| 1 |
.\" -*- mode: troff; coding: utf-8 -*- |
| 2 |
.\" Automatically generated by Pod::Man 5.0102 (Pod::Simple 3.45) |
| 3 |
.\" |
| 4 |
.\" Standard preamble: |
| 5 |
.\" ======================================================================== |
| 6 |
.de Sp \" Vertical space (when we can't use .PP) |
| 7 |
.if t .sp .5v |
| 8 |
.if n .sp |
| 9 |
.. |
| 10 |
.de Vb \" Begin verbatim text |
| 11 |
.ft CW |
| 12 |
.nf |
| 13 |
.ne \\$1 |
| 14 |
.. |
| 15 |
.de Ve \" End verbatim text |
| 16 |
.ft R |
| 17 |
.fi |
| 18 |
.. |
| 19 |
.\" \*(C` and \*(C' are quotes in nroff, nothing in troff, for use with C<>. |
| 20 |
.ie n \{\ |
| 21 |
. ds C` "" |
| 22 |
. ds C' "" |
| 23 |
'br\} |
| 24 |
.el\{\ |
| 25 |
. ds C` |
| 26 |
. ds C' |
| 27 |
'br\} |
| 28 |
.\" |
| 29 |
.\" Escape single quotes in literal strings from groff's Unicode transform. |
| 30 |
.ie \n(.g .ds Aq \(aq |
| 31 |
.el .ds Aq ' |
| 32 |
.\" |
| 33 |
.\" If the F register is >0, we'll generate index entries on stderr for |
| 34 |
.\" titles (.TH), headers (.SH), subsections (.SS), items (.Ip), and index |
| 35 |
.\" entries marked with X<> in POD. Of course, you'll have to process the |
| 36 |
.\" output yourself in some meaningful fashion. |
| 37 |
.\" |
| 38 |
.\" Avoid warning from groff about undefined register 'F'. |
| 39 |
.de IX |
| 40 |
.. |
| 41 |
.nr rF 0 |
| 42 |
.if \n(.g .if rF .nr rF 1 |
| 43 |
.if (\n(rF:(\n(.g==0)) \{\ |
| 44 |
. if \nF \{\ |
| 45 |
. de IX |
| 46 |
. tm Index:\\$1\t\\n%\t"\\$2" |
| 47 |
.. |
| 48 |
. if !\nF==2 \{\ |
| 49 |
. nr % 0 |
| 50 |
. nr F 2 |
| 51 |
. \} |
| 52 |
. \} |
| 53 |
.\} |
| 54 |
.rr rF |
| 55 |
.\" ======================================================================== |
| 56 |
.\" |
| 57 |
.IX Title "PERLUNIINTRO 1" |
| 58 |
.TH PERLUNIINTRO 1 2025-11-16 "perl v5.40.1" "Perl Programmers Reference Guide" |
| 59 |
.\" For nroff, turn off justification. Always turn off hyphenation; it makes |
| 60 |
.\" way too many mistakes in technical documents. |
| 61 |
.if n .ad l |
| 62 |
.nh |
| 63 |
.SH NAME |
| 64 |
perluniintro \- Perl Unicode introduction |
| 65 |
.SH DESCRIPTION |
| 66 |
.IX Header "DESCRIPTION" |
| 67 |
This document gives a general idea of Unicode and how to use Unicode |
| 68 |
in Perl. See "Further Resources" for references to more in-depth |
| 69 |
treatments of Unicode. |
| 70 |
.SS Unicode |
| 71 |
.IX Subsection "Unicode" |
| 72 |
Unicode is a character set standard which plans to codify all of the |
| 73 |
writing systems of the world, plus many other symbols. |
| 74 |
.PP |
| 75 |
Unicode and ISO/IEC 10646 are coordinated standards that unify |
| 76 |
almost all other modern character set standards, |
| 77 |
covering more than 80 writing systems and hundreds of languages, |
| 78 |
including all commercially-important modern languages. All characters |
| 79 |
in the largest Chinese, Japanese, and Korean dictionaries are also |
| 80 |
encoded. The standards will eventually cover almost all characters in |
| 81 |
more than 250 writing systems and thousands of languages. |
| 82 |
Unicode 1.0 was released in October 1991, and 6.0 in October 2010. |
| 83 |
.PP |
| 84 |
A Unicode \fIcharacter\fR is an abstract entity. It is not bound to any |
| 85 |
particular integer width, especially not to the C language \f(CW\*(C`char\*(C'\fR. |
| 86 |
Unicode is language-neutral and display-neutral: it does not encode the |
| 87 |
language of the text, and it does not generally define fonts or other graphical |
| 88 |
layout details. Unicode operates on characters and on text built from |
| 89 |
those characters. |
| 90 |
.PP |
| 91 |
Unicode defines characters like \f(CW\*(C`LATIN CAPITAL LETTER A\*(C'\fR or \f(CW\*(C`GREEK |
| 92 |
SMALL LETTER ALPHA\*(C'\fR and unique numbers for the characters, in this |
| 93 |
case 0x0041 and 0x03B1, respectively. These unique numbers are called |
| 94 |
\&\fIcode points\fR. A code point is essentially the position of the |
| 95 |
character within the set of all possible Unicode characters, and thus in |
| 96 |
Perl, the term \fIordinal\fR is often used interchangeably with it. |
| 97 |
.PP |
| 98 |
The Unicode standard prefers using hexadecimal notation for the code |
| 99 |
points. If numbers like \f(CW0x0041\fR are unfamiliar to you, take a peek |
| 100 |
at a later section, "Hexadecimal Notation". The Unicode standard |
| 101 |
uses the notation \f(CW\*(C`U+0041 LATIN CAPITAL LETTER A\*(C'\fR, to give the |
| 102 |
hexadecimal code point and the normative name of the character. |
| 103 |
.PP |
| 104 |
Unicode also defines various \fIproperties\fR for the characters, like |
| 105 |
"uppercase" or "lowercase", "decimal digit", or "punctuation"; |
| 106 |
these properties are independent of the names of the characters. |
| 107 |
Furthermore, various operations on the characters like uppercasing, |
| 108 |
lowercasing, and collating (sorting) are defined. |
| 109 |
.PP |
| 110 |
A Unicode \fIlogical\fR "character" can actually consist of more than one internal |
| 111 |
\&\fIactual\fR "character" or code point. For Western languages, this is adequately |
| 112 |
modelled by a \fIbase character\fR (like \f(CW\*(C`LATIN CAPITAL LETTER A\*(C'\fR) followed |
| 113 |
by one or more \fImodifiers\fR (like \f(CW\*(C`COMBINING ACUTE ACCENT\*(C'\fR). This sequence of |
| 114 |
base character and modifiers is called a \fIcombining character |
| 115 |
sequence\fR. Some non-western languages require more complicated |
| 116 |
models, so Unicode created the \fIgrapheme cluster\fR concept, which was |
| 117 |
later further refined into the \fIextended grapheme cluster\fR. For |
| 118 |
example, a Korean Hangul syllable is considered a single logical |
| 119 |
character, but most often consists of three actual |
| 120 |
Unicode characters: a leading consonant followed by an interior vowel followed |
| 121 |
by a trailing consonant. |
| 122 |
.PP |
| 123 |
Whether to call these extended grapheme clusters "characters" depends on your |
| 124 |
point of view. If you are a programmer, you probably would tend towards seeing |
| 125 |
each element in the sequences as one unit, or "character". However from |
| 126 |
the user's point of view, the whole sequence could be seen as one |
| 127 |
"character" since that's probably what it looks like in the context of the |
| 128 |
user's language. In this document, we take the programmer's point of |
| 129 |
view: one "character" is one Unicode code point. |
| 130 |
.PP |
| 131 |
For some combinations of base character and modifiers, there are |
| 132 |
\&\fIprecomposed\fR characters. There is a single character equivalent, for |
| 133 |
example, for the sequence \f(CW\*(C`LATIN CAPITAL LETTER A\*(C'\fR followed by |
| 134 |
\&\f(CW\*(C`COMBINING ACUTE ACCENT\*(C'\fR. It is called \f(CW\*(C`LATIN CAPITAL LETTER A WITH |
