3 perlunicode - Unicode support in Perl
7 =head2 Important Caveats
9 Unicode support is an extensive requirement. While Perl does not
10 implement the Unicode standard or the accompanying technical reports
11 from cover to cover, Perl does support many Unicode features.
13 People who want to learn to use Unicode in Perl, should probably read
14 the L<Perl Unicode tutorial, perlunitut|perlunitut> and
15 L<perluniintro>, before reading
16 this reference document.
18 Also, the use of Unicode may present security issues that aren't obvious.
19 Read L<Unicode Security Considerations|http://www.unicode.org/reports/tr36>.
23 =item Safest if you "use feature 'unicode_strings'"
25 In order to preserve backward compatibility, Perl does not turn
26 on full internal Unicode support unless the pragma
27 C<use feature 'unicode_strings'> is specified. (This is automatically
28 selected if you use C<use 5.012> or higher.) Failure to do this can
29 trigger unexpected surprises. See L</The "Unicode Bug"> below.
31 This pragma doesn't affect I/O, and there are still several places
32 where Unicode isn't fully supported, such as in filenames.
34 =item Input and Output Layers
36 Perl knows when a filehandle uses Perl's internal Unicode encodings
37 (UTF-8, or UTF-EBCDIC if in EBCDIC) if the filehandle is opened with
38 the ":encoding(utf8)" layer. Other encodings can be converted to Perl's
39 encoding on input or from Perl's encoding on output by use of the
40 ":encoding(...)" layer. See L<open>.
42 To indicate that Perl source itself is in UTF-8, use C<use utf8;>.
44 =item C<use utf8> still needed to enable UTF-8/UTF-EBCDIC in scripts
46 As a compatibility measure, the C<use utf8> pragma must be explicitly
47 included to enable recognition of UTF-8 in the Perl scripts themselves
48 (in string or regular expression literals, or in identifier names) on
49 ASCII-based machines or to recognize UTF-EBCDIC on EBCDIC-based
50 machines. B<These are the only times when an explicit C<use utf8>
51 is needed.> See L<utf8>.
53 =item BOM-marked scripts and UTF-16 scripts autodetected
55 If a Perl script begins marked with the Unicode BOM (UTF-16LE, UTF16-BE,
56 or UTF-8), or if the script looks like non-BOM-marked UTF-16 of either
57 endianness, Perl will correctly read in the script as Unicode.
58 (BOMless UTF-8 cannot be effectively recognized or differentiated from
59 ISO 8859-1 or other eight-bit encodings.)
61 =item C<use encoding> needed to upgrade non-Latin-1 byte strings
63 By default, there is a fundamental asymmetry in Perl's Unicode model:
64 implicit upgrading from byte strings to Unicode strings assumes that
65 they were encoded in I<ISO 8859-1 (Latin-1)>, but Unicode strings are
66 downgraded with UTF-8 encoding. This happens because the first 256
67 codepoints in Unicode happens to agree with Latin-1.
69 See L</"Byte and Character Semantics"> for more details.
73 =head2 Byte and Character Semantics
75 Beginning with version 5.6, Perl uses logically-wide characters to
76 represent strings internally.
78 Starting in Perl 5.14, Perl-level operations work with
79 characters rather than bytes within the scope of a
80 C<L<use feature 'unicode_strings'|feature>> (or equivalently
81 C<use 5.012> or higher). (This is not true if bytes have been
82 explicitly requested by C<L<use bytes|bytes>>, nor necessarily true
83 for interactions with the platform's operating system.)
85 For earlier Perls, and when C<unicode_strings> is not in effect, Perl
86 provides a fairly safe environment that can handle both types of
87 semantics in programs. For operations where Perl can unambiguously
88 decide that the input data are characters, Perl switches to character
89 semantics. For operations where this determination cannot be made
90 without additional information from the user, Perl decides in favor of
91 compatibility and chooses to use byte semantics.
93 When C<use locale> (but not C<use locale ':not_characters'>) is in
94 effect, Perl uses the semantics associated with the current locale.
95 (C<use locale> overrides C<use feature 'unicode_strings'> in the same scope;
96 while C<use locale ':not_characters'> effectively also selects
97 C<use feature 'unicode_strings'> in its scope; see L<perllocale>.)
98 Otherwise, Perl uses the platform's native
99 byte semantics for characters whose code points are less than 256, and
100 Unicode semantics for those greater than 255. On EBCDIC platforms, this
101 is almost seamless, as the EBCDIC code pages that Perl handles are
102 equivalent to Unicode's first 256 code points. (The exception is that
103 EBCDIC regular expression case-insensitive matching rules are not as
104 as robust as Unicode's.) But on ASCII platforms, Perl uses US-ASCII
105 (or Basic Latin in Unicode terminology) byte semantics, meaning that characters
106 whose ordinal numbers are in the range 128 - 255 are undefined except for their
107 ordinal numbers. This means that none have case (upper and lower), nor are any
108 a member of character classes, like C<[:alpha:]> or C<\w>. (But all do belong
109 to the C<\W> class or the Perl regular expression extension C<[:^alpha:]>.)
111 This behavior preserves compatibility with earlier versions of Perl,
112 which allowed byte semantics in Perl operations only if
113 none of the program's inputs were marked as being a source of Unicode
114 character data. Such data may come from filehandles, from calls to
115 external programs, from information provided by the system (such as %ENV),
116 or from literals and constants in the source text.
118 The C<utf8> pragma is primarily a compatibility device that enables
119 recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
120 Note that this pragma is only required while Perl defaults to byte
121 semantics; when character semantics become the default, this pragma
122 may become a no-op. See L<utf8>.
124 If strings operating under byte semantics and strings with Unicode
125 character data are concatenated, the new string will have
126 character semantics. This can cause surprises: See L</BUGS>, below.
127 You can choose to be warned when this happens. See L<encoding::warnings>.
129 Under character semantics, many operations that formerly operated on
130 bytes now operate on characters. A character in Perl is
131 logically just a number ranging from 0 to 2**31 or so. Larger
132 characters may encode into longer sequences of bytes internally, but
133 this internal detail is mostly hidden for Perl code.
134 See L<perluniintro> for more.
136 =head2 Effects of Character Semantics
138 Character semantics have the following effects:
144 Strings--including hash keys--and regular expression patterns may
145 contain characters that have an ordinal value larger than 255.
147 If you use a Unicode editor to edit your program, Unicode characters may
148 occur directly within the literal strings in UTF-8 encoding, or UTF-16.
149 (The former requires a BOM or C<use utf8>, the latter requires a BOM.)
151 Unicode characters can also be added to a string by using the C<\N{U+...}>
152 notation. The Unicode code for the desired character, in hexadecimal,
153 should be placed in the braces, after the C<U>. For instance, a smiley face is
156 Alternatively, you can use the C<\x{...}> notation for characters 0x100 and
157 above. For characters below 0x100 you may get byte semantics instead of
158 character semantics; see L</The "Unicode Bug">. On EBCDIC machines there is
159 the additional problem that the value for such characters gives the EBCDIC
160 character rather than the Unicode one, thus it is more portable to use
161 C<\N{U+...}> instead.
163 Additionally, you can use the C<\N{...}> notation and put the official
164 Unicode character name within the braces, such as
165 C<\N{WHITE SMILING FACE}>. This automatically loads the L<charnames>
166 module with the C<:full> and C<:short> options. If you prefer different
167 options for this module, you can instead, before the C<\N{...}>,
168 explicitly load it with your desired options; for example,
170 use charnames ':loose';
174 If an appropriate L<encoding> is specified, identifiers within the
175 Perl script may contain Unicode alphanumeric characters, including
176 ideographs. Perl does not currently attempt to canonicalize variable
181 Regular expressions match characters instead of bytes. "." matches
182 a character instead of a byte.
186 Bracketed character classes in regular expressions match characters instead of
187 bytes and match against the character properties specified in the
188 Unicode properties database. C<\w> can be used to match a Japanese
189 ideograph, for instance.
193 Named Unicode properties, scripts, and block ranges may be used (like bracketed
194 character classes) by using the C<\p{}> "matches property" construct and
195 the C<\P{}> negation, "doesn't match property".
196 See L</"Unicode Character Properties"> for more details.
198 You can define your own character properties and use them
199 in the regular expression with the C<\p{}> or C<\P{}> construct.
