.\" Copyright (C) Markus Kuhn, 1996, 2001

.\"

.\" %%%LICENSE_START(GPLv2+_DOC_FULL)

.\" This is free documentation; you can redistribute it and/or

.\" modify it under the terms of the GNU General Public License as

.\" published by the Free Software Foundation; either version 2 of

.\" the License, or (at your option) any later version.

.\"

.\" The GNU General Public License's references to "object code"

.\" and "executables" are to be interpreted as the output of any

.\" document formatting or typesetting system, including

.\" intermediate and printed output.

.\"

.\" This manual is distributed in the hope that it will be useful,

.\" but WITHOUT ANY WARRANTY; without even the implied warranty of

.\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the

.\" GNU General Public License for more details.

.\"

.\" You should have received a copy of the GNU General Public

.\" License along with this manual; if not, see

.\" <http://www.gnu.org/licenses/>.

.\" %%%LICENSE_END

.\"

.\" 1995-11-26  Markus Kuhn <mskuhn@cip.informatik.uni-erlangen.de>

.\"      First version written

.\" 2001-05-11  Markus Kuhn <mgk25@cl.cam.ac.uk>

.\"      Update

.\"

.TH UTF-8 7 2016-07-17 "GNU" "Linux Programmer's Manual"

.SH NAME

UTF-8 \- an ASCII compatible multibyte Unicode encoding

.SH DESCRIPTION

The Unicode 3.0 character set occupies a 16-bit code space.

The most obvious

Unicode encoding (known as UCS-2)

consists of a sequence of 16-bit words.

Such strings can contain\(emas part of many 16-bit characters\(embytes

such as \(aq\\0\(aq or \(aq/\(aq, which have a

special meaning in filenames and other C library function arguments.

In addition, the majority of UNIX tools expect ASCII files and can't

read 16-bit words as characters without major modifications.

For these reasons,

UCS-2 is not a suitable external encoding of Unicode

in filenames, text files, environment variables, and so on.

The ISO 10646 Universal Character Set (UCS),

a superset of Unicode, occupies an even larger code

space\(em31\ bits\(emand the obvious

UCS-4 encoding for it (a sequence of 32-bit words) has the same problems.

.PP

The UTF-8 encoding of Unicode and UCS

does not have these problems and is the common way in which

Unicode is used on UNIX-style operating systems.

.SS Properties

The UTF-8 encoding has the following nice properties:

.TP 0.2i

*

UCS

characters 0x00000000 to 0x0000007f (the classic US-ASCII

characters) are encoded simply as bytes 0x00 to 0x7f (ASCII

compatibility).

This means that files and strings which contain only

7-bit ASCII characters have the same encoding under both

ASCII

and

UTF-8 .

.TP

*

All UCS characters greater than 0x7f are encoded as a multibyte sequence

consisting only of bytes in the range 0x80 to 0xfd, so no ASCII

byte can appear as part of another character and there are no

problems with, for example,  \(aq\\0\(aq or \(aq/\(aq.

.TP

*

The lexicographic sorting order of UCS-4 strings is preserved.

.TP

*

All possible 2^31 UCS codes can be encoded using UTF-8.

.TP

*

The bytes 0xc0, 0xc1, 0xfe, and 0xff are never used in the UTF-8 encoding.

.TP

*

The first byte of a multibyte sequence which represents a single non-ASCII

UCS character is always in the range 0xc2 to 0xfd and indicates how long

this multibyte sequence is.

All further bytes in a multibyte sequence

are in the range 0x80 to 0xbf.

This allows easy resynchronization and

makes the encoding stateless and robust against missing bytes.

.TP

*

UTF-8 encoded UCS characters may be up to six bytes long, however the

Unicode standard specifies no characters above 0x10ffff, so Unicode characters

can be only up to four bytes long in

UTF-8.

.SS Encoding

The following byte sequences are used to represent a character.

