Implement transferToImageBitmap() in WebGLRenderingContext

third_party/WebKit/Source/modules/webgl/WebGLRenderingContextBase.cpp

 /*
  * Copyright (C) 2009 Apple Inc. All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer.
  * 2. Redistributions in binary form must reproduce the above copyright
  *    notice, this list of conditions and the following disclaimer in the
@@ skipping 615 matching lines @@
 PassOwnPtr<WebGraphicsContext3DProvider> WebGLRenderingContextBase::createWebGraphicsContext3DProvider(ScriptState* scriptState, WebGLContextAttributes attributes, unsigned webGLVersion)
 {
     return createContextProviderInternal(nullptr, scriptState, attributes, webGLVersion);
 }
 
 void WebGLRenderingContextBase::forceNextWebGLContextCreationToFail()
 {
     shouldFailContextCreationForTesting = true;
 }
 
+ImageBitmap* WebGLRenderingContextBase::transferToImageBitmapBase()
+{
+    ImageData* imageData = paintRenderingResultsToImageData(BackBuffer);

[Ken Russell (switch to Gerrit), 2016/05/10 23:11:42]

This needs to be much better optimized in the future. This operation should
transfer the DrawingBuffer's texture into a texture mailbox, retrieve that
texture into a new texture object (presumably), and create an ImageBitmap from
that texture ID. Is this how the 2D canvas context works, or will work? Is there
a plan for this? Please, minimally, add a TODO to this that this needs to avoid
readbacks from the GPU, which is how the canvas is getting the data into an
ImageData object.

[xidachen, 2016/05/11 00:17:52]

In the case of 2d canvas context, it contains a ImageBuffer private member, and
creating an ImageBitmap from a ImageBuffer is easy. I will dig into your
suggestion here. If it is not hard to get a new texture object and create an
ImageBitmap, then I will do that in this CL. I agree with you, accessing the
entire image data and creating from that is really in-efficient.

+    if (!imageData)
+        return nullptr;
+    return ImageBitmap::create(imageData, IntRect(0, 0, imageData->width(), imageData->height()));
+}
+
 namespace {
 
 // ES2 enums
 static const GLenum kSupportedInternalFormatsES2[] = {
     GL_RGB,
     GL_RGBA,
     GL_LUMINANCE_ALPHA,
     GL_LUMINANCE,
     GL_ALPHA,
 };
@@ skipping 511 matching lines @@
     m_drawingBuffer.clear();
 }
 
 void WebGLRenderingContextBase::markContextChanged(ContentChangeType changeType)
 {
     if (m_framebufferBinding || isContextLost())
         return;
 
     drawingBuffer()->markContentsChanged();
 
+    if (!canvas())
+        return;
+
     LayoutBox* layoutBox = canvas()->layoutBox();
     if (layoutBox && layoutBox->hasAcceleratedCompositing()) {
         m_markedCanvasDirty = true;
         canvas()->clearCopiedImage();
         layoutBox->contentChanged(changeType);
     } else {
         if (!m_markedCanvasDirty) {
             m_markedCanvasDirty = true;
             canvas()->didDraw(FloatRect(FloatPoint(0, 0), FloatSize(clampedCanvasSize())));
         }
@@ skipping 5184 matching lines @@
         contextGL()->PixelStorei(GL_UNPACK_ALIGNMENT, 1);
 }
 
 void WebGLRenderingContextBase::restoreUnpackParameters()
 {
     if (m_unpackAlignment != 1)
         contextGL()->PixelStorei(GL_UNPACK_ALIGNMENT, m_unpackAlignment);
 }
 
 } // namespace blink