200 See L</"User-Defined Character Properties"> for more details.
204 The special pattern C<\X> matches a logical character, an "extended grapheme
205 cluster" in Standardese. In Unicode what appears to the user to be a single
206 character, for example an accented C<G>, may in fact be composed of a sequence
207 of characters, in this case a C<G> followed by an accent character. C<\X>
208 will match the entire sequence.
212 The C<tr///> operator translates characters instead of bytes. Note
213 that the C<tr///CU> functionality has been removed. For similar
214 functionality see pack('U0', ...) and pack('C0', ...).
218 Case translation operators use the Unicode case translation tables
219 when character input is provided. Note that C<uc()>, or C<\U> in
220 interpolated strings, translates to uppercase, while C<ucfirst>,
221 or C<\u> in interpolated strings, translates to titlecase in languages
222 that make the distinction (which is equivalent to uppercase in languages
223 without the distinction).
227 Most operators that deal with positions or lengths in a string will
228 automatically switch to using character positions, including
229 C<chop()>, C<chomp()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
230 C<sprintf()>, C<write()>, and C<length()>. An operator that
231 specifically does not switch is C<vec()>. Operators that really don't
232 care include operators that treat strings as a bucket of bits such as
233 C<sort()>, and operators dealing with filenames.
237 The C<pack()>/C<unpack()> letter C<C> does I<not> change, since it is often
238 used for byte-oriented formats. Again, think C<char> in the C language.
240 There is a new C<U> specifier that converts between Unicode characters
241 and code points. There is also a C<W> specifier that is the equivalent of
242 C<chr>/C<ord> and properly handles character values even if they are above 255.
246 The C<chr()> and C<ord()> functions work on characters, similar to
247 C<pack("W")> and C<unpack("W")>, I<not> C<pack("C")> and
248 C<unpack("C")>. C<pack("C")> and C<unpack("C")> are methods for
249 emulating byte-oriented C<chr()> and C<ord()> on Unicode strings.
250 While these methods reveal the internal encoding of Unicode strings,
251 that is not something one normally needs to care about at all.
255 The bit string operators, C<& | ^ ~>, can operate on character data.
256 However, for backward compatibility, such as when using bit string
257 operations when characters are all less than 256 in ordinal value, one
258 should not use C<~> (the bit complement) with characters of both
259 values less than 256 and values greater than 256. Most importantly,
260 DeMorgan's laws (C<~($x|$y) eq ~$x&~$y> and C<~($x&$y) eq ~$x|~$y>)
261 will not hold. The reason for this mathematical I<faux pas> is that
262 the complement cannot return B<both> the 8-bit (byte-wide) bit
263 complement B<and> the full character-wide bit complement.
267 There is a CPAN module, L<Unicode::Casing>, which allows you to define
268 your own mappings to be used in C<lc()>, C<lcfirst()>, C<uc()>,
269 C<ucfirst()>, and C<fc> (or their double-quoted string inlined
270 versions such as C<\U>).
271 (Prior to Perl 5.16, this functionality was partially provided
272 in the Perl core, but suffered from a number of insurmountable
273 drawbacks, so the CPAN module was written instead.)
281 And finally, C<scalar reverse()> reverses by character rather than by byte.
285 =head2 Unicode Character Properties
287 (The only time that Perl considers a sequence of individual code
288 points as a single logical character is in the C<\X> construct, already
289 mentioned above. Therefore "character" in this discussion means a single
292 Very nearly all Unicode character properties are accessible through
293 regular expressions by using the C<\p{}> "matches property" construct
294 and the C<\P{}> "doesn't match property" for its negation.
296 For instance, C<\p{Uppercase}> matches any single character with the Unicode
297 "Uppercase" property, while C<\p{L}> matches any character with a
298 General_Category of "L" (letter) property. Brackets are not
299 required for single letter property names, so C<\p{L}> is equivalent to C<\pL>.
301 More formally, C<\p{Uppercase}> matches any single character whose Unicode
302 Uppercase property value is True, and C<\P{Uppercase}> matches any character
303 whose Uppercase property value is False, and they could have been written as
304 C<\p{Uppercase=True}> and C<\p{Uppercase=False}>, respectively.
306 This formality is needed when properties are not binary; that is, if they can
307 take on more values than just True and False. For example, the Bidi_Class (see
308 L</"Bidirectional Character Types"> below), can take on several different
309 values, such as Left, Right, Whitespace, and others. To match these, one needs
310 to specify both the property name (Bidi_Class), AND the value being
312 (Left, Right, etc.). This is done, as in the examples above, by having the
313 two components separated by an equal sign (or interchangeably, a colon), like
314 C<\p{Bidi_Class: Left}>.
316 All Unicode-defined character properties may be written in these compound forms
317 of C<\p{property=value}> or C<\p{property:value}>, but Perl provides some
318 additional properties that are written only in the single form, as well as
319 single-form short-cuts for all binary properties and certain others described
320 below, in which you may omit the property name and the equals or colon
323 Most Unicode character properties have at least two synonyms (or aliases if you
324 prefer): a short one that is easier to type and a longer one that is more
325 descriptive and hence easier to understand. Thus the "L" and "Letter" properties
326 above are equivalent and can be used interchangeably. Likewise,
327 "Upper" is a synonym for "Uppercase", and we could have written
328 C<\p{Uppercase}> equivalently as C<\p{Upper}>. Also, there are typically
329 various synonyms for the values the property can be. For binary properties,
330 "True" has 3 synonyms: "T", "Yes", and "Y"; and "False has correspondingly "F",
331 "No", and "N". But be careful. A short form of a value for one property may
332 not mean the same thing as the same short form for another. Thus, for the
333 General_Category property, "L" means "Letter", but for the Bidi_Class property,
334 "L" means "Left". A complete list of properties and synonyms is in
337 Upper/lower case differences in property names and values are irrelevant;
338 thus C<\p{Upper}> means the same thing as C<\p{upper}> or even C<\p{UpPeR}>.
339 Similarly, you can add or subtract underscores anywhere in the middle of a
340 word, so that these are also equivalent to C<\p{U_p_p_e_r}>. And white space
341 is irrelevant adjacent to non-word characters, such as the braces and the equals
342 or colon separators, so C<\p{ Upper }> and C<\p{ Upper_case : Y }> are
343 equivalent to these as well. In fact, white space and even
344 hyphens can usually be added or deleted anywhere. So even C<\p{ Up-per case = Yes}> is
345 equivalent. All this is called "loose-matching" by Unicode. The few places
346 where stricter matching is used is in the middle of numbers, and in the Perl
347 extension properties that begin or end with an underscore. Stricter matching
348 cares about white space (except adjacent to non-word characters),
349 hyphens, and non-interior underscores.
351 You can also use negation in both C<\p{}> and C<\P{}> by introducing a caret
352 (^) between the first brace and the property name: C<\p{^Tamil}> is
353 equal to C<\P{Tamil}>.
355 Almost all properties are immune to case-insensitive matching. That is,
356 adding a C</i> regular expression modifier does not change what they
357 match. There are two sets that are affected.
361 and C<Titlecase_Letter>,
362 all of which match C<Cased_Letter> under C</i> matching.
363 And the second set is
367 all of which match C<Cased> under C</i> matching.
368 This set also includes its subsets C<PosixUpper> and C<PosixLower> both
369 of which under C</i> matching match C<PosixAlpha>.
370 (The difference between these sets is that some things, such as Roman
371 numerals, come in both upper and lower case so they are C<Cased>, but aren't considered
372 letters, so they aren't C<Cased_Letter>s.)
374 The result is undefined if you try to match a non-Unicode code point
375 (that is, one above 0x10FFFF) against a Unicode property. Currently, a
376 warning is raised, and the match will fail. In some cases, this is
377 counterintuitive, as both these fail:
379 chr(0x110000) =~ \p{ASCII_Hex_Digit=True} # Fails.
380 chr(0x110000) =~ \p{ASCII_Hex_Digit=False} # Fails!
