base: Move StringSplitAlongWhitespace to string_split.h

base/string_split.cc

 // Copyright (c) 2010 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "base/string_split.h"
 
 #include "base/logging.h"
 #include "base/string_util.h"
 #include "base/third_party/icu/icu_utf.h"
 #include "base/utf_string_conversions.h"
@@ skipping 150 matching lines @@
 }
 
 void SplitStringDontTrim(const std::string& str,
                          char c,
                          std::vector<std::string>* r) {
   DCHECK(IsStringUTF8(str));
   DCHECK(c >= 0 && c < 0x7F);
   SplitStringT(str, c, false, r);
 }
 
+template<typename STR>
+void SplitStringAlongWhitespaceT(const STR& str, std::vector<STR>* result) {
+  const size_t length = str.length();
+  if (!length)
+    return;
+
+  bool last_was_ws = false;
+  size_t last_non_ws_start = 0;
+  for (size_t i = 0; i < length; ++i) {
+    switch (str[i]) {
+      // HTML 5 defines whitespace as: space, tab, LF, line tab, FF, or CR.
+      case L' ':
+      case L'\t':
+      case L'\xA':
+      case L'\xB':
+      case L'\xC':
+      case L'\xD':
+        if (!last_was_ws) {
+          if (i > 0) {
+            result->push_back(
+                str.substr(last_non_ws_start, i - last_non_ws_start));
+          }
+          last_was_ws = true;
+        }
+        break;
+
+      default:  // Not a space character.
+        if (last_was_ws) {
+          last_was_ws = false;
+          last_non_ws_start = i;
+        }
+        break;
+    }
+  }
+  if (!last_was_ws) {
+    result->push_back(
+        str.substr(last_non_ws_start, length - last_non_ws_start));
+  }
+}
+
+void SplitStringAlongWhitespace(const std::wstring& str,
+                                std::vector<std::wstring>* result) {
+  SplitStringAlongWhitespaceT(str, result);
+}
+
+#if !defined(WCHAR_T_IS_UTF16)
+void SplitStringAlongWhitespace(const string16& str,
+                                std::vector<string16>* result) {
+  SplitStringAlongWhitespaceT(str, result);
+}
+#endif
+
+void SplitStringAlongWhitespace(const std::string& str,
+                                std::vector<std::string>* result) {
+  SplitStringAlongWhitespaceT(str, result);
+}
+
 }  // namespace base