warmup, remove clamping in append_gamut_transform()

Clamping here seems inconsistent with our color pipeline model,
and with the existing GPU impl.  The SkRasterPipeline store
stages already do clamp when storing unorms, and table-lookup
stages clamp their inputs, so it should be safe.

While refactoring, slim its interface down a bit.

Change-Id: I4772457fdf90e483834d034f02974d7a859cbe24
Reviewed-on: https://skia-review.googlesource.com/130902
Reviewed-by: Brian Osman <brianosman@google.com>
Commit-Queue: Mike Klein <mtklein@chromium.org>

1 files changed, 3 insertions, 3 deletions

diff --git a/src/core/SkConvertPixels.cpp b/src/core/SkConvertPixels.cpp
index 1eafe8a492..ddc04c06b6 100644
--- a/src/core/SkConvertPixels.cpp
+++ b/src/core/SkConvertPixels.cpp
@@ -302,10 +302,10 @@ static void convert_with_pipeline(const SkImageInfo& dstInfo, void* dstRow, size
         }
     }
 
-    float matrix[12];
+    SkSTArenaAlloc<12*sizeof(float)> alloc;
     if (isColorAware) {
-        append_gamut_transform(&pipeline, matrix, srcInfo.colorSpace(), dstInfo.colorSpace(),
-                               premulState);
+        append_gamut_transform(&pipeline, &alloc,
+                               srcInfo.colorSpace(), dstInfo.colorSpace(), premulState);
     }
 
     SkAlphaType dat = dstInfo.alphaType();