382 =head3 B<General_Category>
384 Every Unicode character is assigned a general category, which is the "most
385 usual categorization of a character" (from
386 L<http://www.unicode.org/reports/tr44>).
388 The compound way of writing these is like C<\p{General_Category=Number}>
389 (short, C<\p{gc:n}>). But Perl furnishes shortcuts in which everything up
390 through the equal or colon separator is omitted. So you can instead just write
393 Here are the short and long forms of the General Category properties:
398 LC, L& Cased_Letter (that is: [\p{Ll}\p{Lu}\p{Lt}])
411 Nd Decimal_Number (also Digit)
415 P Punctuation (also Punct)
416 Pc Connector_Punctuation
420 Pi Initial_Punctuation
421 (may behave like Ps or Pe depending on usage)
423 (may behave like Ps or Pe depending on usage)
435 Zp Paragraph_Separator
438 Cc Control (also Cntrl)
444 Single-letter properties match all characters in any of the
445 two-letter sub-properties starting with the same letter.
446 C<LC> and C<L&> are special: both are aliases for the set consisting of everything matched by C<Ll>, C<Lu>, and C<Lt>.
448 =head3 B<Bidirectional Character Types>
450 Because scripts differ in their directionality (Hebrew and Arabic are
451 written right to left, for example) Unicode supplies these properties in
452 the Bidi_Class class:
457 LRE Left-to-Right Embedding
458 LRO Left-to-Right Override
461 RLE Right-to-Left Embedding
462 RLO Right-to-Left Override
463 PDF Pop Directional Format
465 ES European Separator
466 ET European Terminator
471 B Paragraph Separator
476 This property is always written in the compound form.
477 For example, C<\p{Bidi_Class:R}> matches characters that are normally
478 written right to left.
482 The world's languages are written in many different scripts. This sentence
483 (unless you're reading it in translation) is written in Latin, while Russian is
484 written in Cyrillic, and Greek is written in, well, Greek; Japanese mainly in
485 Hiragana or Katakana. There are many more.
487 The Unicode Script and Script_Extensions properties give what script a
488 given character is in. Either property can be specified with the
490 C<\p{Script=Hebrew}> (short: C<\p{sc=hebr}>), or
491 C<\p{Script_Extensions=Javanese}> (short: C<\p{scx=java}>).
492 In addition, Perl furnishes shortcuts for all
493 C<Script> property names. You can omit everything up through the equals
494 (or colon), and simply write C<\p{Latin}> or C<\P{Cyrillic}>.
495 (This is not true for C<Script_Extensions>, which is required to be
496 written in the compound form.)
498 The difference between these two properties involves characters that are
499 used in multiple scripts. For example the digits '0' through '9' are
500 used in many parts of the world. These are placed in a script named
501 C<Common>. Other characters are used in just a few scripts. For
502 example, the "KATAKANA-HIRAGANA DOUBLE HYPHEN" is used in both Japanese
503 scripts, Katakana and Hiragana, but nowhere else. The C<Script>
504 property places all characters that are used in multiple scripts in the
505 C<Common> script, while the C<Script_Extensions> property places those
506 that are used in only a few scripts into each of those scripts; while
507 still using C<Common> for those used in many scripts. Thus both these
510 "0" =~ /\p{sc=Common}/ # Matches
511 "0" =~ /\p{scx=Common}/ # Matches
513 and only the first of these match:
515 "\N{KATAKANA-HIRAGANA DOUBLE HYPHEN}" =~ /\p{sc=Common} # Matches
516 "\N{KATAKANA-HIRAGANA DOUBLE HYPHEN}" =~ /\p{scx=Common} # No match
518 And only the last two of these match:
520 "\N{KATAKANA-HIRAGANA DOUBLE HYPHEN}" =~ /\p{sc=Hiragana} # No match
521 "\N{KATAKANA-HIRAGANA DOUBLE HYPHEN}" =~ /\p{sc=Katakana} # No match
522 "\N{KATAKANA-HIRAGANA DOUBLE HYPHEN}" =~ /\p{scx=Hiragana} # Matches
523 "\N{KATAKANA-HIRAGANA DOUBLE HYPHEN}" =~ /\p{scx=Katakana} # Matches
525 C<Script_Extensions> is thus an improved C<Script>, in which there are
526 fewer characters in the C<Common> script, and correspondingly more in
527 other scripts. It is new in Unicode version 6.0, and its data are likely
528 to change significantly in later releases, as things get sorted out.
530 (Actually, besides C<Common>, the C<Inherited> script, contains
531 characters that are used in multiple scripts. These are modifier
532 characters which modify other characters, and inherit the script value
533 of the controlling character. Some of these are used in many scripts,
534 and so go into C<Inherited> in both C<Script> and C<Script_Extensions>.
535 Others are used in just a few scripts, so are in C<Inherited> in
536 C<Script>, but not in C<Script_Extensions>.)
538 It is worth stressing that there are several different sets of digits in
539 Unicode that are equivalent to 0-9 and are matchable by C<\d> in a
540 regular expression. If they are used in a single language only, they
541 are in that language's C<Script> and C<Script_Extension>. If they are
542 used in more than one script, they will be in C<sc=Common>, but only
543 if they are used in many scripts should they be in C<scx=Common>.
545 A complete list of scripts and their shortcuts is in L<perluniprops>.
547 =head3 B<Use of "Is" Prefix>
549 For backward compatibility (with Perl 5.6), all properties mentioned
550 so far may have C<Is> or C<Is_> prepended to their name, so C<\P{Is_Lu}>, for
551 example, is equal to C<\P{Lu}>, and C<\p{IsScript:Arabic}> is equal to
556 In addition to B<scripts>, Unicode also defines B<blocks> of
557 characters. The difference between scripts and blocks is that the
558 concept of scripts is closer to natural languages, while the concept
559 of blocks is more of an artificial grouping based on groups of Unicode
560 characters with consecutive ordinal values. For example, the "Basic Latin"
561 block is all characters whose ordinals are between 0 and 127, inclusive; in
562 other words, the ASCII characters. The "Latin" script contains some letters
563 from this as well as several other blocks, like "Latin-1 Supplement",
564 "Latin Extended-A", etc., but it does not contain all the characters from
565 those blocks. It does not, for example, contain the digits 0-9, because
566 those digits are shared across many scripts, and hence are in the
569 For more about scripts versus blocks, see UAX#24 "Unicode Script Property":
570 L<http://www.unicode.org/reports/tr24>
572 The C<Script> or C<Script_Extensions> properties are likely to be the
573 ones you want to use when processing
574 natural language; the Block property may occasionally be useful in working
575 with the nuts and bolts of Unicode.
577 Block names are matched in the compound form, like C<\p{Block: Arrows}> or
578 C<\p{Blk=Hebrew}>. Unlike most other properties, only a few block names have a
579 Unicode-defined short name. But Perl does provide a (slight) shortcut: You
580 can say, for example C<\p{In_Arrows}> or C<\p{In_Hebrew}>. For backwards
581 compatibility, the C<In> prefix may be omitted if there is no naming conflict
582 with a script or any other property, and you can even use an C<Is> prefix
583 instead in those cases. But it is not a good idea to do this, for a couple
590 It is confusing. There are many naming conflicts, and you may forget some.
591 For example, C<\p{Hebrew}> means the I<script> Hebrew, and NOT the I<block>
592 Hebrew. But would you remember that 6 months from now?
596 It is unstable. A new version of Unicode may pre-empt the current meaning by
597 creating a property with the same name. There was a time in very early Unicode
598 releases when C<\p{Hebrew}> would have matched the I<block> Hebrew; now it
603 Some people prefer to always use C<\p{Block: foo}> and C<\p{Script: bar}>
604 instead of the shortcuts, whether for clarity, because they can't remember the
605 difference between 'In' and 'Is' anyway, or they aren't confident that those who
606 eventually will read their code will know that difference.
608 A complete list of blocks and their shortcuts is in L<perluniprops>.
610 =head3 B<Other Properties>
612 There are many more properties than the very basic ones described here.
613 A complete list is in L<perluniprops>.
615 Unicode defines all its properties in the compound form, so all single-form
616 properties are Perl extensions. Most of these are just synonyms for the
617 Unicode ones, but some are genuine extensions, including several that are in
618 the compound form. And quite a few of these are actually recommended by Unicode
619 (in L<http://www.unicode.org/reports/tr18>).
621 This section gives some details on all extensions that aren't just
622 synonyms for compound-form Unicode properties
623 (for those properties, you'll have to refer to the
624 L<Unicode Standard|http://www.unicode.org/reports/tr44>.
630 This matches any of the 1_114_112 Unicode code points. It is a synonym for
633 =item B<C<\p{Alnum}>>
635 This matches any C<\p{Alphabetic}> or C<\p{Decimal_Number}> character.
639 This matches any of the 1_114_112 Unicode code points. It is a synonym for
642 =item B<C<\p{ASCII}>>
644 This matches any of the 128 characters in the US-ASCII character set,
645 which is a subset of Unicode.
