Pull up clearCacheExceptFrame to ImageDecoder.

third_party/WebKit/Source/platform/image-decoders/ImageDecoder.h

Index: third_party/WebKit/Source/platform/image-decoders/ImageDecoder.h
diff --git a/third_party/WebKit/Source/platform/image-decoders/ImageDecoder.h b/third_party/WebKit/Source/platform/image-decoders/ImageDecoder.h
index b8b87a52a3eed2acce651642937d42664e2d7cf9..a35ae005ad3b200d0d00cc577b3101971ebe7ae6 100644
--- a/third_party/WebKit/Source/platform/image-decoders/ImageDecoder.h
+++ b/third_party/WebKit/Source/platform/image-decoders/ImageDecoder.h
@@ -242,10 +242,15 @@ class PLATFORM_EXPORT ImageDecoder {
 
   bool failed() const { return m_failed; }
 
-  // Clears decoded pixel data from all frames except the provided frame.
+  // Clears decoded pixel data from all frames except the provided frame. If
+  // subsequent frames depend on this frame's required previous frame, then that
+  // frame is also kept in cache to prevent re-decoding from the beginning.
   // Callers may pass WTF::kNotFound to clear all frames.
   // Note: If |m_frameBufferCache| contains only one frame, it won't be cleared.
   // Returns the number of bytes of frame data actually cleared.
+  //
+  // This is a virtual method because MockImageDecoder needs to override it in
+  // order to run the test ImageFrameGenerator::clearMultiFrameDecode.

[scroggo_chromium, 2016/12/01 14:44:07]

Instead of making this virtual, how about making MockImageDecoder override
clearFrameBuffer? MockImageDecoder will have to be a little more complicated;
instead of reporting the index passed to clearCacheExceptFrame, you'll need to
keep track of all the frames that were cleared. But that's probably what we care
about more anyway.

[joostouwerling, 2016/12/01 15:05:06]

Sgtm but I propose to do that in a different CL. In that way, we can keep this
CL a pure refactor and change the behavior of the clearMultiFrameDecode test in
another CL.

[scroggo_chromium, 2016/12/01 15:08:00]

Okay by me, but can you add a TODO? The danger in leaving it to the next CL is
that we might forget about it.

[joostouwerling, 2016/12/01 15:25:21]

Done.

   virtual size_t clearCacheExceptFrame(size_t);
 
   // If the image has a cursor hot-spot, stores it in the argument
@@ -351,6 +356,26 @@ class PLATFORM_EXPORT ImageDecoder {
   // future decodes to purge old frames as it goes.
   void updateAggressivePurging(size_t index);
 
+  // The method is only relevant for multi-frame images.
+  //
+  // This method indicates whether the provided frame has enough data to decode
+  // successive frames that depend on it. It is used by clearCacheExceptFrame
+  // to determine which frame to keep in cache when the indicated frame is not
+  // yet sufficiently decoded.
+  //
+  // The default condition is that the frame status needs to be FramePartial or
+  // FrameComplete, since the data of previous frames is copied in
+  // initFrameBuffer() before setting the status to FramePartial. For WebP,
+  // however, the status needs to be FrameComplete since the complete buffer is
+  // used to do alpha blending in WEBPImageDecoder::applyPostProcessing().
+  //
+  // Before calling this, verify that frame |index| exists by checking that
+  // |index| is smaller than |m_frameBufferCache|.size().
+  virtual bool frameStatusSufficientForSuccessors(size_t index) {
+    DCHECK(index < m_frameBufferCache.size());
+    return m_frameBufferCache[index].getStatus() != ImageFrame::FrameEmpty;
+  }
+
  private:
   enum class SniffResult { JPEG, PNG, GIF, WEBP, ICO, BMP, Invalid };