man7/utf-8.7

   1 .\" Copyright (C) Markus Kuhn, 1996, 2001
   2 .\"
   3 .\" %%%LICENSE_START(GPLv2+_DOC_FULL)
   4 .\" This is free documentation; you can redistribute it and/or
   5 .\" modify it under the terms of the GNU General Public License as
   6 .\" published by the Free Software Foundation; either version 2 of
   7 .\" the License, or (at your option) any later version.
   8 .\"
   9 .\" The GNU General Public License's references to "object code"
  10 .\" and "executables" are to be interpreted as the output of any
  11 .\" document formatting or typesetting system, including
  12 .\" intermediate and printed output.
  13 .\"
  14 .\" This manual is distributed in the hope that it will be useful,
  15 .\" but WITHOUT ANY WARRANTY; without even the implied warranty of
  16 .\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  17 .\" GNU General Public License for more details.
  18 .\"
  19 .\" You should have received a copy of the GNU General Public
  20 .\" License along with this manual; if not, see
  21 .\" <http://www.gnu.org/licenses/>.
  22 .\" %%%LICENSE_END
  23 .\"
  24 .\" 1995-11-26  Markus Kuhn <mskuhn@cip.informatik.uni-erlangen.de>
  25 .\"      First version written
  26 .\" 2001-05-11  Markus Kuhn <mgk25@cl.cam.ac.uk>
  27 .\"      Update
  28 .\"
  29 .TH UTF-8 7 2019-03-06 "GNU" "Linux Programmer's Manual"
  30 .SH NAME
  31 UTF-8 \- an ASCII compatible multibyte Unicode encoding
  32 .SH DESCRIPTION
  33 The Unicode 3.0 character set occupies a 16-bit code space.
  34 The most obvious
  35 Unicode encoding (known as UCS-2)
  36 consists of a sequence of 16-bit words.
  37 Such strings can contain\(emas part of many 16-bit characters\(embytes
  38 such as \(aq\e0\(aq or \(aq/\(aq, which have a
  39 special meaning in filenames and other C library function arguments.
  40 In addition, the majority of UNIX tools expect ASCII files and can't
  41 read 16-bit words as characters without major modifications.
  42 For these reasons,
  43 UCS-2 is not a suitable external encoding of Unicode
  44 in filenames, text files, environment variables, and so on.
  45 The ISO 10646 Universal Character Set (UCS),
  46 a superset of Unicode, occupies an even larger code
  47 space\(em31\ bits\(emand the obvious
  48 UCS-4 encoding for it (a sequence of 32-bit words) has the same problems.
  49 .PP
  50 The UTF-8 encoding of Unicode and UCS
  51 does not have these problems and is the common way in which
  52 Unicode is used on UNIX-style operating systems.
  53 .SS Properties
  54 The UTF-8 encoding has the following nice properties:
  55 .TP 0.2i
  56 *
  57 UCS
  58 characters 0x00000000 to 0x0000007f (the classic US-ASCII
  59 characters) are encoded simply as bytes 0x00 to 0x7f (ASCII
  60 compatibility).
  61 This means that files and strings which contain only
  62 7-bit ASCII characters have the same encoding under both
  63 ASCII
  64 and
  65 UTF-8 .
  66 .TP
  67 *
  68 All UCS characters greater than 0x7f are encoded as a multibyte sequence
  69 consisting only of bytes in the range 0x80 to 0xfd, so no ASCII
  70 byte can appear as part of another character and there are no
  71 problems with, for example,  \(aq\e0\(aq or \(aq/\(aq.
  72 .TP
  73 *
  74 The lexicographic sorting order of UCS-4 strings is preserved.
  75 .TP
  76 *
  77 All possible 2^31 UCS codes can be encoded using UTF-8.
  78 .TP
  79 *
  80 The bytes 0xc0, 0xc1, 0xfe, and 0xff are never used in the UTF-8 encoding.
  81 .TP
  82 *
  83 The first byte of a multibyte sequence which represents a single non-ASCII
  84 UCS character is always in the range 0xc2 to 0xfd and indicates how long
  85 this multibyte sequence is.
  86 All further bytes in a multibyte sequence
  87 are in the range 0x80 to 0xbf.
  88 This allows easy resynchronization and
  89 makes the encoding stateless and robust against missing bytes.