647 =item B<C<\p{Assigned}>>
649 This matches any assigned code point; that is, any code point whose general
650 category is not Unassigned (or equivalently, not Cn).
652 =item B<C<\p{Blank}>>
654 This is the same as C<\h> and C<\p{HorizSpace}>: A character that changes the
655 spacing horizontally.
657 =item B<C<\p{Decomposition_Type: Non_Canonical}>> (Short: C<\p{Dt=NonCanon}>)
659 Matches a character that has a non-canonical decomposition.
661 To understand the use of this rarely used property=value combination, it is
662 necessary to know some basics about decomposition.
663 Consider a character, say H. It could appear with various marks around it,
664 such as an acute accent, or a circumflex, or various hooks, circles, arrows,
665 I<etc.>, above, below, to one side or the other, etc. There are many
666 possibilities among the world's languages. The number of combinations is
667 astronomical, and if there were a character for each combination, it would
668 soon exhaust Unicode's more than a million possible characters. So Unicode
669 took a different approach: there is a character for the base H, and a
670 character for each of the possible marks, and these can be variously combined
671 to get a final logical character. So a logical character--what appears to be a
672 single character--can be a sequence of more than one individual characters.
673 This is called an "extended grapheme cluster"; Perl furnishes the C<\X>
674 regular expression construct to match such sequences.
676 But Unicode's intent is to unify the existing character set standards and
677 practices, and several pre-existing standards have single characters that
678 mean the same thing as some of these combinations. An example is ISO-8859-1,
679 which has quite a few of these in the Latin-1 range, an example being "LATIN
680 CAPITAL LETTER E WITH ACUTE". Because this character was in this pre-existing
681 standard, Unicode added it to its repertoire. But this character is considered
682 by Unicode to be equivalent to the sequence consisting of the character
683 "LATIN CAPITAL LETTER E" followed by the character "COMBINING ACUTE ACCENT".
685 "LATIN CAPITAL LETTER E WITH ACUTE" is called a "pre-composed" character, and
686 its equivalence with the sequence is called canonical equivalence. All
687 pre-composed characters are said to have a decomposition (into the equivalent
688 sequence), and the decomposition type is also called canonical.
690 However, many more characters have a different type of decomposition, a
691 "compatible" or "non-canonical" decomposition. The sequences that form these
692 decompositions are not considered canonically equivalent to the pre-composed
693 character. An example, again in the Latin-1 range, is the "SUPERSCRIPT ONE".
694 It is somewhat like a regular digit 1, but not exactly; its decomposition
695 into the digit 1 is called a "compatible" decomposition, specifically a
696 "super" decomposition. There are several such compatibility
697 decompositions (see L<http://www.unicode.org/reports/tr44>), including one
698 called "compat", which means some miscellaneous type of decomposition
699 that doesn't fit into the decomposition categories that Unicode has chosen.
701 Note that most Unicode characters don't have a decomposition, so their
702 decomposition type is "None".
704 For your convenience, Perl has added the C<Non_Canonical> decomposition
705 type to mean any of the several compatibility decompositions.
707 =item B<C<\p{Graph}>>
709 Matches any character that is graphic. Theoretically, this means a character
710 that on a printer would cause ink to be used.
712 =item B<C<\p{HorizSpace}>>
714 This is the same as C<\h> and C<\p{Blank}>: a character that changes the
715 spacing horizontally.
719 This is a synonym for C<\p{Present_In=*}>
721 =item B<C<\p{PerlSpace}>>
723 This is the same as C<\s>, restricted to ASCII, namely C<S<[ \f\n\r\t]>>.
725 Mnemonic: Perl's (original) space
727 =item B<C<\p{PerlWord}>>
729 This is the same as C<\w>, restricted to ASCII, namely C<[A-Za-z0-9_]>
731 Mnemonic: Perl's (original) word.
733 =item B<C<\p{Posix...}>>
735 There are several of these, which are equivalents using the C<\p>
736 notation for Posix classes and are described in
737 L<perlrecharclass/POSIX Character Classes>.
739 =item B<C<\p{Present_In: *}>> (Short: C<\p{In=*}>)
741 This property is used when you need to know in what Unicode version(s) a
744 The "*" above stands for some two digit Unicode version number, such as
745 C<1.1> or C<4.0>; or the "*" can also be C<Unassigned>. This property will
746 match the code points whose final disposition has been settled as of the
747 Unicode release given by the version number; C<\p{Present_In: Unassigned}>
748 will match those code points whose meaning has yet to be assigned.
750 For example, C<U+0041> "LATIN CAPITAL LETTER A" was present in the very first
751 Unicode release available, which is C<1.1>, so this property is true for all
752 valid "*" versions. On the other hand, C<U+1EFF> was not assigned until version
753 5.1 when it became "LATIN SMALL LETTER Y WITH LOOP", so the only "*" that
754 would match it are 5.1, 5.2, and later.
756 Unicode furnishes the C<Age> property from which this is derived. The problem
757 with Age is that a strict interpretation of it (which Perl takes) has it
758 matching the precise release a code point's meaning is introduced in. Thus
759 C<U+0041> would match only 1.1; and C<U+1EFF> only 5.1. This is not usually what
762 Some non-Perl implementations of the Age property may change its meaning to be
763 the same as the Perl Present_In property; just be aware of that.
765 Another confusion with both these properties is that the definition is not
766 that the code point has been I<assigned>, but that the meaning of the code point
767 has been I<determined>. This is because 66 code points will always be
768 unassigned, and so the Age for them is the Unicode version in which the decision
769 to make them so was made. For example, C<U+FDD0> is to be permanently
770 unassigned to a character, and the decision to do that was made in version 3.1,
771 so C<\p{Age=3.1}> matches this character, as also does C<\p{Present_In: 3.1}> and up.
773 =item B<C<\p{Print}>>
775 This matches any character that is graphical or blank, except controls.
777 =item B<C<\p{SpacePerl}>>
779 This is the same as C<\s>, including beyond ASCII.
781 Mnemonic: Space, as modified by Perl. (It doesn't include the vertical tab
782 which both the Posix standard and Unicode consider white space.)
784 =item B<C<\p{Title}>> and B<C<\p{Titlecase}>>
786 Under case-sensitive matching, these both match the same code points as
787 C<\p{General Category=Titlecase_Letter}> (C<\p{gc=lt}>). The difference
788 is that under C</i> caseless matching, these match the same as
789 C<\p{Cased}>, whereas C<\p{gc=lt}> matches C<\p{Cased_Letter>).
791 =item B<C<\p{VertSpace}>>
793 This is the same as C<\v>: A character that changes the spacing vertically.
797 This is the same as C<\w>, including over 100_000 characters beyond ASCII.
799 =item B<C<\p{XPosix...}>>
801 There are several of these, which are the standard Posix classes
802 extended to the full Unicode range. They are described in
803 L<perlrecharclass/POSIX Character Classes>.
807 =head2 User-Defined Character Properties
809 You can define your own binary character properties by defining subroutines
810 whose names begin with "In" or "Is". The subroutines can be defined in any
811 package. The user-defined properties can be used in the regular expression
812 C<\p> and C<\P> constructs; if you are using a user-defined property from a
813 package other than the one you are in, you must specify its package in the
814 C<\p> or C<\P> construct.
816 # assuming property Is_Foreign defined in Lang::
817 package main; # property package name required
818 if ($txt =~ /\p{Lang::IsForeign}+/) { ... }
820 package Lang; # property package name not required
821 if ($txt =~ /\p{IsForeign}+/) { ... }
824 Note that the effect is compile-time and immutable once defined.
825 However, the subroutines are passed a single parameter, which is 0 if
826 case-sensitive matching is in effect and non-zero if caseless matching
827 is in effect. The subroutine may return different values depending on
828 the value of the flag, and one set of values will immutably be in effect
829 for all case-sensitive matches, and the other set for all case-insensitive
832 Note that if the regular expression is tainted, then Perl will die rather
833 than calling the subroutine, where the name of the subroutine is
834 determined by the tainted data.