| 135 |
ACUTE\*(C'\fR. These precomposed characters are, however, only available for |
| 136 |
some combinations, and are mainly meant to support round-trip |
| 137 |
conversions between Unicode and legacy standards (like ISO 8859). Using |
| 138 |
sequences, as Unicode does, allows for needing fewer basic building blocks |
| 139 |
(code points) to express many more potential grapheme clusters. To |
| 140 |
support conversion between equivalent forms, various \fInormalization |
| 141 |
forms\fR are also defined. Thus, \f(CW\*(C`LATIN CAPITAL LETTER A WITH ACUTE\*(C'\fR is |
| 142 |
in \fINormalization Form Composed\fR, (abbreviated NFC), and the sequence |
| 143 |
\&\f(CW\*(C`LATIN CAPITAL LETTER A\*(C'\fR followed by \f(CW\*(C`COMBINING ACUTE ACCENT\*(C'\fR |
| 144 |
represents the same character in \fINormalization Form Decomposed\fR (NFD). |
| 145 |
.PP |
| 146 |
Because of backward compatibility with legacy encodings, the "a unique |
| 147 |
number for every character" idea breaks down a bit: instead, there is |
| 148 |
"at least one number for every character". The same character could |
| 149 |
be represented differently in several legacy encodings. The |
| 150 |
converse is not true: some code points do not have an assigned |
| 151 |
character. Firstly, there are unallocated code points within |
| 152 |
otherwise used blocks. Secondly, there are special Unicode control |
| 153 |
characters that do not represent true characters. |
| 154 |
.PP |
| 155 |
When Unicode was first conceived, it was thought that all the world's |
| 156 |
characters could be represented using a 16\-bit word; that is a maximum of |
| 157 |
\&\f(CW0x10000\fR (or 65,536) characters would be needed, from \f(CW0x0000\fR to |
| 158 |
\&\f(CW0xFFFF\fR. This soon proved to be wrong, and since Unicode 2.0 (July |
| 159 |
1996), Unicode has been defined all the way up to 21 bits (\f(CW0x10FFFF\fR), |
| 160 |
and Unicode 3.1 (March 2001) defined the first characters above \f(CW0xFFFF\fR. |
| 161 |
The first \f(CW0x10000\fR characters are called the \fIPlane 0\fR, or the |
| 162 |
\&\fIBasic Multilingual Plane\fR (BMP). With Unicode 3.1, 17 (yes, |
| 163 |
seventeen) planes in all were defined\-\-but they are nowhere near full of |
| 164 |
defined characters, yet. |
| 165 |
.PP |
| 166 |
When a new language is being encoded, Unicode generally will choose a |
| 167 |
\&\f(CW\*(C`block\*(C'\fR of consecutive unallocated code points for its characters. So |
| 168 |
far, the number of code points in these blocks has always been evenly |
| 169 |
divisible by 16. Extras in a block, not currently needed, are left |
| 170 |
unallocated, for future growth. But there have been occasions when |
| 171 |
a later release needed more code points than the available extras, and a |
| 172 |
new block had to allocated somewhere else, not contiguous to the initial |
| 173 |
one, to handle the overflow. Thus, it became apparent early on that |
| 174 |
"block" wasn't an adequate organizing principle, and so the \f(CW\*(C`Script\*(C'\fR |
| 175 |
property was created. (Later an improved script property was added as |
| 176 |
well, the \f(CW\*(C`Script_Extensions\*(C'\fR property.) Those code points that are in |
| 177 |
overflow blocks can still |
| 178 |
have the same script as the original ones. The script concept fits more |
| 179 |
closely with natural language: there is \f(CW\*(C`Latin\*(C'\fR script, \f(CW\*(C`Greek\*(C'\fR |
| 180 |
script, and so on; and there are several artificial scripts, like |
| 181 |
\&\f(CW\*(C`Common\*(C'\fR for characters that are used in multiple scripts, such as |
| 182 |
mathematical symbols. Scripts usually span varied parts of several |
| 183 |
blocks. For more information about scripts, see "Scripts" in perlunicode. |
| 184 |
The division into blocks exists, but it is almost completely |
| 185 |
accidental\-\-an artifact of how the characters have been and still are |
| 186 |
allocated. (Note that this paragraph has oversimplified things for the |
| 187 |
sake of this being an introduction. Unicode doesn't really encode |
| 188 |
languages, but the writing systems for them\-\-their scripts; and one |
| 189 |
script can be used by many languages. Unicode also encodes things that |
| 190 |
aren't really about languages, such as symbols like \f(CW\*(C`BAGGAGE CLAIM\*(C'\fR.) |
| 191 |
.PP |
| 192 |
The Unicode code points are just abstract numbers. To input and |
| 193 |
output these abstract numbers, the numbers must be \fIencoded\fR or |
| 194 |
\&\fIserialised\fR somehow. Unicode defines several \fIcharacter encoding |
| 195 |
forms\fR, of which \fIUTF\-8\fR is the most popular. UTF\-8 is a |
| 196 |
variable length encoding that encodes Unicode characters as 1 to 4 |
| 197 |
bytes. Other encodings |
| 198 |
include UTF\-16 and UTF\-32 and their big\- and little-endian variants |
| 199 |
(UTF\-8 is byte-order independent). The ISO/IEC 10646 defines the UCS\-2 |
| 200 |
and UCS\-4 encoding forms. |
| 201 |
.PP |
| 202 |
For more information about encodings\-\-for instance, to learn what |
| 203 |
\&\fIsurrogates\fR and \fIbyte order marks\fR (BOMs) are\-\-see perlunicode. |
| 204 |
.SS "Perl's Unicode Support" |
| 205 |
.IX Subsection "Perl's Unicode Support" |
| 206 |
Starting from Perl v5.6.0, Perl has had the capacity to handle Unicode |
| 207 |
natively. Perl v5.8.0, however, is the first recommended release for |
| 208 |
serious Unicode work. The maintenance release 5.6.1 fixed many of the |
| 209 |
problems of the initial Unicode implementation, but for example |
| 210 |
regular expressions still do not work with Unicode in 5.6.1. |
| 211 |
Perl v5.14.0 is the first release where Unicode support is |
| 212 |
(almost) seamlessly integratable without some gotchas. (There are a few |
| 213 |
exceptions. Firstly, some differences in quotemeta |
| 214 |
were fixed starting in Perl 5.16.0. Secondly, some differences in |
| 215 |
the range operator were fixed starting in |
| 216 |
Perl 5.26.0. Thirdly, some differences in split were fixed |
| 217 |
started in Perl 5.28.0.) |
| 218 |
.PP |
| 219 |
To enable this |
| 220 |
seamless support, you should \f(CW\*(C`use feature \*(Aqunicode_strings\*(Aq\*(C'\fR (which is |
| 221 |
automatically selected if you \f(CW\*(C`use v5.12\*(C'\fR or higher). See feature. |
| 222 |
(5.14 also fixes a number of bugs and departures from the Unicode |
| 223 |
standard.) |
| 224 |
.PP |
| 225 |
Before Perl v5.8.0, the use of \f(CW\*(C`use utf8\*(C'\fR was used to declare |
| 226 |
that operations in the current block or file would be Unicode-aware. |
| 227 |
This model was found to be wrong, or at least clumsy: the "Unicodeness" |
| 228 |
is now carried with the data, instead of being attached to the |
| 229 |
operations. |
| 230 |