  90 .TP
  91 *
  92 UTF-8 encoded UCS characters may be up to six bytes long, however the
  93 Unicode standard specifies no characters above 0x10ffff, so Unicode characters
  94 can be only up to four bytes long in
  95 UTF-8.
  96 .SS Encoding
  97 The following byte sequences are used to represent a character.
  98 The sequence to be used depends on the UCS code number of the character:
  99 .TP 0.4i
 100 0x00000000 \- 0x0000007F:
 101 .RI 0 xxxxxxx
 102 .TP
 103 0x00000080 \- 0x000007FF:
 104 .RI 110 xxxxx
 105 .RI 10 xxxxxx
 106 .TP
 107 0x00000800 \- 0x0000FFFF:
 108 .RI 1110 xxxx
 109 .RI 10 xxxxxx
 110 .RI 10 xxxxxx
 111 .TP
 112 0x00010000 \- 0x001FFFFF:
 113 .RI 11110 xxx
 114 .RI 10 xxxxxx
 115 .RI 10 xxxxxx
 116 .RI 10 xxxxxx
 117 .TP
 118 0x00200000 \- 0x03FFFFFF:
 119 .RI 111110 xx
 120 .RI 10 xxxxxx
 121 .RI 10 xxxxxx
 122 .RI 10 xxxxxx
 123 .RI 10 xxxxxx
 124 .TP
 125 0x04000000 \- 0x7FFFFFFF:
 126 .RI 1111110 x
 127 .RI 10 xxxxxx
 128 .RI 10 xxxxxx
 129 .RI 10 xxxxxx
 130 .RI 10 xxxxxx
 131 .RI 10 xxxxxx
 132 .PP
 133 The
 134 .I xxx
 135 bit positions are filled with the bits of the character code number in
 136 binary representation, most significant bit first (big-endian).
 137 Only the shortest possible multibyte sequence
 138 which can represent the code number of the character can be used.
 139 .PP
 140 The UCS code values 0xd800\(en0xdfff (UTF-16 surrogates) as well as 0xfffe and
 141 0xffff (UCS noncharacters) should not appear in conforming UTF-8 streams. According
 142 to RFC 3629 no point above U+10FFFF should be used, which limits characters to four
 143 bytes.
 144 .SS Example
 145 The Unicode character 0xa9 = 1010 1001 (the copyright sign) is encoded
 146 in UTF-8 as
 147 .PP
 148 .RS
 149 11000010 10101001 = 0xc2 0xa9
 150 .RE
 151 .PP
 152 and character 0x2260 = 0010 0010 0110 0000 (the "not equal" symbol) is
 153 encoded as:
 154 .PP
 155 .RS
 156 11100010 10001001 10100000 = 0xe2 0x89 0xa0
 157 .RE
 158 .SS Application notes
 159 Users have to select a UTF-8 locale, for example with
 160 .PP
 161 .RS
 162 export LANG=en_GB.UTF-8
 163 .RE
 164 .PP
 165 in order to activate the UTF-8 support in applications.
 166 .PP
 167 Application software that has to be aware of the used character
 168 encoding should always set the locale with for example
 169 .PP
 170 .RS
 171 setlocale(LC_CTYPE, "")
 172 .RE
 173 .PP
 174 and programmers can then test the expression
 175 .PP
 176 .RS
 177 strcmp(nl_langinfo(CODESET), "UTF-8") == 0
 178 .RE
 179 .PP
 180 to determine whether a UTF-8 locale has been selected and whether
 181 therefore all plaintext standard input and output, terminal
 182 communication, plaintext file content, filenames and environment
 183 variables are encoded in UTF-8.
 184 .PP
 185 Programmers accustomed to single-byte encodings such as US-ASCII or ISO 8859
 186 have to be aware that two assumptions made so far are no longer valid
 187 in UTF-8 locales.
 188 Firstly, a single byte does not necessarily correspond any
 189 more to a single character.
 190 Secondly, since modern terminal emulators in UTF-8
 191 mode also support Chinese, Japanese, and Korean
 192 double-width characters as well as nonspacing combining characters,
 193 outputting a single character does not necessarily advance the cursor
 194 by one position as it did in ASCII.
 195 Library functions such as
 196 .BR mbsrtowcs (3)
 197 and
 198 .BR wcswidth (3)
 199 should be used today to count characters and cursor positions.
 200 .PP
 201 The official ESC sequence to switch from an ISO 2022
 202 encoding scheme (as used for instance by VT100 terminals) to
 203 UTF-8 is ESC % G
 204 ("\ex1b%G").
 205 The corresponding return sequence from
 206 UTF-8 to ISO 2022 is ESC % @ ("\ex1b%@").
 207 Other ISO 2022 sequences (such as
 208 for switching the G0 and G1 sets) are not applicable in UTF-8 mode.
 209 .SS Security
 210 The Unicode and UCS standards require that producers of UTF-8
 211 shall use the shortest form possible, for example, producing a two-byte
 212 sequence with first byte 0xc0 is nonconforming.
 213 Unicode 3.1 has added the requirement that conforming programs must not accept
 214 non-shortest forms in their input.
 215 This is for security reasons: if
 216 user input is checked for possible security violations, a program
 217 might check only for the ASCII
 218 version of "/../" or ";" or NUL and overlook that there are many
 219 non-ASCII ways to represent these things in a non-shortest UTF-8
 220 encoding.
 221 .SS Standards
 222 ISO/IEC 10646-1:2000, Unicode 3.1, RFC\ 3629, Plan 9.
 223 .\" .SH AUTHOR
 224 .\" Markus Kuhn <mgk25@cl.cam.ac.uk>
 225 .SH SEE ALSO
 226 .BR locale (1),
 227 .BR nl_langinfo (3),
 228 .BR setlocale (3),
 229 .BR charsets (7),
 230 .BR unicode (7)