836 The subroutines must return a specially-formatted string, with one
837 or more newline-separated lines. Each line must be one of the following:
843 A single hexadecimal number denoting a Unicode code point to include.
847 Two hexadecimal numbers separated by horizontal whitespace (space or
848 tabular characters) denoting a range of Unicode code points to include.
852 Something to include, prefixed by "+": a built-in character
853 property (prefixed by "utf8::") or a fully qualified (including package
854 name) user-defined character property,
855 to represent all the characters in that property; two hexadecimal code
856 points for a range; or a single hexadecimal code point.
860 Something to exclude, prefixed by "-": an existing character
861 property (prefixed by "utf8::") or a fully qualified (including package
862 name) user-defined character property,
863 to represent all the characters in that property; two hexadecimal code
864 points for a range; or a single hexadecimal code point.
868 Something to negate, prefixed "!": an existing character
869 property (prefixed by "utf8::") or a fully qualified (including package
870 name) user-defined character property,
871 to represent all the characters in that property; two hexadecimal code
872 points for a range; or a single hexadecimal code point.
876 Something to intersect with, prefixed by "&": an existing character
877 property (prefixed by "utf8::") or a fully qualified (including package
878 name) user-defined character property,
879 for all the characters except the characters in the property; two
880 hexadecimal code points for a range; or a single hexadecimal code point.
884 For example, to define a property that covers both the Japanese
885 syllabaries (hiragana and katakana), you can define
894 Imagine that the here-doc end marker is at the beginning of the line.
895 Now you can use C<\p{InKana}> and C<\P{InKana}>.
897 You could also have used the existing block property names:
906 Suppose you wanted to match only the allocated characters,
907 not the raw block ranges: in other words, you want to remove
918 The negation is useful for defining (surprise!) negated classes.
928 This will match all non-Unicode code points, since every one of them is
929 not in Kana. You can use intersection to exclude these, if desired, as
930 this modified example shows:
941 C<&utf8::Any> must be the last line in the definition.
943 Intersection is used generally for getting the common characters matched
944 by two (or more) classes. It's important to remember not to use "&" for
945 the first set; that would be intersecting with nothing, resulting in an
948 (Note that official Unicode properties differ from these in that they
949 automatically exclude non-Unicode code points and a warning is raised if
950 a match is attempted on one of those.)
952 =head2 User-Defined Case Mappings (for serious hackers only)
954 B<This feature has been removed as of Perl 5.16.>
955 The CPAN module L<Unicode::Casing> provides better functionality without
956 the drawbacks that this feature had. If you are using a Perl earlier
957 than 5.16, this feature was most fully documented in the 5.14 version of
959 L<http://perldoc.perl.org/5.14.0/perlunicode.html#User-Defined-Case-Mappings-%28for-serious-hackers-only%29>
961 =head2 Character Encodings for Input and Output
965 =head2 Unicode Regular Expression Support Level
967 The following list of Unicode supported features for regular expressions describes
968 all features currently directly supported by core Perl. The references to "Level N"
969 and the section numbers refer to the Unicode Technical Standard #18,
970 "Unicode Regular Expressions", version 13, from August 2008.
976 Level 1 - Basic Unicode Support
978 RL1.1 Hex Notation - done [1]
979 RL1.2 Properties - done [2][3]
980 RL1.2a Compatibility Properties - done [4]
981 RL1.3 Subtraction and Intersection - MISSING [5]
982 RL1.4 Simple Word Boundaries - done [6]
983 RL1.5 Simple Loose Matches - done [7]
984 RL1.6 Line Boundaries - MISSING [8][9]
985 RL1.7 Supplementary Code Points - done [10]
989 [3] supports not only minimal list, but all Unicode character
990 properties (see Unicode Character Properties above)
991 [4] \d \D \s \S \w \W \X [:prop:] [:^prop:]
992 [5] can use regular expression look-ahead [a] or
993 user-defined character properties [b] to emulate set
996 [7] note that Perl does Full case-folding in matching (but with
997 bugs), not Simple: for example U+1F88 is equivalent to
998 U+1F00 U+03B9, instead of just U+1F80. This difference
999 matters mainly for certain Greek capital letters with certain
1000 modifiers: the Full case-folding decomposes the letter,
1001 while the Simple case-folding would map it to a single
1003 [8] should do ^ and $ also on U+000B (\v in C), FF (\f), CR
1004 (\r), CRLF (\r\n), NEL (U+0085), LS (U+2028), and PS
1005 (U+2029); should also affect <>, $., and script line
1006 numbers; should not split lines within CRLF [c] (i.e. there
1007 is no empty line between \r and \n)
1008 [9] Linebreaking conformant with UAX#14 "Unicode Line Breaking
1009 Algorithm" is available through the Unicode::LineBreaking
1011 [10] UTF-8/UTF-EBDDIC used in Perl allows not only U+10000 to
1012 U+10FFFF but also beyond U+10FFFF
1014 [a] You can mimic class subtraction using lookahead.
1015 For example, what UTS#18 might write as
1017 [{Greek}-[{UNASSIGNED}]]
1019 in Perl can be written as:
1021 (?!\p{Unassigned})\p{InGreekAndCoptic}
1022 (?=\p{Assigned})\p{InGreekAndCoptic}
1024 But in this particular example, you probably really want
1028 which will match assigned characters known to be part of the Greek script.
1030 Also see the L<Unicode::Regex::Set> module, it does implement the full
1031 UTS#18 grouping, intersection, union, and removal (subtraction) syntax.
1033 [b] '+' for union, '-' for removal (set-difference), '&' for intersection
1034 (see L</"User-Defined Character Properties">)
1036 [c] Try the C<:crlf> layer (see L<PerlIO>).
1040 Level 2 - Extended Unicode Support
1042 RL2.1 Canonical Equivalents - MISSING [10][11]
1043 RL2.2 Default Grapheme Clusters - MISSING [12]
1044 RL2.3 Default Word Boundaries - MISSING [14]
1045 RL2.4 Default Loose Matches - MISSING [15]
1046 RL2.5 Name Properties - DONE
1047 RL2.6 Wildcard Properties - MISSING
1049 [10] see UAX#15 "Unicode Normalization Forms"
1050 [11] have Unicode::Normalize but not integrated to regexes
1051 [12] have \X but we don't have a "Grapheme Cluster Mode"
1052 [14] see UAX#29, Word Boundaries
1053 [15] This is covered in Chapter 3.13 (in Unicode 6.0)
1057 Level 3 - Tailored Support
1059 RL3.1 Tailored Punctuation - MISSING
1060 RL3.2 Tailored Grapheme Clusters - MISSING [17][18]
1061 RL3.3 Tailored Word Boundaries - MISSING
1062 RL3.4 Tailored Loose Matches - MISSING
1063 RL3.5 Tailored Ranges - MISSING
1064 RL3.6 Context Matching - MISSING [19]
1065 RL3.7 Incremental Matches - MISSING
1066 ( RL3.8 Unicode Set Sharing )
1067 RL3.9 Possible Match Sets - MISSING
1068 RL3.10 Folded Matching - MISSING [20]
1069 RL3.11 Submatchers - MISSING
1071 [17] see UAX#10 "Unicode Collation Algorithms"
1072 [18] have Unicode::Collate but not integrated to regexes
1073 [19] have (?<=x) and (?=x), but look-aheads or look-behinds
1074 should see outside of the target substring
1075 [20] need insensitive matching for linguistic features other
1076 than case; for example, hiragana to katakana, wide and
1077 narrow, simplified Han to traditional Han (see UTR#30
1078 "Character Foldings")
1082 =head2 Unicode Encodings
1084 Unicode characters are assigned to I<code points>, which are abstract
1085 numbers. To use these numbers, various encodings are needed.
1093 UTF-8 is a variable-length (1 to 4 bytes), byte-order independent
1094 encoding. For ASCII (and we really do mean 7-bit ASCII, not another
1095 8-bit encoding), UTF-8 is transparent.
1097 The following table is from Unicode 3.2.
1099 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
1101 U+0000..U+007F 00..7F
1102 U+0080..U+07FF * C2..DF 80..BF
1103 U+0800..U+0FFF E0 * A0..BF 80..BF
1104 U+1000..U+CFFF E1..EC 80..BF 80..BF
1105 U+D000..U+D7FF ED 80..9F 80..BF
1106 U+D800..U+DFFF +++++ utf16 surrogates, not legal utf8 +++++
1107 U+E000..U+FFFF EE..EF 80..BF 80..BF
1108 U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF
1109 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
1110 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
1112 Note the gaps marked by "*" before several of the byte entries above. These are
1113 caused by legal UTF-8 avoiding non-shortest encodings: it is technically
1114 possible to UTF-8-encode a single code point in different ways, but that is
1115 explicitly forbidden, and the shortest possible encoding should always be used
1116 (and that is what Perl does).
