to make sure that
|
|
* they equal the doubles that we got in. */
|
|
bool
|
|
gfxUtils::GfxRectToIntRect(const gfxRect& aIn, nsIntRect* aOut)
|
|
{
|
|
*aOut = nsIntRect(int32_t(aIn.X()), int32_t(aIn.Y()),
|
|
int32_t(aIn.Width()), int32_t(aIn.Height()));
|
|
return gfxRect(aOut->x, aOut->y, aOut->width, aOut->height).IsEqualEdges(aIn);
|
|
}
|
|
|
|
void
|
|
gfxUtils::GetYCbCrToRGBDestFormatAndSize(const PlanarYCbCrData& aData,
|
|
gfxImageFormat& aSuggestedFormat,
|
|
gfxIntSize& aSuggestedSize)
|
|
{
|
|
YUVType yuvtype =
|
|
TypeFromSize(aData.mYSize.width,
|
|
aData.mYSize.height,
|
|
aData.mCbCrSize.width,
|
|
aData.mCbCrSize.height);
|
|
|
|
// 'prescale' is true if the scaling is to be done as part of the
|
|
// YCbCr to RGB conversion rather than on the RGB data when rendered.
|
|
bool prescale = aSuggestedSize.width > 0 && aSuggestedSize.height > 0 &&
|
|
ToIntSize(aSuggestedSize) != aData.mPicSize;
|
|
|
|
if (aSuggestedFormat == gfxImageFormat::RGB16_565) {
|
|
#if defined(HAVE_YCBCR_TO_RGB565)
|
|
if (prescale &&
|
|
!IsScaleYCbCrToRGB565Fast(aData.mPicX,
|
|
aData.mPicY,
|
|
aData.mPicSize.width,
|
|
aData.mPicSize.height,
|
|
aSuggestedSize.width,
|
|
aSuggestedSize.height,
|
|
yuvtype,
|
|
FILTER_BILINEAR) &&
|
|
IsConvertYCbCrToRGB565Fast(aData.mPicX,
|
|
aData.mPicY,
|
|
aData.mPicSize.width,
|
|
aData.mPicSize.height,
|
|
yuvtype)) {
|
|
prescale = false;
|
|
}
|
|
#else
|
|
// yuv2rgb16 function not available
|
|
aSuggestedFormat = gfxImageFormat::RGB24;
|
|
#endif
|
|
}
|
|
else if (aSuggestedFormat != gfxImageFormat::RGB24) {
|
|
// No other formats are currently supported.
|
|
aSuggestedFormat = gfxImageFormat::RGB24;
|
|
}
|
|
if (aSuggestedFormat == gfxImageFormat::RGB24) {
|
|
/* ScaleYCbCrToRGB32 does not support a picture offset, nor 4:4:4 data.
|
|
See bugs 639415 and 640073. */
|
|
if (aData.mPicX != 0 || aData.mPicY != 0 || yuvtype == YV24)
|
|
prescale = false;
|
|
}
|
|
if (!prescale) {
|
|
ToIntSize(aSuggestedSize) = aData.mPicSize;
|
|
}
|
|
}
|
|
|
|
void
|
|
gfxUtils::ConvertYCbCrToRGB(const PlanarYCbCrData& aData,
|
|
const gfxImageFormat& aDestFormat,
|
|
const gfxIntSize& aDestSize,
|
|
unsigned char* aDestBuffer,
|
|
int32_t aStride)
|
|
{
|
|
// ConvertYCbCrToRGB et al. assume the chroma planes are rounded up if the
|
|
// luma plane is odd sized.
|
|
MOZ_ASSERT((aData.mCbCrSize.width == aData.mYSize.width ||
|
|
aData.mCbCrSize.width == (aData.mYSize.width + 1) >> 1) &&
|
|
(aData.mCbCrSize.height == aData.mYSize.height ||
|
|
aData.mCbCrSize.height == (aData.mYSize.height + 1) >> 1));
|
|
YUVType yuvtype =
|
|
TypeFromSize(aData.mYSize.width,
|
|
aData.mYSize.height,
|
|
aData.mCbCrSize.width,
|
|
aData.mCbCrSize.height);
|
|
|
|
// Convert from YCbCr to RGB now, scaling the image if needed.
