Extern in .cpp files is a code smell.

core/src/fpdftext/unicodenormalization.cpp

 // 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
 
 #include "../../include/fxcrt/fx_string.h"
+#include "unicodenormalizationdata.h"
 
-extern const FX_WCHAR g_UnicodeData_Normalization[];
-extern const FX_WCHAR g_UnicodeData_Normalization_Map1[];
-extern const FX_WCHAR g_UnicodeData_Normalization_Map2[];
-extern const FX_WCHAR g_UnicodeData_Normalization_Map3[];
-extern const FX_WCHAR g_UnicodeData_Normalization_Map4[];
-const FX_WCHAR* g_UnicodeData_Normalization_Maps[5] = {
-    NULL, g_UnicodeData_Normalization_Map1, g_UnicodeData_Normalization_Map2,
-    g_UnicodeData_Normalization_Map3, g_UnicodeData_Normalization_Map4};
+const FX_WCHAR* const g_UnicodeData_Normalization_Maps[5] = {
+    nullptr,
+    g_UnicodeData_Normalization_Map1,
+    g_UnicodeData_Normalization_Map2,
+    g_UnicodeData_Normalization_Map3,
+    g_UnicodeData_Normalization_Map4};

[Lei Zhang, 2015/08/18 22:24:19]

Is clang-format being weird? I think if you put a comma at the end of the line,
it will put the closing brace on the next line.

+
 FX_STRSIZE FX_Unicode_GetNormalization(FX_WCHAR wch, FX_WCHAR* pDst) {
   wch = wch & 0xFFFF;
   FX_WCHAR wFind = g_UnicodeData_Normalization[wch];
   if (!wFind) {
     if (pDst) {
       *pDst = wch;
     }
     return 1;
   }
   if (wFind >= 0x8000) {
@@ skipping 35 matching lines @@
                                           CFX_WideString& wsDst) {
   FX_STRSIZE nLen = FX_WideString_GetNormalization(wsSrc, (FX_WCHAR*)NULL);
   if (!nLen) {
     return 0;
   }
   FX_WCHAR* pBuf = wsDst.GetBuffer(nLen);
   FX_WideString_GetNormalization(wsSrc, pBuf);
   wsDst.ReleaseBuffer(nLen);
   return nLen;
 }