Starting with Perl v5.8.0, only one case remains where an explicit \f(CW\*(C`use |
| 231 |
utf8\*(C'\fR is needed: if your Perl script itself is encoded in UTF\-8, you can |
| 232 |
use UTF\-8 in your identifier names, and in string and regular expression |
| 233 |
literals, by saying \f(CW\*(C`use utf8\*(C'\fR. This is not the default because |
| 234 |
scripts with legacy 8\-bit data in them would break. See utf8. |
| 235 |
.SS "Perl's Unicode Model" |
| 236 |
.IX Subsection "Perl's Unicode Model" |
| 237 |
Perl supports both pre\-5.6 strings of eight-bit native bytes, and |
| 238 |
strings of Unicode characters. The general principle is that Perl tries |
| 239 |
to keep its data as eight-bit bytes for as long as possible, but as soon |
| 240 |
as Unicodeness cannot be avoided, the data is transparently upgraded |
| 241 |
to Unicode. Prior to Perl v5.14.0, the upgrade was not completely |
| 242 |
transparent (see "The "Unicode Bug"" in perlunicode), and for backwards |
| 243 |
compatibility, full transparency is not gained unless \f(CWuse feature |
| 244 |
\&\*(Aqunicode_strings\*(Aq\fR (see feature) or \f(CW\*(C`use v5.12\*(C'\fR (or higher) is |
| 245 |
selected. |
| 246 |
.PP |
| 247 |
Internally, Perl currently uses either whatever the native eight-bit |
| 248 |
character set of the platform (for example Latin\-1) is, defaulting to |
| 249 |
UTF\-8, to encode Unicode strings. Specifically, if all code points in |
| 250 |
the string are \f(CW0xFF\fR or less, Perl uses the native eight-bit |
| 251 |
character set. Otherwise, it uses UTF\-8. |
| 252 |
.PP |
| 253 |
A user of Perl does not normally need to know nor care how Perl |
| 254 |
happens to encode its internal strings, but it becomes relevant when |
| 255 |
outputting Unicode strings to a stream without a PerlIO layer (one with |
| 256 |
the "default" encoding). In such a case, the raw bytes used internally |
| 257 |
(the native character set or UTF\-8, as appropriate for each string) |
| 258 |
will be used, and a "Wide character" warning will be issued if those |
| 259 |
strings contain a character beyond 0x00FF. |
| 260 |
.PP |
| 261 |
For example, |
| 262 |
.PP |
| 263 |
.Vb 1 |
| 264 |
\& perl \-e \*(Aqprint "\ex{DF}\en", "\ex{0100}\ex{DF}\en"\*(Aq |
| 265 |
.Ve |
| 266 |
.PP |
| 267 |
produces a fairly useless mixture of native bytes and UTF\-8, as well |
| 268 |
as a warning: |
| 269 |
.PP |
| 270 |
.Vb 1 |
| 271 |
\& Wide character in print at ... |
| 272 |
.Ve |
| 273 |
.PP |
| 274 |
To output UTF\-8, use the \f(CW\*(C`:encoding\*(C'\fR or \f(CW\*(C`:utf8\*(C'\fR output layer. Prepending |
| 275 |
.PP |
| 276 |
.Vb 1 |
| 277 |
\& binmode(STDOUT, ":utf8"); |
| 278 |
.Ve |
| 279 |
.PP |
| 280 |
to this sample program ensures that the output is completely UTF\-8, |
| 281 |
and removes the program's warning. |
| 282 |
.PP |
| 283 |
You can enable automatic UTF\-8\-ification of your standard file |
| 284 |
handles, default \f(CWopen()\fR layer, and \f(CW@ARGV\fR by using either |
| 285 |
the \f(CW\*(C`\-C\*(C'\fR command line switch or the \f(CW\*(C`PERL_UNICODE\*(C'\fR environment |
| 286 |
variable, see perlrun for the |
| 287 |
documentation of the \f(CW\*(C`\-C\*(C'\fR switch. |
| 288 |
.PP |
| 289 |
Note that this means that Perl expects other software to work the same |
| 290 |
way: |
| 291 |
if Perl has been led to believe that STDIN should be UTF\-8, but then |
| 292 |
STDIN coming in from another command is not UTF\-8, Perl will likely |
| 293 |
complain about the malformed UTF\-8. |
| 294 |
.PP |
| 295 |
All features that combine Unicode and I/O also require using the new |
| 296 |
PerlIO feature. Almost all Perl 5.8 platforms do use PerlIO, though: |
| 297 |
you can see whether yours is by running "perl \-V" and looking for |
| 298 |
\&\f(CW\*(C`useperlio=define\*(C'\fR. |
| 299 |
.SS "Unicode and EBCDIC" |
| 300 |
.IX Subsection "Unicode and EBCDIC" |
| 301 |
Perl 5.8.0 added support for Unicode on EBCDIC platforms. This support |
| 302 |
was allowed to lapse in later releases, but was revived in 5.22. |
| 303 |
Unicode support is somewhat more complex to implement since additional |
| 304 |
conversions are needed. See perlebcdic for more information. |
| 305 |
.PP |
| 306 |
On EBCDIC platforms, the internal Unicode encoding form is UTF-EBCDIC |
| 307 |
instead of UTF\-8. The difference is that as UTF\-8 is "ASCII-safe" in |
| 308 |
that ASCII characters encode to UTF\-8 as-is, while UTF-EBCDIC is |
| 309 |
"EBCDIC-safe", in that all the basic characters (which includes all |
| 310 |
those that have ASCII equivalents (like \f(CW"A"\fR, \f(CW"0"\fR, \f(CW"%"\fR, \fIetc.\fR) |
| 311 |
are the same in both EBCDIC and UTF-EBCDIC. Often, documentation |
| 312 |
will use the term "UTF\-8" to mean UTF-EBCDIC as well. This is the case |
| 313 |
in this document. |
| 314 |
.SS "Creating Unicode" |
| 315 |
.IX Subsection "Creating Unicode" |
| 316 |
This section applies fully to Perls starting with v5.22. Various |
| 317 |
caveats for earlier releases are in the "Earlier releases caveats" |
| 318 |
subsection below. |
| 319 |
.PP |
| 320 |
To create Unicode characters in literals, |
| 321 |
use the \f(CW\*(C`\eN{...}\*(C'\fR notation in double-quoted strings: |
| 322 |
.PP |
| 323 |
.Vb 2 |
| 324 |
\& my $smiley_from_name = "\eN{WHITE SMILING FACE}"; |
| 325 |
\& my $smiley_from_code_point = "\eN{U+263a}"; |
| 326 |
.Ve |
| 327 |
.PP |
| 328 |
Similarly, they can be used in regular expression literals |
| 329 |
.PP |
| 330 |
.Vb 2 |
| 331 |
\& $smiley =~ /\eN{WHITE SMILING FACE}/; |
| 332 |
\& $smiley =~ /\eN{U+263a}/; |
| 333 |
.Ve |
| 334 |
.PP |
| 335 |
or, starting in v5.32: |
| 336 |
.PP |
| 337 |
.Vb 2 |
| 338 |
\& $smiley =~ /\ep{Name=WHITE SMILING FACE}/; |
| 339 |
\& $smiley =~ /\ep{Name=whitesmilingface}/; |
| 340 |
.Ve |
| 341 |
.PP |
| 342 |
At run-time you can use: |
| 343 |
.PP |
| 344 |
.Vb 4 |
| 345 |
\& use charnames (); |
| 346 |
\& my $hebrew_alef_from_name |
| 347 |
\& = charnames::string_vianame("HEBREW LETTER ALEF"); |
| 348 |
\& my $hebrew_alef_from_code_point = charnames::string_vianame("U+05D0"); |
| 349 |
.Ve |
| 350 |
.PP |
| 351 |
Naturally, \f(CWord()\fR will do the reverse: it turns a character into |
| 352 |
a code point. |
| 353 |
.PP |
| 354 |
There are other runtime options as well. You can use \f(CWpack()\fR: |
| 355 |
.PP |
| 356 |
.Vb 1 |
| 357 |
\& my $hebrew_alef_from_code_point = pack("U", 0x05d0); |
| 358 |
.Ve |
| 359 |
.PP |
| 360 |
Or you can use \f(CWchr()\fR, though it is less convenient in the general |
| 361 |
case: |
| 362 |
.PP |
| 363 |
.Vb 2 |
| 364 |
\& $hebrew_alef_from_code_point = chr(utf8::unicode_to_native(0x05d0)); |
| 365 |
\& utf8::upgrade($hebrew_alef_from_code_point); |
| 366 |
.Ve |
| 367 |
.PP |
| 368 |
The \f(CWutf8::unicode_to_native()\fR and \f(CWutf8::upgrade()\fR aren't needed if |