|
|
if (ToIntSize(aDestSize) != aData.mPicSize) {
|
|
#if defined(HAVE_YCBCR_TO_RGB565)
|
|
if (aDestFormat == gfxImageFormat::RGB16_565) {
|
|
ScaleYCbCrToRGB565(aData.mYChannel,
|
|
aData.mCbChannel,
|
|
aData.mCrChannel,
|
|
aDestBuffer,
|
|
aData.mPicX,
|
|
aData.mPicY,
|
|
aData.mPicSize.width,
|
|
aData.mPicSize.height,
|
|
aDestSize.width,
|
|
aDestSize.height,
|
|
aData.mYStride,
|
|
aData.mCbCrStride,
|
|
aStride,
|
|
yuvtype,
|
|
FILTER_BILINEAR);
|
|
} else
|
|
#endif
|
|
ScaleYCbCrToRGB32(aData.mYChannel,
|
|
aData.mCbChannel,
|
|
aData.mCrChannel,
|
|
aDestBuffer,
|
|
aData.mPicSize.width,
|
|
aData.mPicSize.height,
|
|
aDestSize.width,
|
|
aDestSize.height,
|
|
aData.mYStride,
|
|
aData.mCbCrStride,
|
|
aStride,
|
|
yuvtype,
|
|
ROTATE_0,
|
|
FILTER_BILINEAR);
|
|
} else { // no prescale
|
|
#if defined(HAVE_YCBCR_TO_RGB565)
|
|
if (aDestFormat == gfxImageFormat::RGB16_565) {
|
|
ConvertYCbCrToRGB565(aData.mYChannel,
|
|
aData.mCbChannel,
|
|
aData.mCrChannel,
|
|
aDestBuffer,
|
|
aData.mPicX,
|
|
aData.mPicY,
|
|
aData.mPicSize.width,
|
|
aData.mPicSize.height,
|
|
aData.mYStride,
|
|
aData.mCbCrStride,
|
|
aStride,
|
|
yuvtype);
|
|
} else // aDestFormat != gfxImageFormat::RGB16_565
|
|
#endif
|
|
ConvertYCbCrToRGB32(aData.mYChannel,
|
|
aData.mCbChannel,
|
|
aData.mCrChannel,
|
|
aDestBuffer,
|
|
aData.mPicX,
|
|
aData.mPicY,
|
|
aData.mPicSize.width,
|
|
aData.mPicSize.height,
|
|
aData.mYStride,
|
|
aData.mCbCrStride,
|
|
aStride,
|
|
yuvtype);
|
|
}
|
|
}
|
|
|
|
/* static */ TemporaryRef<DataSourceSurface>
|
|
gfxUtils::CopySurfaceToDataSourceSurfaceWithFormat(SourceSurface* aSurface,
|
|
SurfaceFormat aFormat)
|
|
{
|
|
MOZ_ASSERT(aFormat != aSurface->GetFormat(),
|
|
"Unnecessary - and very expersive - surface format conversion");
|
|
|
|
Rect bounds(0, 0, aSurface->GetSize().width, aSurface->GetSize().height);
|
|
|
|
if (aSurface->GetType() != SurfaceType::DATA) {
|
|
// If the surface is NOT of type DATA then its data is not mapped into main
|
|
// memory. Format conversion is probably faster on the GPU, and by doing it
|
|
// there we can avoid any expensive uploads/readbacks except for (possibly)
|
|
// a single readback due to the unavoidable GetDataSurface() call. Using
|
|
// CreateOffscreenContentDrawTarget ensures the conversion happens on the
|
|
// GPU.
|
|
RefPtr<DrawTarget> dt = gfxPlatform::GetPlatform()->
|
|
CreateOffscreenContentDrawTarget(aSurface->GetSize(), aFormat);
|
|
// Using DrawSurface() here rather than CopySurface() because CopySurface
|
|
// is optimized for memcpy and therefore isn't good for format conversion.
|
|
// Using OP_OVER since in our case it's equivalent to OP_SOURCE and
|
|
// generally more optimized.
