ceph/src/common/utf8.h

   1 // -*- mode:C++; tab-width:8; c-basic-offset:2; indent-tabs-mode:t -*-
   2 // vim: ts=8 sw=2 smarttab
   3 /*
   4  * Ceph - scalable distributed file system
   5  *
   6  * Copyright (C) 2011 New Dream Network
   7  *
   8  * This is free software; you can redistribute it and/or
   9  * modify it under the terms of the GNU Lesser General Public
  10  * License version 2.1, as published by the Free Software
  11  * Foundation.  See file COPYING.
  12  *
  13  */
  14
  15 #ifndef CEPH_COMMON_UTF8_H
  16 #define CEPH_COMMON_UTF8_H
  17
  18 #define MAX_UTF8_SZ 6
  19 #define INVALID_UTF8_CHAR 0xfffffffful
  20
  21 #ifdef __cplusplus
  22 extern "C" {
  23 #endif
  24
  25 /* Checks if a buffer is valid UTF-8.
  26  * Returns 0 if it is, and one plus the offset of the first invalid byte
  27  * if it is not.
  28  */
  29 int check_utf8(const char *buf, int len);
  30
  31 /* Checks if a null-terminated string is valid UTF-8.
  32  * Returns 0 if it is, and one plus the offset of the first invalid byte
  33  * if it is not.
  34  */
  35 int check_utf8_cstr(const char *buf);
  36
  37 /* Returns true if 'ch' is a control character.
  38  * We do count newline as a control character, but not NULL.
  39  */
  40 int is_control_character(int ch);
  41
  42 /* Checks if a buffer contains control characters.
  43  */
  44 int check_for_control_characters(const char *buf, int len);
  45
  46 /* Checks if a null-terminated string contains control characters.
  47  */
  48 int check_for_control_characters_cstr(const char *buf);
  49
  50 /* Encode a 31-bit UTF8 code point to 'buf'.
  51  * Assumes buf is of size MAX_UTF8_SZ
  52  * Returns -1 on failure; number of bytes in the encoded value otherwise.
  53  */
  54 int encode_utf8(unsigned long u, unsigned char *buf);
  55
  56 /*
  57  * Decode a UTF8 character from an array of bytes. Return character code.
  58  * Upon error, return INVALID_UTF8_CHAR.
  59  */
  60 unsigned long decode_utf8(unsigned char *buf, int nbytes);
  61
  62 #ifdef __cplusplus
  63 }
  64 #endif
  65
  66 #endif