Use new jpeg_crop_scanlines() API to optimize jpeg subset decodes

dm/DMSrcSink.h

 /*
  * Copyright 2015 Google Inc.
  *
  * Use of this source code is governed by a BSD-style license that can be
  * found in the LICENSE file.
  */
 
 #ifndef DMSrcSink_DEFINED
 #define DMSrcSink_DEFINED
 
@@ skipping 91 matching lines @@
 
 class CodecSrc : public Src {
 public:
     enum Mode {
         kCodec_Mode,
         // We choose to test only one mode with zero initialized memory.
         // This will exercise all of the interesting cases in SkSwizzler
         // without doubling the size of our test suite.
         kCodecZeroInit_Mode,
         kScanline_Mode,
-        kStripe_Mode, // Tests the skipping of scanlines
+        kTile_Mode, // Tests jpeg subset decoding optimizations

[scroggo, 2016/02/22 15:55:03]

"Tile" doesn't seem like it describes this. kPartialScanline? kCroppedScanline?

Also, have we potentially lost some useful coverage by not testing other decodes
in stripe mode?

[msarett, 2016/02/22 17:38:16]

You're right, I think I've gone too far.

My concern is that the kCroppedScanline test is quite slow, and there are
sporadic complaints about the codec tests running for too long.  But I think
it's ok to keep the stripe test and only run the kCroppedScanline test for
jpegs.

         kSubset_Mode, // For codecs that support subsets directly.
         kGen_Mode,    // Test SkCodecImageGenerator (includes YUV)
     };
     enum DstColorType {
         kGetFromCanvas_DstColorType,
         kIndex8_Always_DstColorType,
         kGrayscale_Always_DstColorType,
     };
     CodecSrc(Path, Mode, DstColorType, SkAlphaType, float);
 
@@ skipping 234 matching lines @@
 
 class ViaMojo : public Via {
 public:
     explicit ViaMojo(Sink* sink) : Via(sink) {}
     Error draw(const Src&, SkBitmap*, SkWStream*, SkString*) const override;
 };
 
 }  // namespace DM
 
 #endif//DMSrcSink_DEFINED