1UTF-8(7) Linux Programmer's Manual UTF-8(7)
2
6 UTF-8 - an ASCII compatible multi-byte Unicode encoding
7
9 The Unicode 3.0 character set occupies a 16-bit code space. The most
10 obvious Unicode encoding (known as UCS-2) consists of a sequence of
11 16-bit words. Such strings can contain as parts of many 16-bit charac‐
12 ters bytes like '\0' or '/' which have a special meaning in filenames
13 and other C library function parameters. In addition, the majority of
14 UNIX tools expects ASCII files and can't read 16-bit words as charac‐
15 ters without major modifications. For these reasons, UCS-2 is not a
16 suitable external encoding of Unicode in filenames, text files, envi‐
17 ronment variables, etc. The ISO 10646 Universal Character Set (UCS), a
18 superset of Unicode, occupies even a 31-bit code space and the obvious
19 UCS-4 encoding for it (a sequence of 32-bit words) has the same prob‐
20 lems.
21 The UTF-8 encoding of Unicode and UCS does not have these problems and
23 is the common way in which Unicode is used on Unix-style operating sys‐
24 tems.
25
27 The UTF-8 encoding has the following nice properties:
28 * UCS characters 0x00000000 to 0x0000007f (the classic US-ASCII charac‐
30 ters) are encoded simply as bytes 0x00 to 0x7f (ASCII compatibility).
31 This means that files and strings which contain only 7-bit ASCII
32 characters have the same encoding under both ASCII and UTF-8.
33 * All UCS characters > 0x7f are encoded as a multi-byte sequence con‐
35 sisting only of bytes in the range 0x80 to 0xfd, so no ASCII byte can
36 appear as part of another character and there are no problems with
37 e.g. '\0' or '/'.
38 * The lexicographic sorting order of UCS-4 strings is preserved.
40 * All possible 2^31 UCS codes can be encoded using UTF-8.
42 * The bytes 0xfe and 0xff are never used in the UTF-8 encoding.
44 * The first byte of a multi-byte sequence which represents a single
46 non-ASCII UCS character is always in the range 0xc0 to 0xfd and indi‐
47 cates how long this multi-byte sequence is. All further bytes in a
48 multi-byte sequence are in the range 0x80 to 0xbf. This allows easy
49 resynchronization and makes the encoding stateless and robust against
50 missing bytes.
51 * UTF-8 encoded UCS characters may be up to six bytes long, however the
53 Unicode standard specifies no characters above 0x10ffff, so Unicode
54 characters can only be up to four bytes long in UTF-8.
55
57 The following byte sequences are used to represent a character. The
58 sequence to be used depends on the UCS code number of the character:
59 0x00000000 - 0x0000007F:
61 0xxxxxxx
62 0x00000080 - 0x000007FF:
64 110xxxxx 10xxxxxx
65 0x00000800 - 0x0000FFFF:
67 1110xxxx 10xxxxxx 10xxxxxx
68 0x00010000 - 0x001FFFFF:
70 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
71 0x00200000 - 0x03FFFFFF:
73 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
74 0x04000000 - 0x7FFFFFFF:
76 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
77 The xxx bit positions are filled with the bits of the character code
79 number in binary representation. Only the shortest possible multi-byte
80 sequence which can represent the code number of the character can be
81 used.
82 The UCS code values 0xd800–0xdfff (UTF-16 surrogates) as well as 0xfffe
84 and 0xffff (UCS non-characters) should not appear in conforming UTF-8
85 streams.
86
88 The Unicode character 0xa9 = 1010 1001 (the copyright sign) is encoded
89 in UTF-8 as
90 11000010 10101001 = 0xc2 0xa9
92 and character 0x2260 = 0010 0010 0110 0000 (the "not equal" symbol) is
94 encoded as:
95 11100010 10001001 10100000 = 0xe2 0x89 0xa0
97
99 Users have to select a UTF-8 locale, for example with
100 export LANG=en_GB.UTF-8
102 in order to activate the UTF-8 support in applications.
104 Application software that has to be aware of the used character encod‐
106 ing should always set the locale with for example
107 setlocale(LC_CTYPE, "")
109 and programmers can then test the expression
111 strcmp(nl_langinfo(CODESET), "UTF-8") == 0
113 to determine whether a UTF-8 locale has been selected and whether
115 therefore all plaintext standard input and output, terminal communica‐
116 tion, plaintext file content, filenames and environment variables are
117 encoded in UTF-8.
118 Programmers accustomed to single-byte encodings such as US-ASCII or ISO
120 8859 have to be aware that two assumptions made so far are no longer
121 valid in UTF-8 locales. Firstly, a single byte does not necessarily
122 correspond any more to a single character. Secondly, since modern ter‐
123 minal emulators in UTF-8 mode also support Chinese, Japanese, and
124 Korean double-width characters as well as non-spacing combining charac‐
125 ters, outputting a single character does not necessarily advance the
126 cursor by one position as it did in ASCII. Library functions such as
127 mbsrtowcs(3) and wcswidth(3) should be used today to count characters
128 and cursor positions.
129 The official ESC sequence to switch from an ISO 2022 encoding scheme
131 (as used for instance by VT100 terminals) to UTF-8 is ESC % G
132 ("\x1b%G"). The corresponding return sequence from UTF-8 to ISO 2022 is
133 ESC % @ ("\x1b%@"). Other ISO 2022 sequences (such as for switching the
134 G0 and G1 sets) are not applicable in UTF-8 mode.
135 It can be hoped that in the foreseeable future, UTF-8 will replace
137 ASCII and ISO 8859 at all levels as the common character encoding on
138 POSIX systems, leading to a significantly richer environment for han‐
139 dling plain text.
140
142 The Unicode and UCS standards require that producers of UTF-8 shall use
143 the shortest form possible, e.g., producing a two-byte sequence with
144 first byte 0xc0 is non-conforming. Unicode 3.1 has added the require‐
145 ment that conforming programs must not accept non-shortest forms in
146 their input. This is for security reasons: if user input is checked for
147 possible security violations, a program might check only for the ASCII
148 version of "/../" or ";" or NUL and overlook that there are many non-
149 ASCII ways to represent these things in a non-shortest UTF-8 encoding.
150
152 ISO/IEC 10646-1:2000, Unicode 3.1, RFC 2279, Plan 9.
153
155 Markus Kuhn <mgk25@cl.cam.ac.uk>
156
158 nl_langinfo(3), setlocale(3), charsets(7), unicode(7)
159GNU 2001-05-11 UTF-8(7)