| 369 |
the argument is above 0xFF, so the above could have been written as |
| 370 |
.PP |
| 371 |
.Vb 1 |
| 372 |
\& $hebrew_alef_from_code_point = chr(0x05d0); |
| 373 |
.Ve |
| 374 |
.PP |
| 375 |
since 0x5d0 is above 255. |
| 376 |
.PP |
| 377 |
\&\f(CW\*(C`\ex{}\*(C'\fR and \f(CW\*(C`\eo{}\*(C'\fR can also be used to specify code points at compile |
| 378 |
time in double-quotish strings, but, for backward compatibility with |
| 379 |
older Perls, the same rules apply as with \f(CWchr()\fR for code points less |
| 380 |
than 256. |
| 381 |
.PP |
| 382 |
\&\f(CWutf8::unicode_to_native()\fR is used so that the Perl code is portable |
| 383 |
to EBCDIC platforms. You can omit it if you're \fIreally\fR sure no one |
| 384 |
will ever want to use your code on a non-ASCII platform. Starting in |
| 385 |
Perl v5.22, calls to it on ASCII platforms are optimized out, so there's |
| 386 |
no performance penalty at all in adding it. Or you can simply use the |
| 387 |
other constructs that don't require it. |
| 388 |
.PP |
| 389 |
See "Further Resources" for how to find all these names and numeric |
| 390 |
codes. |
| 391 |
.PP |
| 392 |
\fIEarlier releases caveats\fR |
| 393 |
.IX Subsection "Earlier releases caveats" |
| 394 |
.PP |
| 395 |
On EBCDIC platforms, prior to v5.22, using \f(CW\*(C`\eN{U+...}\*(C'\fR doesn't work |
| 396 |
properly. |
| 397 |
.PP |
| 398 |
Prior to v5.16, using \f(CW\*(C`\eN{...}\*(C'\fR with a character name (as opposed to a |
| 399 |
\&\f(CW\*(C`U+...\*(C'\fR code point) required a \f(CW\*(C`use\ charnames\ :full\*(C'\fR. |
| 400 |
.PP |
| 401 |
Prior to v5.14, there were some bugs in \f(CW\*(C`\eN{...}\*(C'\fR with a character name |
| 402 |
(as opposed to a \f(CW\*(C`U+...\*(C'\fR code point). |
| 403 |
.PP |
| 404 |
\&\f(CWcharnames::string_vianame()\fR was introduced in v5.14. Prior to that, |
| 405 |
\&\f(CWcharnames::vianame()\fR should work, but only if the argument is of the |
| 406 |
form \f(CW"U+..."\fR. Your best bet there for runtime Unicode by character |
| 407 |
name is probably: |
| 408 |
.PP |
| 409 |
.Vb 3 |
| 410 |
\& use charnames (); |
| 411 |
\& my $hebrew_alef_from_name |
| 412 |
\& = pack("U", charnames::vianame("HEBREW LETTER ALEF")); |
| 413 |
.Ve |
| 414 |
.SS "Handling Unicode" |
| 415 |
.IX Subsection "Handling Unicode" |
| 416 |
Handling Unicode is for the most part transparent: just use the |
| 417 |
strings as usual. Functions like \f(CWindex()\fR, \f(CWlength()\fR, and |
| 418 |
\&\f(CWsubstr()\fR will work on the Unicode characters; regular expressions |
| 419 |
will work on the Unicode characters (see perlunicode and perlretut). |
| 420 |
.PP |
| 421 |
Note that Perl considers grapheme clusters to be separate characters, so for |
| 422 |
example |
| 423 |
.PP |
| 424 |
.Vb 2 |
| 425 |
\& print length("\eN{LATIN CAPITAL LETTER A}\eN{COMBINING ACUTE ACCENT}"), |
| 426 |
\& "\en"; |
| 427 |
.Ve |
| 428 |
.PP |
| 429 |
will print 2, not 1. The only exception is that regular expressions |
| 430 |
have \f(CW\*(C`\eX\*(C'\fR for matching an extended grapheme cluster. (Thus \f(CW\*(C`\eX\*(C'\fR in a |
| 431 |
regular expression would match the entire sequence of both the example |
| 432 |
characters.) |
| 433 |
.PP |
| 434 |
Life is not quite so transparent, however, when working with legacy |
| 435 |
encodings, I/O, and certain special cases: |
| 436 |
.SS "Legacy Encodings" |
| 437 |
.IX Subsection "Legacy Encodings" |
| 438 |
When you combine legacy data and Unicode, the legacy data needs |
| 439 |
to be upgraded to Unicode. Normally the legacy data is assumed to be |
| 440 |
ISO 8859\-1 (or EBCDIC, if applicable). |
| 441 |
.PP |
| 442 |
The \f(CW\*(C`Encode\*(C'\fR module knows about many encodings and has interfaces |
| 443 |
for doing conversions between those encodings: |
| 444 |
.PP |
| 445 |
.Vb 2 |
| 446 |
\& use Encode \*(Aqdecode\*(Aq; |
| 447 |
\& $data = decode("iso\-8859\-3", $data); # convert from legacy |
| 448 |
.Ve |
| 449 |
.SS "Unicode I/O" |
| 450 |
.IX Subsection "Unicode I/O" |
| 451 |
Normally, writing out Unicode data |
| 452 |
.PP |
| 453 |
.Vb 1 |
| 454 |
\& print FH $some_string_with_unicode, "\en"; |
| 455 |
.Ve |
| 456 |
.PP |
| 457 |
produces raw bytes that Perl happens to use to internally encode the |
| 458 |
Unicode string. Perl's internal encoding depends on the system as |
| 459 |
well as what characters happen to be in the string at the time. If |
| 460 |
any of the characters are at code points \f(CW0x100\fR or above, you will get |
| 461 |
a warning. To ensure that the output is explicitly rendered in the |
| 462 |
encoding you desire\-\-and to avoid the warning\-\-open the stream with |
| 463 |
the desired encoding. Some examples: |
| 464 |
.PP |
| 465 |
.Vb 1 |
| 466 |
\& open FH, ">:utf8", "file"; |
| 467 |
\& |
| 468 |
\& open FH, ">:encoding(ucs2)", "file"; |
| 469 |
\& open FH, ">:encoding(UTF\-8)", "file"; |
| 470 |
\& open FH, ">:encoding(shift_jis)", "file"; |
| 471 |
.Ve |
| 472 |
.PP |
| 473 |
and on already open streams, use \f(CWbinmode()\fR: |
| 474 |
.PP |
| 475 |
.Vb 1 |
| 476 |
\& binmode(STDOUT, ":utf8"); |
| 477 |
\& |
| 478 |
\& binmode(STDOUT, ":encoding(ucs2)"); |
| 479 |
\& binmode(STDOUT, ":encoding(UTF\-8)"); |
| 480 |
\& binmode(STDOUT, ":encoding(shift_jis)"); |
| 481 |
.Ve |
| 482 |
.PP |
| 483 |
The matching of encoding names is loose: case does not matter, and |
| 484 |
many encodings have several aliases. Note that the \f(CW\*(C`:utf8\*(C'\fR layer |
| 485 |
must always be specified exactly like that; it is \fInot\fR subject to |
| 486 |
the loose matching of encoding names. Also note that currently \f(CW\*(C`:utf8\*(C'\fR is unsafe for |
| 487 |
input, because it accepts the data without validating that it is indeed valid |
| 488 |
UTF\-8; you should instead use \f(CW:encoding(UTF\-8)\fR (with or without a |
| 489 |
hyphen). |
| 490 |
.PP |
| 491 |
See PerlIO for the \f(CW\*(C`:utf8\*(C'\fR layer, PerlIO::encoding and |
| 492 |
Encode::PerlIO for the \f(CW:encoding()\fR layer, and |
| 493 |
Encode::Supported for many encodings supported by the \f(CW\*(C`Encode\*(C'\fR |
| 494 |
module. |
| 495 |
.PP |
| 496 |
Reading in a file that you know happens to be encoded in one of the |
| 497 |
Unicode or legacy encodings does not magically turn the data into |
| 498 |
Unicode in Perl's eyes. To do that, specify the appropriate |
| 499 |
layer when opening files |
| 500 |
.PP |
| 501 |
.Vb 2 |
| 502 |
\& open(my $fh,\*(Aq<:encoding(UTF\-8)\*(Aq, \*(Aqanything\*(Aq); |
| 503 |
\& my $line_of_unicode = <$fh>; |
| 504 |
\& |
| 505 |
\& open(my $fh,\*(Aq<:encoding(Big5)\*(Aq, \*(Aqanything\*(Aq); |
| 506 |
\& my $line_of_unicode = <$fh>; |
| 507 |
.Ve |
| 508 |
.PP |
| 509 |
The I/O layers can also be specified more flexibly with |
| 510 |
