michael@0: // Copyright (c) 2011 The Chromium Authors. All rights reserved.
michael@0: // Use of this source code is governed by a BSD-style license that can be
michael@0: // found in the LICENSE file.
michael@0: 
michael@0: #ifndef BASE_STRINGS_UTF_STRING_CONVERSION_UTILS_H_
michael@0: #define BASE_STRINGS_UTF_STRING_CONVERSION_UTILS_H_
michael@0: 
michael@0: // This should only be used by the various UTF string conversion files.
michael@0: 
michael@0: #include "base/base_export.h"
michael@0: #include "base/strings/string16.h"
michael@0: 
michael@0: namespace base {
michael@0: 
michael@0: inline bool IsValidCodepoint(uint32 code_point) {
michael@0:   // Excludes the surrogate code points ([0xD800, 0xDFFF]) and
michael@0:   // codepoints larger than 0x10FFFF (the highest codepoint allowed).
michael@0:   // Non-characters and unassigned codepoints are allowed.
michael@0:   return code_point < 0xD800u ||
michael@0:          (code_point >= 0xE000u && code_point <= 0x10FFFFu);
michael@0: }
michael@0: 
michael@0: inline bool IsValidCharacter(uint32 code_point) {
michael@0:   // Excludes non-characters (U+FDD0..U+FDEF, and all codepoints ending in
michael@0:   // 0xFFFE or 0xFFFF) from the set of valid code points.
michael@0:   return code_point < 0xD800u || (code_point >= 0xE000u &&
michael@0:       code_point < 0xFDD0u) || (code_point > 0xFDEFu &&
michael@0:       code_point <= 0x10FFFFu && (code_point & 0xFFFEu) != 0xFFFEu);
michael@0: }
michael@0: 
michael@0: // ReadUnicodeCharacter --------------------------------------------------------
michael@0: 
michael@0: // Reads a UTF-8 stream, placing the next code point into the given output
michael@0: // |*code_point|. |src| represents the entire string to read, and |*char_index|
michael@0: // is the character offset within the string to start reading at. |*char_index|
michael@0: // will be updated to index the last character read, such that incrementing it
michael@0: // (as in a for loop) will take the reader to the next character.
michael@0: //
michael@0: // Returns true on success. On false, |*code_point| will be invalid.
michael@0: BASE_EXPORT bool ReadUnicodeCharacter(const char* src,
michael@0:                                       int32 src_len,
michael@0:                                       int32* char_index,
michael@0:                                       uint32* code_point_out);
michael@0: 
michael@0: // Reads a UTF-16 character. The usage is the same as the 8-bit version above.
michael@0: BASE_EXPORT bool ReadUnicodeCharacter(const char16* src,
michael@0:                                       int32 src_len,
michael@0:                                       int32* char_index,
michael@0:                                       uint32* code_point);
michael@0: 
michael@0: #if defined(WCHAR_T_IS_UTF32)
michael@0: // Reads UTF-32 character. The usage is the same as the 8-bit version above.
michael@0: BASE_EXPORT bool ReadUnicodeCharacter(const wchar_t* src,
michael@0:                                       int32 src_len,
michael@0:                                       int32* char_index,
michael@0:                                       uint32* code_point);
michael@0: #endif  // defined(WCHAR_T_IS_UTF32)
michael@0: 
michael@0: // WriteUnicodeCharacter -------------------------------------------------------
michael@0: 
michael@0: // Appends a UTF-8 character to the given 8-bit string.  Returns the number of
michael@0: // bytes written.
michael@0: // TODO(brettw) Bug 79631: This function should not be exposed.
michael@0: BASE_EXPORT size_t WriteUnicodeCharacter(uint32 code_point,
michael@0:                                          std::string* output);
michael@0: 
michael@0: // Appends the given code point as a UTF-16 character to the given 16-bit
michael@0: // string.  Returns the number of 16-bit values written.
michael@0: BASE_EXPORT size_t WriteUnicodeCharacter(uint32 code_point, string16* output);
michael@0: 
michael@0: #if defined(WCHAR_T_IS_UTF32)
michael@0: // Appends the given UTF-32 character to the given 32-bit string.  Returns the
michael@0: // number of 32-bit values written.
michael@0: inline size_t WriteUnicodeCharacter(uint32 code_point, std::wstring* output) {
michael@0:   // This is the easy case, just append the character.
michael@0:   output->push_back(code_point);
michael@0:   return 1;
michael@0: }
michael@0: #endif  // defined(WCHAR_T_IS_UTF32)
michael@0: 
michael@0: // Generalized Unicode converter -----------------------------------------------
michael@0: 
michael@0: // Guesses the length of the output in UTF-8 in bytes, clears that output
michael@0: // string, and reserves that amount of space.  We assume that the input
michael@0: // character types are unsigned, which will be true for UTF-16 and -32 on our
michael@0: // systems.
michael@0: template<typename CHAR>
michael@0: void PrepareForUTF8Output(const CHAR* src, size_t src_len, std::string* output);
michael@0: 
michael@0: // Prepares an output buffer (containing either UTF-16 or -32 data) given some
michael@0: // UTF-8 input that will be converted to it.  See PrepareForUTF8Output().
michael@0: template<typename STRING>
michael@0: void PrepareForUTF16Or32Output(const char* src, size_t src_len, STRING* output);
michael@0: 
michael@0: }  // namespace base
michael@0: 
michael@0: #endif  // BASE_STRINGS_UTF_STRING_CONVERSION_UTILS_H_