1118 Another way to look at it is via bits:
1120 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
1123 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
1124 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
1125 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
1127 As you can see, the continuation bytes all begin with "10", and the
1128 leading bits of the start byte tell how many bytes there are in the
1131 The original UTF-8 specification allowed up to 6 bytes, to allow
1132 encoding of numbers up to 0x7FFF_FFFF. Perl continues to allow those,
1133 and has extended that up to 13 bytes to encode code points up to what
1134 can fit in a 64-bit word. However, Perl will warn if you output any of
1135 these as being non-portable; and under strict UTF-8 input protocols,
1138 The Unicode non-character code points are also disallowed in UTF-8 in
1139 "open interchange". See L</Non-character code points>.
1145 Like UTF-8 but EBCDIC-safe, in the way that UTF-8 is ASCII-safe.
1149 UTF-16, UTF-16BE, UTF-16LE, Surrogates, and BOMs (Byte Order Marks)
1151 The followings items are mostly for reference and general Unicode
1152 knowledge, Perl doesn't use these constructs internally.
1154 Like UTF-8, UTF-16 is a variable-width encoding, but where
1155 UTF-8 uses 8-bit code units, UTF-16 uses 16-bit code units.
1156 All code points occupy either 2 or 4 bytes in UTF-16: code points
1157 C<U+0000..U+FFFF> are stored in a single 16-bit unit, and code
1158 points C<U+10000..U+10FFFF> in two 16-bit units. The latter case is
1159 using I<surrogates>, the first 16-bit unit being the I<high
1160 surrogate>, and the second being the I<low surrogate>.
1162 Surrogates are code points set aside to encode the C<U+10000..U+10FFFF>
1163 range of Unicode code points in pairs of 16-bit units. The I<high
1164 surrogates> are the range C<U+D800..U+DBFF> and the I<low surrogates>
1165 are the range C<U+DC00..U+DFFF>. The surrogate encoding is
1167 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
1168 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
1172 $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
1174 Because of the 16-bitness, UTF-16 is byte-order dependent. UTF-16
1175 itself can be used for in-memory computations, but if storage or
1176 transfer is required either UTF-16BE (big-endian) or UTF-16LE
1177 (little-endian) encodings must be chosen.
1179 This introduces another problem: what if you just know that your data
1180 is UTF-16, but you don't know which endianness? Byte Order Marks, or
1181 BOMs, are a solution to this. A special character has been reserved
1182 in Unicode to function as a byte order marker: the character with the
1183 code point C<U+FEFF> is the BOM.
1185 The trick is that if you read a BOM, you will know the byte order,
1186 since if it was written on a big-endian platform, you will read the
1187 bytes C<0xFE 0xFF>, but if it was written on a little-endian platform,
1188 you will read the bytes C<0xFF 0xFE>. (And if the originating platform
1189 was writing in UTF-8, you will read the bytes C<0xEF 0xBB 0xBF>.)
1191 The way this trick works is that the character with the code point
1192 C<U+FFFE> is not supposed to be in input streams, so the
1193 sequence of bytes C<0xFF 0xFE> is unambiguously "BOM, represented in
1194 little-endian format" and cannot be C<U+FFFE>, represented in big-endian
1197 Surrogates have no meaning in Unicode outside their use in pairs to
1198 represent other code points. However, Perl allows them to be
1199 represented individually internally, for example by saying
1200 C<chr(0xD801)>, so that all code points, not just those valid for open
1202 representable. Unicode does define semantics for them, such as their
1203 General Category is "Cs". But because their use is somewhat dangerous,
1204 Perl will warn (using the warning category "surrogate", which is a
1205 sub-category of "utf8") if an attempt is made
1206 to do things like take the lower case of one, or match
1207 case-insensitively, or to output them. (But don't try this on Perls
1212 UTF-32, UTF-32BE, UTF-32LE
1214 The UTF-32 family is pretty much like the UTF-16 family, expect that
1215 the units are 32-bit, and therefore the surrogate scheme is not
1216 needed. UTF-32 is a fixed-width encoding. The BOM signatures are
1217 C<0x00 0x00 0xFE 0xFF> for BE and C<0xFF 0xFE 0x00 0x00> for LE.
1223 Legacy, fixed-width encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
1224 encoding. Unlike UTF-16, UCS-2 is not extensible beyond C<U+FFFF>,
1225 because it does not use surrogates. UCS-4 is a 32-bit encoding,
1226 functionally identical to UTF-32 (the difference being that
1227 UCS-4 forbids neither surrogates nor code points larger than 0x10_FFFF).
1233 A seven-bit safe (non-eight-bit) encoding, which is useful if the
1234 transport or storage is not eight-bit safe. Defined by RFC 2152.
1238 =head2 Non-character code points
1240 66 code points are set aside in Unicode as "non-character code points".
1241 These all have the Unassigned (Cn) General Category, and they never will
1242 be assigned. These are never supposed to be in legal Unicode input
1243 streams, so that code can use them as sentinels that can be mixed in
1244 with character data, and they always will be distinguishable from that data.
1245 To keep them out of Perl input streams, strict UTF-8 should be
1246 specified, such as by using the layer C<:encoding('UTF-8')>. The
1247 non-character code points are the 32 between U+FDD0 and U+FDEF, and the
1248 34 code points U+FFFE, U+FFFF, U+1FFFE, U+1FFFF, ... U+10FFFE, U+10FFFF.
1249 Some people are under the mistaken impression that these are "illegal",
1250 but that is not true. An application or cooperating set of applications
1251 can legally use them at will internally; but these code points are
1252 "illegal for open interchange". Therefore, Perl will not accept these
1253 from input streams unless lax rules are being used, and will warn
1254 (using the warning category "nonchar", which is a sub-category of "utf8") if
1255 an attempt is made to output them.
1257 =head2 Beyond Unicode code points
1259 The maximum Unicode code point is U+10FFFF. But Perl accepts code
1260 points up to the maximum permissible unsigned number available on the
1261 platform. However, Perl will not accept these from input streams unless
1262 lax rules are being used, and will warn (using the warning category
1263 "non_unicode", which is a sub-category of "utf8") if an attempt is made to
1264 operate on or output them. For example, C<uc(0x11_0000)> will generate
1265 this warning, returning the input parameter as its result, as the upper
1266 case of every non-Unicode code point is the code point itself.
1268 =head2 Security Implications of Unicode
1270 Read L<Unicode Security Considerations|http://www.unicode.org/reports/tr36>.
1271 Also, note the following:
1279 Unfortunately, the original specification of UTF-8 leaves some room for
1280 interpretation of how many bytes of encoded output one should generate
1281 from one input Unicode character. Strictly speaking, the shortest
1282 possible sequence of UTF-8 bytes should be generated,
1283 because otherwise there is potential for an input buffer overflow at
1284 the receiving end of a UTF-8 connection. Perl always generates the
1285 shortest length UTF-8, and with warnings on, Perl will warn about
1286 non-shortest length UTF-8 along with other malformations, such as the
1287 surrogates, which are not Unicode code points valid for interchange.
1291 Regular expression pattern matching may surprise you if you're not
1292 accustomed to Unicode. Starting in Perl 5.14, several pattern
1293 modifiers are available to control this, called the character set
1294 modifiers. Details are given in L<perlre/Character set modifiers>.
1298 As discussed elsewhere, Perl has one foot (two hooves?) planted in
1299 each of two worlds: the old world of bytes and the new world of
1300 characters, upgrading from bytes to characters when necessary.
1301 If your legacy code does not explicitly use Unicode, no automatic
1302 switch-over to characters should happen. Characters shouldn't get
1303 downgraded to bytes, either. It is possible to accidentally mix bytes
1304 and characters, however (see L<perluniintro>), in which case C<\w> in
1305 regular expressions might start behaving differently (unless the C</a>
1306 modifier is in effect). Review your code. Use warnings and the C<strict> pragma.