the \f(CW\*(C`open\*(C'\fR pragma. See open, or look at the following example. |
| 511 |
.PP |
| 512 |
.Vb 8 |
| 513 |
\& use open \*(Aq:encoding(UTF\-8)\*(Aq; # input/output default encoding will be |
| 514 |
\& # UTF\-8 |
| 515 |
\& open X, ">file"; |
| 516 |
\& print X chr(0x100), "\en"; |
| 517 |
\& close X; |
| 518 |
\& open Y, "<file"; |
| 519 |
\& printf "%#x\en", ord(<Y>); # this should print 0x100 |
| 520 |
\& close Y; |
| 521 |
.Ve |
| 522 |
.PP |
| 523 |
With the \f(CW\*(C`open\*(C'\fR pragma you can use the \f(CW\*(C`:locale\*(C'\fR layer |
| 524 |
.PP |
| 525 |
.Vb 10 |
| 526 |
\& BEGIN { $ENV{LC_ALL} = $ENV{LANG} = \*(Aqru_RU.KOI8\-R\*(Aq } |
| 527 |
\& # the :locale will probe the locale environment variables like |
| 528 |
\& # LC_ALL |
| 529 |
\& use open OUT => \*(Aq:locale\*(Aq; # russki parusski |
| 530 |
\& open(O, ">koi8"); |
| 531 |
\& print O chr(0x430); # Unicode CYRILLIC SMALL LETTER A = KOI8\-R 0xc1 |
| 532 |
\& close O; |
| 533 |
\& open(I, "<koi8"); |
| 534 |
\& printf "%#x\en", ord(<I>), "\en"; # this should print 0xc1 |
| 535 |
\& close I; |
| 536 |
.Ve |
| 537 |
.PP |
| 538 |
These methods install a transparent filter on the I/O stream that |
| 539 |
converts data from the specified encoding when it is read in from the |
| 540 |
stream. The result is always Unicode. |
| 541 |
.PP |
| 542 |
The open pragma affects all the \f(CWopen()\fR calls after the pragma by |
| 543 |
setting default layers. If you want to affect only certain |
| 544 |
streams, use explicit layers directly in the \f(CWopen()\fR call. |
| 545 |
.PP |
| 546 |
You can switch encodings on an already opened stream by using |
| 547 |
\&\f(CWbinmode()\fR; see "binmode" in perlfunc. |
| 548 |
.PP |
| 549 |
The \f(CW\*(C`:locale\*(C'\fR does not currently work with |
| 550 |
\&\f(CWopen()\fR and \f(CWbinmode()\fR, only with the \f(CW\*(C`open\*(C'\fR pragma. The |
| 551 |
\&\f(CW\*(C`:utf8\*(C'\fR and \f(CW:encoding(...)\fR methods do work with all of \f(CWopen()\fR, |
| 552 |
\&\f(CWbinmode()\fR, and the \f(CW\*(C`open\*(C'\fR pragma. |
| 553 |
.PP |
| 554 |
Similarly, you may use these I/O layers on output streams to |
| 555 |
automatically convert Unicode to the specified encoding when it is |
| 556 |
written to the stream. For example, the following snippet copies the |
| 557 |
contents of the file "text.jis" (encoded as ISO\-2022\-JP, aka JIS) to |
| 558 |
the file "text.utf8", encoded as UTF\-8: |
| 559 |
.PP |
| 560 |
.Vb 3 |
| 561 |
\& open(my $nihongo, \*(Aq<:encoding(iso\-2022\-jp)\*(Aq, \*(Aqtext.jis\*(Aq); |
| 562 |
\& open(my $unicode, \*(Aq>:utf8\*(Aq, \*(Aqtext.utf8\*(Aq); |
| 563 |
\& while (<$nihongo>) { print $unicode $_ } |
| 564 |
.Ve |
| 565 |
.PP |
| 566 |
The naming of encodings, both by the \f(CWopen()\fR and by the \f(CW\*(C`open\*(C'\fR |
| 567 |
pragma allows for flexible names: \f(CW\*(C`koi8\-r\*(C'\fR and \f(CW\*(C`KOI8R\*(C'\fR will both be |
| 568 |
understood. |
| 569 |
.PP |
| 570 |
Common encodings recognized by ISO, MIME, IANA, and various other |
| 571 |
standardisation organisations are recognised; for a more detailed |
| 572 |
list see Encode::Supported. |
| 573 |
.PP |
| 574 |
\&\f(CWread()\fR reads characters and returns the number of characters. |
| 575 |
\&\f(CWseek()\fR and \f(CWtell()\fR operate on byte counts, as does \f(CWsysseek()\fR. |
| 576 |
.PP |
| 577 |
\&\f(CWsysread()\fR and \f(CWsyswrite()\fR should not be used on file handles with |
| 578 |
character encoding layers, they behave badly, and that behaviour has |
| 579 |
been deprecated since perl 5.24. |
| 580 |
.PP |
| 581 |
Notice that because of the default behaviour of not doing any |
| 582 |
conversion upon input if there is no default layer, |
| 583 |
it is easy to mistakenly write code that keeps on expanding a file |
| 584 |
by repeatedly encoding the data: |
| 585 |
.PP |
| 586 |
.Vb 8 |
| 587 |
\& # BAD CODE WARNING |
| 588 |
\& open F, "file"; |
| 589 |
\& local $/; ## read in the whole file of 8\-bit characters |
| 590 |
\& $t = <F>; |
| 591 |
\& close F; |
| 592 |
\& open F, ">:encoding(UTF\-8)", "file"; |
| 593 |
\& print F $t; ## convert to UTF\-8 on output |
| 594 |
\& close F; |
| 595 |
.Ve |
| 596 |
.PP |
| 597 |
If you run this code twice, the contents of the \fIfile\fR will be twice |
| 598 |
UTF\-8 encoded. A \f(CW\*(C`use open \*(Aq:encoding(UTF\-8)\*(Aq\*(C'\fR would have avoided the |
| 599 |
bug, or explicitly opening also the \fIfile\fR for input as UTF\-8. |
| 600 |
.PP |
| 601 |
\&\fBNOTE\fR: the \f(CW\*(C`:utf8\*(C'\fR and \f(CW\*(C`:encoding\*(C'\fR features work only if your |
| 602 |
Perl has been built with PerlIO, which is the default |
| 603 |
on most systems. |
| 604 |
.SS "Displaying Unicode As Text" |
| 605 |
.IX Subsection "Displaying Unicode As Text" |
| 606 |
Sometimes you might want to display Perl scalars containing Unicode as |
| 607 |
simple ASCII (or EBCDIC) text. The following subroutine converts |
| 608 |
its argument so that Unicode characters with code points greater than |
| 609 |
255 are displayed as \f(CW\*(C`\ex{...}\*(C'\fR, control characters (like \f(CW\*(C`\en\*(C'\fR) are |
| 610 |
displayed as \f(CW\*(C`\ex..\*(C'\fR, and the rest of the characters as themselves: |
| 611 |
.PP |
| 612 |
.Vb 9 |
| 613 |
\& sub nice_string { |
| 614 |
\& join("", |
| 615 |
\& map { $_ > 255 # if wide character... |
| 616 |
\& ? sprintf("\e\ex{%04X}", $_) # \ex{...} |
| 617 |
\& : chr($_) =~ /[[:cntrl:]]/ # else if control character... |
| 618 |
\& ? sprintf("\e\ex%02X", $_) # \ex.. |
| 619 |
\& : quotemeta(chr($_)) # else quoted or as themselves |
| 620 |
\& } unpack("W*", $_[0])); # unpack Unicode characters |
| 621 |
\& } |
| 622 |
.Ve |
| 623 |
.PP |
| 624 |
For example, |
| 625 |
.PP |
| 626 |
.Vb 1 |
| 627 |
\& nice_string("foo\ex{100}bar\en") |
| 628 |
.Ve |
| 629 |
.PP |
| 630 |
returns the string |
| 631 |
.PP |
| 632 |
.Vb 1 |
| 633 |
\& \*(Aqfoo\ex{0100}bar\ex0A\*(Aq |
| 634 |
.Ve |
| 635 |
.PP |
| 636 |
which is ready to be printed. |
| 637 |
.PP |
| 638 |
(\f(CW\*(C`\e\ex{}\*(C'\fR is used here instead of \f(CW\*(C`\e\eN{}\*(C'\fR, since it's most likely that |
| 639 |
you want to see what the native values are.) |
| 640 |
.SS "Special Cases" |
| 641 |
.IX Subsection "Special Cases" |
| 642 |
.IP \(bu 4 |
| 643 |
Starting in Perl 5.28, it is illegal for bit operators, like \f(CW\*(C`~\*(C'\fR, to |
| 644 |
operate on strings containing code points above 255. |
| 645 |
.IP \(bu 4 |
| 646 |
The \fBvec()\fR function may produce surprising results if |
| 647 |
used on strings containing characters with ordinal values above |
| 648 |
255. In such a case, the results are consistent with the internal |
| 649 |
encoding of the characters, but not with much else. So don't do |
| 650 |
that, and starting in Perl 5.28, a deprecation message is issued if you |