|
|
dt->DrawSurface(aSurface, bounds, bounds, DrawSurfaceOptions(),
|
|
DrawOptions(1.0f, CompositionOp::OP_OVER));
|
|
RefPtr<SourceSurface> surface = dt->Snapshot();
|
|
return surface->GetDataSurface();
|
|
}
|
|
|
|
// If the surface IS of type DATA then it may or may not be in main memory
|
|
// depending on whether or not it has been mapped yet. We have no way of
|
|
// knowing, so we can't be sure if it's best to create a data wrapping
|
|
// DrawTarget for the conversion or an offscreen content DrawTarget. We could
|
|
// guess it's not mapped and create an offscreen content DrawTarget, but if
|
|
// it is then we'll end up uploading the surface data, and most likely the
|
|
// caller is going to be accessing the resulting surface data, resulting in a
|
|
// readback (both very expensive operations). Alternatively we could guess
|
|
// the data is mapped and create a data wrapping DrawTarget and, if the
|
|
// surface is not in main memory, then we will incure a readback. The latter
|
|
// of these two "wrong choices" is the least costly (a readback, vs an
|
|
// upload and a readback), and more than likely the DATA surface that we've
|
|
// been passed actually IS in main memory anyway. For these reasons it's most
|
|
// likely best to create a data wrapping DrawTarget here to do the format
|
|
// conversion.
|
|
RefPtr<DataSourceSurface> dataSurface =
|
|
Factory::CreateDataSourceSurface(aSurface->GetSize(), aFormat);
|
|
DataSourceSurface::MappedSurface map;
|
|
if (!dataSurface ||
|
|
!dataSurface->Map(DataSourceSurface::MapType::READ_WRITE, &map)) {
|
|
return nullptr;
|
|
}
|
|
RefPtr<DrawTarget> dt =
|
|
Factory::CreateDrawTargetForData(BackendType::CAIRO,
|
|
map.mData,
|
|
dataSurface->GetSize(),
|
|
map.mStride,
|
|
aFormat);
|
|
if (!dt) {
|
|
dataSurface->Unmap();
|
|
return nullptr;
|
|
}
|
|
// Using DrawSurface() here rather than CopySurface() because CopySurface
|
|
// is optimized for memcpy and therefore isn't good for format conversion.
|
|
// Using OP_OVER since in our case it's equivalent to OP_SOURCE and
|
|
// generally more optimized.
|
|
dt->DrawSurface(aSurface, bounds, bounds, DrawSurfaceOptions(),
|
|
DrawOptions(1.0f, CompositionOp::OP_OVER));
|
|
dataSurface->Unmap();
|
|
return dataSurface.forget();
|
|
}
|
|
|
|
const uint32_t gfxUtils::sNumFrameColors = 8;
|
|
|
|
/* static */ const gfx::Color&
|
|
gfxUtils::GetColorForFrameNumber(uint64_t aFrameNumber)
|
|
{
|
|
static bool initialized = false;
|
|
static gfx::Color colors[sNumFrameColors];
|
|
|
|
if (!initialized) {
|
|
uint32_t i = 0;
|
|
colors[i++] = gfx::Color::FromABGR(0xffff0000);
|
|
colors[i++] = gfx::Color::FromABGR(0xffcc00ff);
|
|
colors[i++] = gfx::Color::FromABGR(0xff0066cc);
|
|
colors[i++] = gfx::Color::FromABGR(0xff00ff00);
|
|
colors[i++] = gfx::Color::FromABGR(0xff33ffff);
|
|
colors[i++] = gfx::Color::FromABGR(0xffff0099);
|
|