1308 =head2 Unicode in Perl on EBCDIC
1310 The way Unicode is handled on EBCDIC platforms is still
1311 experimental. On such platforms, references to UTF-8 encoding in this
1312 document and elsewhere should be read as meaning the UTF-EBCDIC
1313 specified in Unicode Technical Report 16, unless ASCII vs. EBCDIC issues
1314 are specifically discussed. There is no C<utfebcdic> pragma or
1315 ":utfebcdic" layer; rather, "utf8" and ":utf8" are reused to mean
1316 the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
1317 for more discussion of the issues.
1321 See L<perllocale/Unicode and UTF-8>
1323 =head2 When Unicode Does Not Happen
1325 While Perl does have extensive ways to input and output in Unicode,
1326 and a few other "entry points" like the @ARGV array (which can sometimes be
1327 interpreted as UTF-8), there are still many places where Unicode
1328 (in some encoding or another) could be given as arguments or received as
1329 results, or both, but it is not.
1331 The following are such interfaces. Also, see L</The "Unicode Bug">.
1332 For all of these interfaces Perl
1333 currently (as of 5.8.3) simply assumes byte strings both as arguments
1334 and results, or UTF-8 strings if the (problematic) C<encoding> pragma has been used.
1336 One reason that Perl does not attempt to resolve the role of Unicode in
1337 these situations is that the answers are highly dependent on the operating
1338 system and the file system(s). For example, whether filenames can be
1339 in Unicode and in exactly what kind of encoding, is not exactly a
1340 portable concept. Similarly for C<qx> and C<system>: how well will the
1341 "command-line interface" (and which of them?) handle Unicode?
1347 chdir, chmod, chown, chroot, exec, link, lstat, mkdir,
1348 rename, rmdir, stat, symlink, truncate, unlink, utime, -X
1360 open, opendir, sysopen
1364 qx (aka the backtick operator), system
1372 =head2 The "Unicode Bug"
1374 The term, "Unicode bug" has been applied to an inconsistency
1375 on ASCII platforms with the
1376 Unicode code points in the Latin-1 Supplement block, that
1377 is, between 128 and 255. Without a locale specified, unlike all other
1378 characters or code points, these characters have very different semantics in
1379 byte semantics versus character semantics, unless
1380 C<use feature 'unicode_strings'> is specified, directly or indirectly.
1381 (It is indirectly specified by a C<use v5.12> or higher.)
1383 In character semantics these upper-Latin1 characters are interpreted as
1384 Unicode code points, which means
1385 they have the same semantics as Latin-1 (ISO-8859-1).
1387 In byte semantics (without C<unicode_strings>), they are considered to
1388 be unassigned characters, meaning that the only semantics they have is
1389 their ordinal numbers, and that they are
1390 not members of various character classes. None are considered to match C<\w>
1391 for example, but all match C<\W>.
1393 Perl 5.12.0 added C<unicode_strings> to force character semantics on
1394 these code points in some circumstances, which fixed portions of the
1395 bug; Perl 5.14.0 fixed almost all of it; and Perl 5.16.0 fixed the
1396 remainder (so far as we know, anyway). The lesson here is to enable
1397 C<unicode_strings> to avoid the headaches described below.
1399 The old, problematic behavior affects these areas:
1405 Changing the case of a scalar, that is, using C<uc()>, C<ucfirst()>, C<lc()>,
1406 and C<lcfirst()>, or C<\L>, C<\U>, C<\u> and C<\l> in double-quotish
1407 contexts, such as regular expression substitutions.
1408 Under C<unicode_strings> starting in Perl 5.12.0, character semantics are
1409 generally used. See L<perlfunc/lc> for details on how this works
1410 in combination with various other pragmas.
1414 Using caseless (C</i>) regular expression matching.
1415 Starting in Perl 5.14.0, regular expressions compiled within
1416 the scope of C<unicode_semantics> use character semantics
1417 even when executed or compiled into larger
1418 regular expressions outside the scope.
1422 Matching any of several properties in regular expressions, namely C<\b>,
1423 C<\B>, C<\s>, C<\S>, C<\w>, C<\W>, and all the Posix character classes
1424 I<except> C<[[:ascii:]]>.
1425 Starting in Perl 5.14.0, regular expressions compiled within
1426 the scope of C<unicode_semantics> use character semantics
1427 even when executed or compiled into larger
1428 regular expressions outside the scope.
1432 In C<quotemeta> or its inline equivalent C<\Q>, no code points above 127
1433 are quoted in UTF-8 encoded strings, but in byte encoded strings, code
1434 points between 128-255 are always quoted.
1435 Starting in Perl 5.16.0, consistent quoting rules are used within the
1436 scope of C<unicode_strings>, as described in L<perlfunc/quotemeta>.
1440 This behavior can lead to unexpected results in which a string's semantics
1441 suddenly change if a code point above 255 is appended to or removed from it,
1442 which changes the string's semantics from byte to character or vice versa. As
1443 an example, consider the following program and its output:
1446 no feature 'unicode_strings';
1449 for ($s1, $s2, $s1.$s2) {
1457 If there's no C<\w> in C<s1> or in C<s2>, why does their concatenation have one?
1459 This anomaly stems from Perl's attempt to not disturb older programs that
1460 didn't use Unicode, and hence had no semantics for characters outside of the
1461 ASCII range (except in a locale), along with Perl's desire to add Unicode
1462 support seamlessly. The result wasn't seamless: these characters were
1465 For Perls earlier than those described above, or when a string is passed
1466 to a function outside the subpragma's scope, a workaround is to always
1467 call C<utf8::upgrade($string)>,
1468 or to use the standard module L<Encode>. Also, a scalar that has any characters
1469 whose ordinal is above 0x100, or which were specified using either of the
1470 C<\N{...}> notations, will automatically have character semantics.
1472 =head2 Forcing Unicode in Perl (Or Unforcing Unicode in Perl)
1474 Sometimes (see L</"When Unicode Does Not Happen"> or L</The "Unicode Bug">)
1475 there are situations where you simply need to force a byte
1476 string into UTF-8, or vice versa. The low-level calls
1477 utf8::upgrade($bytestring) and utf8::downgrade($utf8string[, FAIL_OK]) are
1480 Note that utf8::downgrade() can fail if the string contains characters
1481 that don't fit into a byte.
1483 Calling either function on a string that already is in the desired state is a
1486 =head2 Using Unicode in XS
1488 If you want to handle Perl Unicode in XS extensions, you may find the
1489 following C APIs useful. See also L<perlguts/"Unicode Support"> for an
1490 explanation about Unicode at the XS level, and L<perlapi> for the API
1497 C<DO_UTF8(sv)> returns true if the C<UTF8> flag is on and the bytes
1498 pragma is not in effect. C<SvUTF8(sv)> returns true if the C<UTF8>
1499 flag is on; the bytes pragma is ignored. The C<UTF8> flag being on
1500 does B<not> mean that there are any characters of code points greater
1501 than 255 (or 127) in the scalar or that there are even any characters
1502 in the scalar. What the C<UTF8> flag means is that the sequence of
1503 octets in the representation of the scalar is the sequence of UTF-8
1504 encoded code points of the characters of a string. The C<UTF8> flag
1505 being off means that each octet in this representation encodes a
1506 single character with code point 0..255 within the string. Perl's
1507 Unicode model is not to use UTF-8 until it is absolutely necessary.
1511 C<uvchr_to_utf8(buf, chr)> writes a Unicode character code point into
1512 a buffer encoding the code point as UTF-8, and returns a pointer
1513 pointing after the UTF-8 bytes. It works appropriately on EBCDIC machines.
1517 C<utf8_to_uvchr_buf(buf, bufend, lenp)> reads UTF-8 encoded bytes from a
1519 returns the Unicode character code point and, optionally, the length of
1520 the UTF-8 byte sequence. It works appropriately on EBCDIC machines.
1524 C<utf8_length(start, end)> returns the length of the UTF-8 encoded buffer
1525 in characters. C<sv_len_utf8(sv)> returns the length of the UTF-8 encoded
1530 C<sv_utf8_upgrade(sv)> converts the string of the scalar to its UTF-8
1531 encoded form. C<sv_utf8_downgrade(sv)> does the opposite, if
1532 possible. C<sv_utf8_encode(sv)> is like sv_utf8_upgrade except that
1533 it does not set the C<UTF8> flag. C<sv_utf8_decode()> does the
1534 opposite of C<sv_utf8_encode()>. Note that none of these are to be
1535 used as general-purpose encoding or decoding interfaces: C<use Encode>
1536 for that. C<sv_utf8_upgrade()> is affected by the encoding pragma
1537 but C<sv_utf8_downgrade()> is not (since the encoding pragma is
1538 designed to be a one-way street).