| 651 |
do so, becoming illegal in Perl 5.32. |
| 652 |
.IP \(bu 4 |
| 653 |
Peeking At Perl's Internal Encoding |
| 654 |
.Sp |
| 655 |
Normal users of Perl should never care how Perl encodes any particular |
| 656 |
Unicode string (because the normal ways to get at the contents of a |
| 657 |
string with Unicode\-\-via input and output\-\-should always be via |
| 658 |
explicitly-defined I/O layers). But if you must, there are two |
| 659 |
ways of looking behind the scenes. |
| 660 |
.Sp |
| 661 |
One way of peeking inside the internal encoding of Unicode characters |
| 662 |
is to use \f(CW\*(C`unpack("C*", ...\*(C'\fR to get the bytes of whatever the string |
| 663 |
encoding happens to be, or \f(CW\*(C`unpack("U0..", ...)\*(C'\fR to get the bytes of the |
| 664 |
UTF\-8 encoding: |
| 665 |
.Sp |
| 666 |
.Vb 2 |
| 667 |
\& # this prints c4 80 for the UTF\-8 bytes 0xc4 0x80 |
| 668 |
\& print join(" ", unpack("U0(H2)*", pack("U", 0x100))), "\en"; |
| 669 |
.Ve |
| 670 |
.Sp |
| 671 |
Yet another way would be to use the Devel::Peek module: |
| 672 |
.Sp |
| 673 |
.Vb 1 |
| 674 |
\& perl \-MDevel::Peek \-e \*(AqDump(chr(0x100))\*(Aq |
| 675 |
.Ve |
| 676 |
.Sp |
| 677 |
That shows the \f(CW\*(C`UTF8\*(C'\fR flag in FLAGS and both the UTF\-8 bytes |
| 678 |
and Unicode characters in \f(CW\*(C`PV\*(C'\fR. See also later in this document |
| 679 |
the discussion about the \f(CWutf8::is_utf8()\fR function. |
| 680 |
.SS "Advanced Topics" |
| 681 |
.IX Subsection "Advanced Topics" |
| 682 |
.IP \(bu 4 |
| 683 |
String Equivalence |
| 684 |
.Sp |
| 685 |
The question of string equivalence turns somewhat complicated |
| 686 |
in Unicode: what do you mean by "equal"? |
| 687 |
.Sp |
| 688 |
(Is \f(CW\*(C`LATIN CAPITAL LETTER A WITH ACUTE\*(C'\fR equal to |
| 689 |
\&\f(CW\*(C`LATIN CAPITAL LETTER A\*(C'\fR?) |
| 690 |
.Sp |
| 691 |
The short answer is that by default Perl compares equivalence (\f(CW\*(C`eq\*(C'\fR, |
| 692 |
\&\f(CW\*(C`ne\*(C'\fR) based only on code points of the characters. In the above |
| 693 |
case, the answer is no (because 0x00C1 != 0x0041). But sometimes, any |
| 694 |
CAPITAL LETTER A's should be considered equal, or even A's of any case. |
| 695 |
.Sp |
| 696 |
The long answer is that you need to consider character normalization |
| 697 |
and casing issues: see Unicode::Normalize, Unicode Technical Report #15, |
| 698 |
Unicode Normalization Forms <https://www.unicode.org/reports/tr15> and |
| 699 |
sections on case mapping in the Unicode Standard <https://www.unicode.org>. |
| 700 |
.Sp |
| 701 |
As of Perl 5.8.0, the "Full" case-folding of \fICase |
| 702 |
Mappings/SpecialCasing\fR is implemented, but bugs remain in \f(CW\*(C`qr//i\*(C'\fR with them, |
| 703 |
mostly fixed by 5.14, and essentially entirely by 5.18. |
| 704 |
.IP \(bu 4 |
| 705 |
String Collation |
| 706 |
.Sp |
| 707 |
People like to see their strings nicely sorted\-\-or as Unicode |
| 708 |
parlance goes, collated. But again, what do you mean by collate? |
| 709 |
.Sp |
| 710 |
(Does \f(CW\*(C`LATIN CAPITAL LETTER A WITH ACUTE\*(C'\fR come before or after |
| 711 |
\&\f(CW\*(C`LATIN CAPITAL LETTER A WITH GRAVE\*(C'\fR?) |
| 712 |
.Sp |
| 713 |
The short answer is that by default, Perl compares strings (\f(CW\*(C`lt\*(C'\fR, |
| 714 |
\&\f(CW\*(C`le\*(C'\fR, \f(CW\*(C`cmp\*(C'\fR, \f(CW\*(C`ge\*(C'\fR, \f(CW\*(C`gt\*(C'\fR) based only on the code points of the |
| 715 |
characters. In the above case, the answer is "after", since |
| 716 |
\&\f(CW0x00C1\fR > \f(CW0x00C0\fR. |
| 717 |
.Sp |
| 718 |
The long answer is that "it depends", and a good answer cannot be |
| 719 |
given without knowing (at the very least) the language context. |
| 720 |
See Unicode::Collate, and \fIUnicode Collation Algorithm\fR |
| 721 |
<https://www.unicode.org/reports/tr10/> |
| 722 |
.SS Miscellaneous |
| 723 |
.IX Subsection "Miscellaneous" |
| 724 |
.IP \(bu 4 |
| 725 |
Character Ranges and Classes |
| 726 |
.Sp |
| 727 |
Character ranges in regular expression bracketed character classes ( e.g., |
| 728 |
\&\f(CW\*(C`/[a\-z]/\*(C'\fR) and in the \f(CW\*(C`tr///\*(C'\fR (also known as \f(CW\*(C`y///\*(C'\fR) operator are not |
| 729 |
magically Unicode-aware. What this means is that \f(CW\*(C`[A\-Za\-z]\*(C'\fR will not |
| 730 |
magically start to mean "all alphabetic letters" (not that it does mean that |
| 731 |
even for 8\-bit characters; for those, if you are using locales (perllocale), |
| 732 |
use \f(CW\*(C`/[[:alpha:]]/\*(C'\fR; and if not, use the 8\-bit\-aware property \f(CW\*(C`\ep{alpha}\*(C'\fR). |
| 733 |
.Sp |
| 734 |
All the properties that begin with \f(CW\*(C`\ep\*(C'\fR (and its inverse \f(CW\*(C`\eP\*(C'\fR) are actually |
| 735 |
character classes that are Unicode-aware. There are dozens of them, see |
| 736 |
perluniprops. |
| 737 |
.Sp |
| 738 |
Starting in v5.22, you can use Unicode code points as the end points of |
| 739 |
regular expression pattern character ranges, and the range will include |
| 740 |
all Unicode code points that lie between those end points, inclusive. |
| 741 |
.Sp |
| 742 |
.Vb 1 |
| 743 |
\& qr/ [ \eN{U+03} \- \eN{U+20} ] /xx |
| 744 |
.Ve |
| 745 |
.Sp |
| 746 |
includes the code points |
| 747 |
\&\f(CW\*(C`\eN{U+03}\*(C'\fR, \f(CW\*(C`\eN{U+04}\*(C'\fR, ..., \f(CW\*(C`\eN{U+20}\*(C'\fR. |
| 748 |
.Sp |
| 749 |
This also works for ranges in \f(CW\*(C`tr///\*(C'\fR starting in Perl v5.24. |
| 750 |
.IP \(bu 4 |
| 751 |
String-To-Number Conversions |
| 752 |
.Sp |
| 753 |
Unicode does define several other decimal\-\-and numeric\-\-characters |
| 754 |
besides the familiar 0 to 9, such as the Arabic and Indic digits. |
| 755 |
Perl does not support string-to-number conversion for digits other |
| 756 |
than ASCII \f(CW0\fR to \f(CW9\fR (and ASCII \f(CW\*(C`a\*(C'\fR to \f(CW\*(C`f\*(C'\fR for hexadecimal). |
| 757 |
To get safe conversions from any Unicode string, use |
| 758 |
"\fBnum()\fR" in Unicode::UCD. |
| 759 |
.SS "Questions With Answers" |
| 760 |
.IX Subsection "Questions With Answers" |
| 761 |
.IP \(bu 4 |
| 762 |
Will My Old Scripts Break? |
| 763 |
.Sp |
| 764 |
Very probably not. Unless you are generating Unicode characters |
| 765 |
somehow, old behaviour should be preserved. About the only behaviour |
| 766 |
that has changed and which could start generating Unicode is the old |
| 767 |
behaviour of \f(CWchr()\fR where supplying an argument more than 255 |
| 768 |
produced a character modulo 255. \f(CWchr(300)\fR, for example, was equal |
| 769 |
to \f(CWchr(45)\fR or "\-" (in ASCII), now it is LATIN CAPITAL LETTER I WITH |
| 770 |
BREVE. |
| 771 |
.IP \(bu 4 |
| 772 |
How Do I Make My Scripts Work With Unicode? |
| 773 |
.Sp |
| 774 |
Very little work should be needed since nothing changes until you |
| 775 |
