Move xfa/include/fxgraphics/fx_graphics.h to xfa/fxgraphics.

xfa/fxgraphics/cfx_shading.cpp

+// Copyright 2016 PDFium Authors. All rights reserved.
+// Use of this source code is governed by a BSD-style license that can be
+// found in the LICENSE file.
+
+// Original code copyright 2014 Foxit Software Inc. http://www.foxitsoftware.com
+
+#include "xfa/fxgraphics/cfx_shading.h"
+
+CFX_Shading::CFX_Shading(const CFX_PointF& beginPoint,
+                         const CFX_PointF& endPoint,
+                         FX_BOOL isExtendedBegin,
+                         FX_BOOL isExtendedEnd,
+                         const FX_ARGB beginArgb,
+                         const FX_ARGB endArgb)
+    : m_type(FX_SHADING_Axial),
+      m_beginPoint(beginPoint),
+      m_endPoint(endPoint),
+      m_isExtendedBegin(isExtendedBegin),
+      m_beginArgb(beginArgb),
+      m_endArgb(endArgb) {
+  InitArgbArray();
+}
+
+CFX_Shading::CFX_Shading(const CFX_PointF& beginPoint,
+                         const CFX_PointF& endPoint,
+                         const FX_FLOAT beginRadius,
+                         const FX_FLOAT endRadius,
+                         FX_BOOL isExtendedBegin,
+                         FX_BOOL isExtendedEnd,
+                         const FX_ARGB beginArgb,
+                         const FX_ARGB endArgb)
+    : m_type(FX_SHADING_Radial),
+      m_beginPoint(beginPoint),
+      m_endPoint(endPoint),
+      m_beginRadius(beginRadius),
+      m_endRadius(endRadius),
+      m_isExtendedBegin(isExtendedBegin),
+      m_isExtendedEnd(isExtendedEnd),
+      m_beginArgb(beginArgb),
+      m_endArgb(endArgb) {
+  InitArgbArray();
+}
+
+CFX_Shading::~CFX_Shading() {}
+
+void CFX_Shading::InitArgbArray() {
+  int32_t a1;
+  int32_t r1;
+  int32_t g1;
+  int32_t b1;
+  ArgbDecode(m_beginArgb, a1, r1, g1, b1);
+
+  int32_t a2;
+  int32_t r2;
+  int32_t g2;
+  int32_t b2;
+  ArgbDecode(m_endArgb, a2, r2, g2, b2);
+
+  FX_FLOAT f = (FX_FLOAT)(FX_SHADING_Steps - 1);
+  FX_FLOAT aScale = (FX_FLOAT)(1.0 * (a2 - a1) / f);

[Tom Sepez, 2016/03/16 20:53:18]

These casts are pointless as we assume an int's range fits in a float (modulo
loss of accurracy).

[dsinclair, 2016/03/16 21:13:39]

Done.

+  FX_FLOAT rScale = (FX_FLOAT)(1.0 * (r2 - r1) / f);
+  FX_FLOAT gScale = (FX_FLOAT)(1.0 * (g2 - g1) / f);
+  FX_FLOAT bScale = (FX_FLOAT)(1.0 * (b2 - b1) / f);
+
+  int32_t a3;

[Tom Sepez, 2016/03/16 20:53:18]

nit: move these inside loop.  They needn't persist across it.

[dsinclair, 2016/03/16 21:13:39]

Done.

+  int32_t r3;
+  int32_t g3;
+  int32_t b3;
+  for (int32_t i = 0; i < FX_SHADING_Steps; i++) {
+    a3 = (int32_t)(i * aScale);

[Tom Sepez, 2016/03/16 20:53:17]

nit: static_cast<>, just to emphasize that we think the float will fit.

[dsinclair, 2016/03/16 21:13:39]

Done.

+    r3 = (int32_t)(i * rScale);
+    g3 = (int32_t)(i * gScale);
+    b3 = (int32_t)(i * bScale);

[Tom Sepez, 2016/03/16 20:53:18]

Pity we dont define addition multiplicaton subtraction and division for argbs.  

[dsinclair, 2016/03/16 21:13:39]

Done.

+    m_argbArray[i] =
+        FXARGB_TODIB(FXARGB_MAKE((a1 + a3), (r1 + r3), (g1 + g3), (b1 + b3)));

[Tom Sepez, 2016/03/16 20:53:18]

nit: overparenthesized, unless FXARGB_MAKE is not careful with argument
expansion.

[dsinclair, 2016/03/16 21:13:39]

Done.

+  }
+}