colors[i++] = gfx::Color::FromABGR(0xff0000ff);
|
|
colors[i++] = gfx::Color::FromABGR(0xff999999);
|
|
MOZ_ASSERT(i == sNumFrameColors);
|
|
initialized = true;
|
|
}
|
|
|
|
return colors[aFrameNumber % sNumFrameColors];
|
|
}
|
|
|
|
#ifdef MOZ_DUMP_PAINTING
|
|
/* static */ void
|
|
gfxUtils::WriteAsPNG(DrawTarget* aDT, const char* aFile)
|
|
{
|
|
aDT->Flush();
|
|
nsRefPtr<gfxASurface> surf = gfxPlatform::GetPlatform()->GetThebesSurfaceForDrawTarget(aDT);
|
|
if (surf) {
|
|
surf->WriteAsPNG(aFile);
|
|
} else {
|
|
NS_WARNING("Failed to get Thebes surface!");
|
|
}
|
|
}
|
|
|
|
/* static */ void
|
|
gfxUtils::DumpAsDataURL(DrawTarget* aDT)
|
|
{
|
|
aDT->Flush();
|
|
nsRefPtr<gfxASurface> surf = gfxPlatform::GetPlatform()->GetThebesSurfaceForDrawTarget(aDT);
|
|
if (surf) {
|
|
surf->DumpAsDataURL();
|
|
} else {
|
|
NS_WARNING("Failed to get Thebes surface!");
|
|
}
|
|
}
|
|
|
|
/* static */ void
|
|
gfxUtils::CopyAsDataURL(DrawTarget* aDT)
|
|
{
|
|
aDT->Flush();
|
|
nsRefPtr<gfxASurface> surf = gfxPlatform::GetPlatform()->GetThebesSurfaceForDrawTarget(aDT);
|
|
if (surf) {
|
|
surf->CopyAsDataURL();
|
|
} else {
|
|
NS_WARNING("Failed to get Thebes surface!");
|
|
}
|
|
}
|
|
|
|
/* static */ void
|
|
gfxUtils::WriteAsPNG(gfx::SourceSurface* aSourceSurface, const char* aFile)
|
|
{
|
|
RefPtr<gfx::DataSourceSurface> dataSurface = aSourceSurface->GetDataSurface();
|
|
RefPtr<gfx::DrawTarget> dt
|
|
= gfxPlatform::GetPlatform()
|
|
->CreateDrawTargetForData(dataSurface->GetData(),
|
|
dataSurface->GetSize(),
|
|
dataSurface->Stride(),
|
|
aSourceSurface->GetFormat());
|
|
gfxUtils::WriteAsPNG(dt.get(), aFile);
|
|
}
|
|
|
|
/* static */ void
|
|
gfxUtils::DumpAsDataURL(gfx::SourceSurface* aSourceSurface)
|
|
{
|
|
RefPtr<gfx::DataSourceSurface> dataSurface = aSourceSurface->GetDataSurface();
|
|
RefPtr<gfx::DrawTarget> dt
|
|
= gfxPlatform::GetPlatform()
|
|
->CreateDrawTargetForData(dataSurface->GetData(),
|
|
dataSurface->GetSize(),
|
|
dataSurface->Stride(),
|
|
aSourceSurface->GetFormat());
|
|
gfxUtils::DumpAsDataURL(dt.get());
|
|
}
|
|
|
|
/* static */ void
|
|
gfxUtils::CopyAsDataURL(gfx::SourceSurface* aSourceSurface)
|
|
{
|
|
RefPtr<gfx::DataSourceSurface> dataSurface = aSourceSurface->GetDataSurface();
|
|
RefPtr<gfx::DrawTarget> dt
|
|
= gfxPlatform::GetPlatform()
|
|
->CreateDrawTargetForData(dataSurface->GetData(),
|
|
dataSurface->GetSize(),
|
|
dataSurface->Stride(),
|
|
aSourceSurface->GetFormat());
|
|
|
|
gfxUtils::CopyAsDataURL(dt.get());
|
|
}
|
|
|
|
static bool sDumpPaintList = getenv("MOZ_DUMP_PAINT_LIST") != 0;
|
|
|
|
/* static */ bool
|
|
gfxUtils::DumpPaintList() {
|
|
return sDumpPaintList || gfxPrefs::LayoutDumpDisplayList();
|
|
}
|
|
|
|
bool gfxUtils::sDumpPainting = getenv("MOZ_DUMP_PAINT") != 0;
|
|
bool gfxUtils::sDumpPaintingToFile = getenv("MOZ_DUMP_PAINT_TO_FILE") != 0;
|
|
FILE *gfxUtils::sDumpPaintFile = nullptr;
|
|
#endif
|