generate Unicode data. The most important thing is getting input as |
| 776 |
Unicode; for that, see the earlier I/O discussion. |
| 777 |
To get full seamless Unicode support, add |
| 778 |
\&\f(CW\*(C`use feature \*(Aqunicode_strings\*(Aq\*(C'\fR (or \f(CW\*(C`use v5.12\*(C'\fR or higher) to your |
| 779 |
script. |
| 780 |
.IP \(bu 4 |
| 781 |
How Do I Know Whether My String Is In Unicode? |
| 782 |
.Sp |
| 783 |
You shouldn't have to care. But you may if your Perl is before 5.14.0 |
| 784 |
or you haven't specified \f(CW\*(C`use feature \*(Aqunicode_strings\*(Aq\*(C'\fR or \f(CWuse |
| 785 |
5.012\fR (or higher) because otherwise the rules for the code points |
| 786 |
in the range 128 to 255 are different depending on |
| 787 |
whether the string they are contained within is in Unicode or not. |
| 788 |
(See "When Unicode Does Not Happen" in perlunicode.) |
| 789 |
.Sp |
| 790 |
To determine if a string is in Unicode, use: |
| 791 |
.Sp |
| 792 |
.Vb 1 |
| 793 |
\& print utf8::is_utf8($string) ? 1 : 0, "\en"; |
| 794 |
.Ve |
| 795 |
.Sp |
| 796 |
But note that this doesn't mean that any of the characters in the |
| 797 |
string are necessary UTF\-8 encoded, or that any of the characters have |
| 798 |
code points greater than 0xFF (255) or even 0x80 (128), or that the |
| 799 |
string has any characters at all. All the \f(CWis_utf8()\fR does is to |
| 800 |
return the value of the internal "utf8ness" flag attached to the |
| 801 |
\&\f(CW$string\fR. If the flag is off, the bytes in the scalar are interpreted |
| 802 |
as a single byte encoding. If the flag is on, the bytes in the scalar |
| 803 |
are interpreted as the (variable-length, potentially multi-byte) UTF\-8 encoded |
| 804 |
code points of the characters. Bytes added to a UTF\-8 encoded string are |
| 805 |
automatically upgraded to UTF\-8. If mixed non\-UTF\-8 and UTF\-8 scalars |
| 806 |
are merged (double-quoted interpolation, explicit concatenation, or |
| 807 |
printf/sprintf parameter substitution), the result will be UTF\-8 encoded |
| 808 |
as if copies of the byte strings were upgraded to UTF\-8: for example, |
| 809 |
.Sp |
| 810 |
.Vb 3 |
| 811 |
\& $x = "ab\ex80c"; |
| 812 |
\& $y = "\ex{100}"; |
| 813 |
\& print "$x = $y\en"; |
| 814 |
.Ve |
| 815 |
.Sp |
| 816 |
the output string will be UTF\-8\-encoded \f(CW\*(C`ab\ex80c = \ex{100}\en\*(C'\fR, but |
| 817 |
\&\f(CW$x\fR will stay byte-encoded. |
| 818 |
.Sp |
| 819 |
Sometimes you might really need to know the byte length of a string |
| 820 |
instead of the character length. For that use the \f(CW\*(C`bytes\*(C'\fR pragma |
| 821 |
and the \f(CWlength()\fR function: |
| 822 |
.Sp |
| 823 |
.Vb 6 |
| 824 |
\& my $unicode = chr(0x100); |
| 825 |
\& print length($unicode), "\en"; # will print 1 |
| 826 |
\& use bytes; |
| 827 |
\& print length($unicode), "\en"; # will print 2 |
| 828 |
\& # (the 0xC4 0x80 of the UTF\-8) |
| 829 |
\& no bytes; |
| 830 |
.Ve |
| 831 |
.IP \(bu 4 |
| 832 |
How Do I Find Out What Encoding a File Has? |
| 833 |
.Sp |
| 834 |
You might try Encode::Guess, but it has a number of limitations. |
| 835 |
.IP \(bu 4 |
| 836 |
How Do I Detect Data That's Not Valid In a Particular Encoding? |
| 837 |
.Sp |
| 838 |
Use the \f(CW\*(C`Encode\*(C'\fR package to try converting it. |
| 839 |
For example, |
| 840 |
.Sp |
| 841 |
.Vb 1 |
| 842 |
\& use Encode \*(Aqdecode\*(Aq; |
| 843 |
\& |
| 844 |
\& if (eval { decode(\*(AqUTF\-8\*(Aq, $string, Encode::FB_CROAK); 1 }) { |
| 845 |
\& # $string is valid UTF\-8 |
| 846 |
\& } else { |
| 847 |
\& # $string is not valid UTF\-8 |
| 848 |
\& } |
| 849 |
.Ve |
| 850 |
.Sp |
| 851 |
Or use \f(CW\*(C`unpack\*(C'\fR to try decoding it: |
| 852 |
.Sp |
| 853 |
.Vb 2 |
| 854 |
\& use warnings; |
| 855 |
\& @chars = unpack("C0U*", $string_of_bytes_that_I_think_is_utf8); |
| 856 |
.Ve |
| 857 |
.Sp |
| 858 |
If invalid, a \f(CW\*(C`Malformed UTF\-8 character\*(C'\fR warning is produced. The "C0" means |
| 859 |
"process the string character per character". Without that, the |
| 860 |
\&\f(CW\*(C`unpack("U*", ...)\*(C'\fR would work in \f(CW\*(C`U0\*(C'\fR mode (the default if the format |
| 861 |
string starts with \f(CW\*(C`U\*(C'\fR) and it would return the bytes making up the UTF\-8 |
| 862 |
encoding of the target string, something that will always work. |
| 863 |
.IP \(bu 4 |
| 864 |
How Do I Convert Binary Data Into a Particular Encoding, Or Vice Versa? |
| 865 |
.Sp |
| 866 |
This probably isn't as useful as you might think. |
| 867 |
Normally, you shouldn't need to. |
| 868 |
.Sp |
| 869 |
In one sense, what you are asking doesn't make much sense: encodings |
| 870 |
are for characters, and binary data are not "characters", so converting |
| 871 |
"data" into some encoding isn't meaningful unless you know in what |
| 872 |
character set and encoding the binary data is in, in which case it's |
| 873 |
not just binary data, now is it? |
| 874 |
.Sp |
| 875 |
If you have a raw sequence of bytes that you know should be |
| 876 |
interpreted via a particular encoding, you can use \f(CW\*(C`Encode\*(C'\fR: |
| 877 |
.Sp |
| 878 |
.Vb 2 |
| 879 |
\& use Encode \*(Aqfrom_to\*(Aq; |
| 880 |
\& from_to($data, "iso\-8859\-1", "UTF\-8"); # from latin\-1 to UTF\-8 |
| 881 |
.Ve |
| 882 |
.Sp |
| 883 |
The call to \f(CWfrom_to()\fR changes the bytes in \f(CW$data\fR, but nothing |
| 884 |
material about the nature of the string has changed as far as Perl is |
| 885 |
concerned. Both before and after the call, the string \f(CW$data\fR |
| 886 |
contains just a bunch of 8\-bit bytes. As far as Perl is concerned, |
| 887 |
the encoding of the string remains as "system-native 8\-bit bytes". |
| 888 |
.Sp |
| 889 |
You might relate this to a fictional 'Translate' module: |
| 890 |
.Sp |
| 891 |
.Vb 4 |
| 892 |
\& use Translate; |
| 893 |
\& my $phrase = "Yes"; |
| 894 |
\& Translate::from_to($phrase, \*(Aqenglish\*(Aq, \*(Aqdeutsch\*(Aq); |
| 895 |
\& ## phrase now contains "Ja" |
| 896 |
.Ve |
| 897 |
.Sp |
| 898 |
The contents of the string changes, but not the nature of the string. |
| 899 |
Perl doesn't know any more after the call than before that the |
| 900 |
contents of the string indicates the affirmative. |
| 901 |
.Sp |
| 902 |
Back to converting data. If you have (or want) data in your system's |
| 903 |
native 8\-bit encoding (e.g. Latin\-1, EBCDIC, etc.), you can use |
| 904 |
pack/unpack to convert to/from Unicode. |
| 905 |
.Sp |
| 906 |
.Vb 2 |
| 907 |
\& $native_string = pack("W*", unpack("U*", $Unicode_string)); |
| 908 |
\& $Unicode_string = pack("U*", unpack("W*", $native_string)); |
| 909 |
.Ve |
| 910 |
.Sp |
| 911 |
If you have a sequence of bytes you \fBknow\fR is valid UTF\-8, |
| 912 |
but Perl doesn't know it yet, you can make Perl a believer, too: |
| 913 |
.Sp |
| 914 |
.Vb 2 |
| 915 |
\& $Unicode = $bytes; |
| 916 |
