Clean up IFX_BidiChar

- Replace IFX_BidiChar with just CFX_BidiChar
- Document implementation
- Change out parameters to pointers
- Remove dead code
- Add an enum for bidi directions
- Move several externs to a header
- Add unit tests

R=tsepez@chromium.org

Review URL: https://codereview.chromium.org/1197643002 .

3 files changed, 88 insertions, 109 deletions

diff --git a/core/include/fxcrt/fx_arb.h b/core/include/fxcrt/fx_arb.h
deleted file mode 100644
index 0f6de64b32..0000000000
--- a/core/include/fxcrt/fx_arb.h
+++ /dev/null
@@ -1,24 +0,0 @@
-// Copyright 2014 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
-
-#ifndef CORE_INCLUDE_FXCRT_FX_ARB_H_
-#define CORE_INCLUDE_FXCRT_FX_ARB_H_
-
-#include "fx_system.h"
-
-class IFX_BidiChar {
- public:
-  static IFX_BidiChar* Create();
-  virtual ~IFX_BidiChar() {}
-
-  virtual void SetPolicy(FX_BOOL bSeparateNeutral = TRUE) = 0;
-  virtual FX_BOOL AppendChar(FX_WCHAR wch) = 0;
-  virtual FX_BOOL EndChar() = 0;
-  virtual int32_t GetBidiInfo(int32_t& iStart, int32_t& iCount) = 0;
-  virtual void Reset() = 0;
-};
-
-#endif  // CORE_INCLUDE_FXCRT_FX_ARB_H_
diff --git a/core/include/fxcrt/fx_bidi.h b/core/include/fxcrt/fx_bidi.h
new file mode 100644
index 0000000000..a55ce6cfd2
--- /dev/null
+++ b/core/include/fxcrt/fx_bidi.h
@@ -0,0 +1,59 @@
+// Copyright 2014 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
+
+#ifndef CORE_INCLUDE_FXCRT_FX_BIDI_H_
+#define CORE_INCLUDE_FXCRT_FX_BIDI_H_
+
+#include "fx_system.h"
+
+// Processes characters and group them into segments based on text direction.
+class CFX_BidiChar {
+ public:
+  enum Direction { NEUTRAL, LEFT, RIGHT };
+
+  CFX_BidiChar();
+  ~CFX_BidiChar();
+
+  // Append a character and classify it as left, right, or neutral.
+  // Returns true if the character has a different direction than the
+  // existing direction to indicate there is a segment to process.
+  bool AppendChar(FX_WCHAR wch);
+
+  // Call this after the last character has been appended. AppendChar()
+  // must not be called after this.
+  // Returns true if there is still a segment to process.
+  bool EndChar();
+
+  // Get information about the segment to process.
+  // The segment's start position and character count is returned in |iStart|
+  // and |iCount|, respectively. Pass in null pointers if the information is
+  // not needed.
+  // Returns the segment direction.
+  Direction GetBidiInfo(int32_t* iStart, int32_t* iCount) const;
+
+ private:
+  void SaveCurrentStateToLastState();
+
+  // Position of the current segment.
+  int32_t m_iCurStart;
+
+  // Number of characters in the current segment.
+  int32_t m_iCurCount;
+
+  // Direction of the current segment.
+  Direction m_CurBidi;
+
+  // Number of characters in the last segment.
+  int32_t m_iLastStart;
+
+  // Number of characters in the last segment.
+  int32_t m_iLastCount;
+
+  // Direction of the last segment.
+  Direction m_LastBidi;
+};
+
+#endif  // CORE_INCLUDE_FXCRT_FX_BIDI_H_
diff --git a/core/include/fxcrt/fx_ucd.h b/core/include/fxcrt/fx_ucd.h
index 8bc5930dd4..62c2dfdf67 100644
--- a/core/include/fxcrt/fx_ucd.h
+++ b/core/include/fxcrt/fx_ucd.h
@@ -9,97 +9,41 @@
 
 #include "fx_system.h"
 