The sequence to be used depends on the UCS code number of the character:

.TP 0.4i

0x00000000 \- 0x0000007F:

.RI 0 xxxxxxx

.TP

0x00000080 \- 0x000007FF:

.RI 110 xxxxx

.RI 10 xxxxxx

.TP

0x00000800 \- 0x0000FFFF:

.RI 1110 xxxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.TP

0x00010000 \- 0x001FFFFF:

.RI 11110 xxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.TP

0x00200000 \- 0x03FFFFFF:

.RI 111110 xx

.RI 10 xxxxxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.TP

0x04000000 \- 0x7FFFFFFF:

.RI 1111110 x

.RI 10 xxxxxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.RI 10 xxxxxx

.PP

The

.I xxx

bit positions are filled with the bits of the character code number in

binary representation, most significant bit first (big-endian).

Only the shortest possible multibyte sequence

which can represent the code number of the character can be used.

.PP

The UCS code values 0xd800\(en0xdfff (UTF-16 surrogates) as well as 0xfffe and

0xffff (UCS noncharacters) should not appear in conforming UTF-8 streams. According

to RFC 3629 no point above U+10FFFF should be used, which limits characters to four

bytes.

.SS Example

The Unicode character 0xa9 = 1010 1001 (the copyright sign) is encoded

in UTF-8 as

.PP

.RS

11000010 10101001 = 0xc2 0xa9

.RE

.PP

and character 0x2260 = 0010 0010 0110 0000 (the "not equal" symbol) is

encoded as:

.PP

.RS

11100010 10001001 10100000 = 0xe2 0x89 0xa0

.RE

.SS Application notes

Users have to select a UTF-8 locale, for example with

.PP

.RS

export LANG=en_GB.UTF-8

.RE

.PP

in order to activate the UTF-8 support in applications.

.PP

Application software that has to be aware of the used character

encoding should always set the locale with for example

.PP

.RS

setlocale(LC_CTYPE, "")

.RE

.PP

and programmers can then test the expression

.PP

.RS

strcmp(nl_langinfo(CODESET), "UTF-8") == 0

.RE

.PP

to determine whether a UTF-8 locale has been selected and whether

therefore all plaintext standard input and output, terminal

communication, plaintext file content, filenames and environment

variables are encoded in UTF-8.

.PP

Programmers accustomed to single-byte encodings such as US-ASCII or ISO 8859

have to be aware that two assumptions made so far are no longer valid

in UTF-8 locales.

Firstly, a single byte does not necessarily correspond any

more to a single character.

Secondly, since modern terminal emulators in UTF-8

mode also support Chinese, Japanese, and Korean

double-width characters as well as nonspacing combining characters,

outputting a single character does not necessarily advance the cursor

by one position as it did in ASCII.

Library functions such as

.BR mbsrtowcs (3)

and

.BR wcswidth (3)

should be used today to count characters and cursor positions.

.PP

The official ESC sequence to switch from an ISO 2022

encoding scheme (as used for instance by VT100 terminals) to

UTF-8 is ESC % G

("\\x1b%G").

The corresponding return sequence from

UTF-8 to ISO 2022 is ESC % @ ("\\x1b%@").

Other ISO 2022 sequences (such as

for switching the G0 and G1 sets) are not applicable in UTF-8 mode.

.SS Security

The Unicode and UCS standards require that producers of UTF-8

shall use the shortest form possible, for example, producing a two-byte

sequence with first byte 0xc0 is nonconforming.

Unicode 3.1 has added the requirement that conforming programs must not accept

non-shortest forms in their input.

This is for security reasons: if

user input is checked for possible security violations, a program

might check only for the ASCII

version of "/../" or ";" or NUL and overlook that there are many

non-ASCII ways to represent these things in a non-shortest UTF-8

encoding.

.SS Standards

ISO/IEC 10646-1:2000, Unicode 3.1, RFC\ 3629, Plan 9.

.\" .SH AUTHOR

.\" Markus Kuhn <mgk25@cl.cam.ac.uk>

.SH SEE ALSO

.BR locale (1),

.BR nl_langinfo (3),

.BR setlocale (3),

.BR charsets (7),

.BR unicode (7)

.SH COLOPHON

This page is part of release 4.15 of the Linux

.I man-pages

project.

A description of the project,

information about reporting bugs,

and the latest version of this page,

can be found at

\%https://www.kernel.org/doc/man\-pages/.