\& utf8::decode($Unicode); |
| 917 |
.Ve |
| 918 |
.Sp |
| 919 |
or: |
| 920 |
.Sp |
| 921 |
.Vb 1 |
| 922 |
\& $Unicode = pack("U0a*", $bytes); |
| 923 |
.Ve |
| 924 |
.Sp |
| 925 |
You can find the bytes that make up a UTF\-8 sequence with |
| 926 |
.Sp |
| 927 |
.Vb 1 |
| 928 |
\& @bytes = unpack("C*", $Unicode_string) |
| 929 |
.Ve |
| 930 |
.Sp |
| 931 |
and you can create well-formed Unicode with |
| 932 |
.Sp |
| 933 |
.Vb 1 |
| 934 |
\& $Unicode_string = pack("U*", 0xff, ...) |
| 935 |
.Ve |
| 936 |
.IP \(bu 4 |
| 937 |
How Do I Display Unicode? How Do I Input Unicode? |
| 938 |
.Sp |
| 939 |
See <http://www.alanwood.net/unicode/> and |
| 940 |
<http://www.cl.cam.ac.uk/~mgk25/unicode.html> |
| 941 |
.IP \(bu 4 |
| 942 |
How Does Unicode Work With Traditional Locales? |
| 943 |
.Sp |
| 944 |
If your locale is a UTF\-8 locale, starting in Perl v5.26, Perl works |
| 945 |
well for all categories; before this, starting with Perl v5.20, it works |
| 946 |
for all categories but \f(CW\*(C`LC_COLLATE\*(C'\fR, which deals with |
| 947 |
sorting and the \f(CW\*(C`cmp\*(C'\fR operator. But note that the standard |
| 948 |
\&\f(CW\*(C`Unicode::Collate\*(C'\fR and \f(CW\*(C`Unicode::Collate::Locale\*(C'\fR modules offer |
| 949 |
much more powerful solutions to collation issues, and work on earlier |
| 950 |
releases. |
| 951 |
.Sp |
| 952 |
For other locales, starting in Perl 5.16, you can specify |
| 953 |
.Sp |
| 954 |
.Vb 1 |
| 955 |
\& use locale \*(Aq:not_characters\*(Aq; |
| 956 |
.Ve |
| 957 |
.Sp |
| 958 |
to get Perl to work well with them. The catch is that you |
| 959 |
have to translate from the locale character set to/from Unicode |
| 960 |
yourself. See "Unicode I/O" above for how to |
| 961 |
.Sp |
| 962 |
.Vb 1 |
| 963 |
\& use open \*(Aq:locale\*(Aq; |
| 964 |
.Ve |
| 965 |
.Sp |
| 966 |
to accomplish this, but full details are in "Unicode and |
| 967 |
UTF\-8" in perllocale, including gotchas that happen if you don't specify |
| 968 |
\&\f(CW\*(C`:not_characters\*(C'\fR. |
| 969 |
.SS "Hexadecimal Notation" |
| 970 |
.IX Subsection "Hexadecimal Notation" |
| 971 |
The Unicode standard prefers using hexadecimal notation because |
| 972 |
that more clearly shows the division of Unicode into blocks of 256 characters. |
| 973 |
Hexadecimal is also simply shorter than decimal. You can use decimal |
| 974 |
notation, too, but learning to use hexadecimal just makes life easier |
| 975 |
with the Unicode standard. The \f(CW\*(C`U+HHHH\*(C'\fR notation uses hexadecimal, |
| 976 |
for example. |
| 977 |
.PP |
| 978 |
The \f(CW\*(C`0x\*(C'\fR prefix means a hexadecimal number, the digits are 0\-9 \fIand\fR |
| 979 |
a\-f (or A\-F, case doesn't matter). Each hexadecimal digit represents |
| 980 |
four bits, or half a byte. \f(CW\*(C`print 0x..., "\en"\*(C'\fR will show a |
| 981 |
hexadecimal number in decimal, and \f(CW\*(C`printf "%x\en", $decimal\*(C'\fR will |
| 982 |
show a decimal number in hexadecimal. If you have just the |
| 983 |
"hex digits" of a hexadecimal number, you can use the \f(CWhex()\fR function. |
| 984 |
.PP |
| 985 |
.Vb 6 |
| 986 |
\& print 0x0009, "\en"; # 9 |
| 987 |
\& print 0x000a, "\en"; # 10 |
| 988 |
\& print 0x000f, "\en"; # 15 |
| 989 |
\& print 0x0010, "\en"; # 16 |
| 990 |
\& print 0x0011, "\en"; # 17 |
| 991 |
\& print 0x0100, "\en"; # 256 |
| 992 |
\& |
| 993 |
\& print 0x0041, "\en"; # 65 |
| 994 |
\& |
| 995 |
\& printf "%x\en", 65; # 41 |
| 996 |
\& printf "%#x\en", 65; # 0x41 |
| 997 |
\& |
| 998 |
\& print hex("41"), "\en"; # 65 |
| 999 |
.Ve |
| 1000 |
.SS "Further Resources" |
| 1001 |
.IX Subsection "Further Resources" |
| 1002 |
.IP \(bu 4 |
| 1003 |
Unicode Consortium |
| 1004 |
.Sp |
| 1005 |
<https://www.unicode.org/> |
| 1006 |
.IP \(bu 4 |
| 1007 |
Unicode FAQ |
| 1008 |
.Sp |
| 1009 |
<https://www.unicode.org/faq/> |
| 1010 |
.IP \(bu 4 |
| 1011 |
Unicode Glossary |
| 1012 |
.Sp |
| 1013 |
<https://www.unicode.org/glossary/> |
| 1014 |
.IP \(bu 4 |
| 1015 |
Unicode Recommended Reading List |
| 1016 |
.Sp |
| 1017 |
The Unicode Consortium has a list of articles and books, some of which |
| 1018 |
give a much more in depth treatment of Unicode: |
| 1019 |
<http://unicode.org/resources/readinglist.html> |
| 1020 |
.IP \(bu 4 |
| 1021 |
Unicode Useful Resources |
| 1022 |
.Sp |
| 1023 |
<https://www.unicode.org/unicode/onlinedat/resources.html> |
| 1024 |
.IP \(bu 4 |
| 1025 |
Unicode and Multilingual Support in HTML, Fonts, Web Browsers and Other Applications |
| 1026 |
.Sp |
| 1027 |
<http://www.alanwood.net/unicode/> |
| 1028 |
.IP \(bu 4 |
| 1029 |
UTF\-8 and Unicode FAQ for Unix/Linux |
| 1030 |
.Sp |
| 1031 |
<http://www.cl.cam.ac.uk/~mgk25/unicode.html> |
| 1032 |
.IP \(bu 4 |
| 1033 |
Legacy Character Sets |
| 1034 |
.Sp |
| 1035 |
<http://www.czyborra.com/> |
| 1036 |
<http://www.eki.ee/letter/> |
| 1037 |
.IP \(bu 4 |
| 1038 |
You can explore various information from the Unicode data files using |
| 1039 |
the \f(CW\*(C`Unicode::UCD\*(C'\fR module. |
| 1040 |
.SH "UNICODE IN OLDER PERLS" |
| 1041 |
.IX Header "UNICODE IN OLDER PERLS" |
| 1042 |
If you cannot upgrade your Perl to 5.8.0 or later, you can still |
| 1043 |
do some Unicode processing by using the modules \f(CW\*(C`Unicode::String\*(C'\fR, |
| 1044 |
\&\f(CW\*(C`Unicode::Map8\*(C'\fR, and \f(CW\*(C`Unicode::Map\*(C'\fR, available from CPAN. |
| 1045 |
If you have the GNU recode installed, you can also use the |
| 1046 |
Perl front-end \f(CW\*(C`Convert::Recode\*(C'\fR for character conversions. |
| 1047 |
.PP |
| 1048 |
The following are fast conversions from ISO 8859\-1 (Latin\-1) bytes |
| 1049 |
to UTF\-8 bytes and back, the code works even with older Perl 5 versions. |
| 1050 |
.PP |
| 1051 |
.Vb 2 |
| 1052 |
\& # ISO 8859\-1 to UTF\-8 |
| 1053 |
\& s/([\ex80\-\exFF])/chr(0xC0|ord($1)>>6).chr(0x80|ord($1)&0x3F)/eg; |
| 1054 |
\& |
| 1055 |
\& # UTF\-8 to ISO 8859\-1 |
| 1056 |
\& s/([\exC2\exC3])([\ex80\-\exBF])/chr(ord($1)<<6&0xC0|ord($2)&0x3F)/eg; |
| 1057 |
.Ve |
| 1058 |
.SH "SEE ALSO" |
| 1059 |
.IX Header "SEE ALSO" |
| 1060 |
perlunitut, perlunicode, Encode, open, utf8, bytes, |
| 1061 |
perlretut, perlrun, Unicode::Collate, Unicode::Normalize, |
| 1062 |
Unicode::UCD |
| 1063 |
.SH ACKNOWLEDGMENTS |
| 1064 |
.IX Header "ACKNOWLEDGMENTS" |
| 1065 |
Thanks to the kind readers of the perl5\[email protected], |
| 1066 |
perl\[email protected], linux\[email protected], and [email protected] |
| 1067 |
mailing lists for their valuable feedback. |
| 1068 |
.SH "AUTHOR, COPYRIGHT, AND LICENSE" |
| 1069 |
.IX Header "AUTHOR, COPYRIGHT, AND LICENSE" |
| 1070 |
Copyright 2001\-2011 Jarkko Hietaniemi <[email protected]>. |
| 1071 |
Now maintained by Perl 5 Porters. |
| 1072 |
.PP |
| 1073 |
This document may be distributed under the same terms as Perl itself. |