-enum FX_CHARBREAKPROP {
-  FX_CBP_OP = 0,
-  FX_CBP_CL = 1,
-  FX_CBP_QU = 2,
-  FX_CBP_GL = 3,
-  FX_CBP_NS = 4,
-  FX_CBP_EX = 5,
-  FX_CBP_SY = 6,
-  FX_CBP_IS = 7,
-  FX_CBP_PR = 8,
-  FX_CBP_PO = 9,
-  FX_CBP_NU = 10,
-  FX_CBP_AL = 11,
-  FX_CBP_ID = 12,
-  FX_CBP_IN = 13,
-  FX_CBP_HY = 14,
-  FX_CBP_BA = 15,
-  FX_CBP_BB = 16,
-  FX_CBP_B2 = 17,
-  FX_CBP_ZW = 18,
-  FX_CBP_CM = 19,
-  FX_CBP_WJ = 20,
-  FX_CBP_H2 = 21,
-  FX_CBP_H3 = 22,
-  FX_CBP_JL = 23,
-  FX_CBP_JV = 24,
-  FX_CBP_JT = 25,
-
-  FX_CBP_BK = 26,
-  FX_CBP_CR = 27,
-  FX_CBP_LF = 28,
-  FX_CBP_NL = 29,
-  FX_CBP_SA = 30,
-  FX_CBP_SG = 31,
-  FX_CBP_CB = 32,
-  FX_CBP_XX = 33,
-  FX_CBP_AI = 34,
-  FX_CBP_SP = 35,
-  FX_CBP_TB = 37,
-  FX_CBP_NONE = 36,
-};
 #define FX_BIDICLASSBITS 6
 #define FX_BIDICLASSBITSMASK (31 << FX_BIDICLASSBITS)
 enum FX_BIDICLASS {
-  FX_BIDICLASS_ON = 0,
-  FX_BIDICLASS_L = 1,
-  FX_BIDICLASS_R = 2,
-  FX_BIDICLASS_AN = 3,
-  FX_BIDICLASS_EN = 4,
-  FX_BIDICLASS_AL = 5,
-  FX_BIDICLASS_NSM = 6,
-  FX_BIDICLASS_CS = 7,
-  FX_BIDICLASS_ES = 8,
-  FX_BIDICLASS_ET = 9,
-  FX_BIDICLASS_BN = 10,
-  FX_BIDICLASS_S = 11,
-  FX_BIDICLASS_WS = 12,
-  FX_BIDICLASS_B = 13,
-  FX_BIDICLASS_RLO = 14,
-  FX_BIDICLASS_RLE = 15,
-  FX_BIDICLASS_LRO = 16,
-  FX_BIDICLASS_LRE = 17,
-  FX_BIDICLASS_PDF = 18,
+  FX_BIDICLASS_ON = 0,    // Other Neutral
+  FX_BIDICLASS_L = 1,     // Left Letter
+  FX_BIDICLASS_R = 2,     // Right Letter
+  FX_BIDICLASS_AN = 3,    // Arabic Number
+  FX_BIDICLASS_EN = 4,    // European Number
+  FX_BIDICLASS_AL = 5,    // Arabic Letter
+  FX_BIDICLASS_NSM = 6,   // Non-spacing Mark
+  FX_BIDICLASS_CS = 7,    // Common Number Separator
+  FX_BIDICLASS_ES = 8,    // European Separator
+  FX_BIDICLASS_ET = 9,    // European Number Terminator
+  FX_BIDICLASS_BN = 10,   // Boundary Neutral
+  FX_BIDICLASS_S = 11,    // Segment Separator
+  FX_BIDICLASS_WS = 12,   // Whitespace
+  FX_BIDICLASS_B = 13,    // Paragraph Separator
+  FX_BIDICLASS_RLO = 14,  // Right-to-Left Override
+  FX_BIDICLASS_RLE = 15,  // Right-to-Left Embedding
+  FX_BIDICLASS_LRO = 16,  // Left-to-Right Override
+  FX_BIDICLASS_LRE = 17,  // Left-to-Right Embedding
+  FX_BIDICLASS_PDF = 18,  // Pop Directional Format
   FX_BIDICLASS_N = FX_BIDICLASS_ON,
 };
-#define FX_CHARTYPEBITS 11
-#define FX_CHARTYPEBITSMASK (15 << FX_CHARTYPEBITS)
-enum FX_CHARTYPE {
-  FX_CHARTYPE_Unknown = 0,
-  FX_CHARTYPE_Tab = (1 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_Space = (2 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_Control = (3 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_Combination = (4 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_Numeric = (5 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_Normal = (6 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_ArabicAlef = (7 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_ArabicSpecial = (8 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_ArabicDistortion = (9 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_ArabicNormal = (10 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_ArabicForm = (11 << FX_CHARTYPEBITS),
-  FX_CHARTYPE_Arabic = (12 << FX_CHARTYPEBITS),
-};
+
+extern const FX_DWORD kTextLayoutCodeProperties[];
+extern const size_t kTextLayoutCodePropertiesSize;
+
+extern const FX_WCHAR kFXTextLayoutVerticalMirror[];
+extern const size_t kFXTextLayoutVerticalMirrorSize;
+
+extern const FX_WCHAR kFXTextLayoutBidiMirror[];
+extern const size_t kFXTextLayoutBidiMirrorSize;
+
 FX_DWORD FX_GetUnicodeProperties(FX_WCHAR wch);
-FX_BOOL FX_IsCtrlCode(FX_WCHAR ch);
-FX_BOOL FX_IsRotationCode(FX_WCHAR ch);
-FX_BOOL FX_IsCombinationChar(FX_WCHAR wch);
-FX_BOOL FX_IsBidiChar(FX_WCHAR wch);
 FX_WCHAR FX_GetMirrorChar(FX_WCHAR wch, FX_BOOL bRTL, FX_BOOL bVertical);
-FX_WCHAR FX_GetMirrorChar(FX_WCHAR wch,
-                          FX_DWORD dwProps,
-                          FX_BOOL bRTL,
-                          FX_BOOL bVertical);
 
 #endif  // CORE_INCLUDE_FXCRT_FX_UCD_H_