1542 C<is_utf8_string(buf, len)> returns true if C<len> bytes of the buffer
1547 C<is_utf8_char(s)> returns true if the pointer points to a valid UTF-8
1548 character. However, this function should not be used because of
1549 security concerns. Instead, use C<is_utf8_string()>.
1553 C<UTF8SKIP(buf)> will return the number of bytes in the UTF-8 encoded
1554 character in the buffer. C<UNISKIP(chr)> will return the number of bytes
1555 required to UTF-8-encode the Unicode character code point. C<UTF8SKIP()>
1556 is useful for example for iterating over the characters of a UTF-8
1557 encoded buffer; C<UNISKIP()> is useful, for example, in computing
1558 the size required for a UTF-8 encoded buffer.
1562 C<utf8_distance(a, b)> will tell the distance in characters between the
1563 two pointers pointing to the same UTF-8 encoded buffer.
1567 C<utf8_hop(s, off)> will return a pointer to a UTF-8 encoded buffer
1568 that is C<off> (positive or negative) Unicode characters displaced
1569 from the UTF-8 buffer C<s>. Be careful not to overstep the buffer:
1570 C<utf8_hop()> will merrily run off the end or the beginning of the
1571 buffer if told to do so.
1575 C<pv_uni_display(dsv, spv, len, pvlim, flags)> and
1576 C<sv_uni_display(dsv, ssv, pvlim, flags)> are useful for debugging the
1577 output of Unicode strings and scalars. By default they are useful
1578 only for debugging--they display B<all> characters as hexadecimal code
1579 points--but with the flags C<UNI_DISPLAY_ISPRINT>,
1580 C<UNI_DISPLAY_BACKSLASH>, and C<UNI_DISPLAY_QQ> you can make the
1581 output more readable.
1585 C<foldEQ_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2)> can be used to
1586 compare two strings case-insensitively in Unicode. For case-sensitive
1587 comparisons you can just use C<memEQ()> and C<memNE()> as usual, except
1588 if one string is in utf8 and the other isn't.
1592 For more information, see L<perlapi>, and F<utf8.c> and F<utf8.h>
1593 in the Perl source code distribution.
1595 =head2 Hacking Perl to work on earlier Unicode versions (for very serious hackers only)
1597 Perl by default comes with the latest supported Unicode version built in, but
1598 you can change to use any earlier one.
1600 Download the files in the desired version of Unicode from the Unicode web
1601 site L<http://www.unicode.org>). These should replace the existing files in
1602 F<lib/unicore> in the Perl source tree. Follow the instructions in
1603 F<README.perl> in that directory to change some of their names, and then build
1604 perl (see L<INSTALL>).
1608 =head2 Interaction with Locales
1610 See L<perllocale/Unicode and UTF-8>
1612 =head2 Problems with characters in the Latin-1 Supplement range
1614 See L</The "Unicode Bug">
1616 =head2 Interaction with Extensions
1618 When Perl exchanges data with an extension, the extension should be
1619 able to understand the UTF8 flag and act accordingly. If the
1620 extension doesn't recognize that flag, it's likely that the extension
1621 will return incorrectly-flagged data.
1623 So if you're working with Unicode data, consult the documentation of
1624 every module you're using if there are any issues with Unicode data
1625 exchange. If the documentation does not talk about Unicode at all,
1626 suspect the worst and probably look at the source to learn how the
1627 module is implemented. Modules written completely in Perl shouldn't
1628 cause problems. Modules that directly or indirectly access code written
1629 in other programming languages are at risk.
1631 For affected functions, the simple strategy to avoid data corruption is
1632 to always make the encoding of the exchanged data explicit. Choose an
1633 encoding that you know the extension can handle. Convert arguments passed
1634 to the extensions to that encoding and convert results back from that
1635 encoding. Write wrapper functions that do the conversions for you, so
1636 you can later change the functions when the extension catches up.
1638 To provide an example, let's say the popular Foo::Bar::escape_html
1639 function doesn't deal with Unicode data yet. The wrapper function
1640 would convert the argument to raw UTF-8 and convert the result back to
1641 Perl's internal representation like so:
1643 sub my_escape_html ($) {
1645 return unless defined $what;
1646 Encode::decode_utf8(Foo::Bar::escape_html(
1647 Encode::encode_utf8($what)));
1650 Sometimes, when the extension does not convert data but just stores
1651 and retrieves them, you will be able to use the otherwise
1652 dangerous Encode::_utf8_on() function. Let's say the popular
1653 C<Foo::Bar> extension, written in C, provides a C<param> method that
1654 lets you store and retrieve data according to these prototypes:
1656 $self->param($name, $value); # set a scalar
1657 $value = $self->param($name); # retrieve a scalar
1659 If it does not yet provide support for any encoding, one could write a
1660 derived class with such a C<param> method:
1663 my($self,$name,$value) = @_;
1664 utf8::upgrade($name); # make sure it is UTF-8 encoded
1665 if (defined $value) {
1666 utf8::upgrade($value); # make sure it is UTF-8 encoded
1667 return $self->SUPER::param($name,$value);
1669 my $ret = $self->SUPER::param($name);
1670 Encode::_utf8_on($ret); # we know, it is UTF-8 encoded
1675 Some extensions provide filters on data entry/exit points, such as
1676 DB_File::filter_store_key and family. Look out for such filters in
1677 the documentation of your extensions, they can make the transition to
1678 Unicode data much easier.
1682 Some functions are slower when working on UTF-8 encoded strings than
1683 on byte encoded strings. All functions that need to hop over
1684 characters such as length(), substr() or index(), or matching regular
1685 expressions can work B<much> faster when the underlying data are
1688 In Perl 5.8.0 the slowness was often quite spectacular; in Perl 5.8.1
1689 a caching scheme was introduced which will hopefully make the slowness
1690 somewhat less spectacular, at least for some operations. In general,
1691 operations with UTF-8 encoded strings are still slower. As an example,
1692 the Unicode properties (character classes) like C<\p{Nd}> are known to
1693 be quite a bit slower (5-20 times) than their simpler counterparts
1694 like C<\d> (then again, there are hundreds of Unicode characters matching C<Nd>
1695 compared with the 10 ASCII characters matching C<d>).
1697 =head2 Problems on EBCDIC platforms
1699 There are several known problems with Perl on EBCDIC platforms. If you
1700 want to use Perl there, send email to perlbug@perl.org.
1702 In earlier versions, when byte and character data were concatenated,
1703 the new string was sometimes created by
1704 decoding the byte strings as I<ISO 8859-1 (Latin-1)>, even if the
1705 old Unicode string used EBCDIC.
1707 If you find any of these, please report them as bugs.
1709 =head2 Porting code from perl-5.6.X
1711 Perl 5.8 has a different Unicode model from 5.6. In 5.6 the programmer
1712 was required to use the C<utf8> pragma to declare that a given scope
1713 expected to deal with Unicode data and had to make sure that only
1714 Unicode data were reaching that scope. If you have code that is
1715 working with 5.6, you will need some of the following adjustments to
1716 your code. The examples are written such that the code will continue
1717 to work under 5.6, so you should be safe to try them out.
1723 A filehandle that should read or write UTF-8
1726 binmode $fh, ":encoding(utf8)";
1731 A scalar that is going to be passed to some extension
1733 Be it Compress::Zlib, Apache::Request or any extension that has no
1734 mention of Unicode in the manpage, you need to make sure that the
1735 UTF8 flag is stripped off. Note that at the time of this writing
1736 (October 2002) the mentioned modules are not UTF-8-aware. Please
1737 check the documentation to verify if this is still true.
1741 $val = Encode::encode_utf8($val); # make octets
1746 A scalar we got back from an extension
1748 If you believe the scalar comes back as UTF-8, you will most likely
1749 want the UTF8 flag restored:
1753 $val = Encode::decode_utf8($val);
1758 Same thing, if you are really sure it is UTF-8
1762 Encode::_utf8_on($val);
1767 A wrapper for fetchrow_array and fetchrow_hashref
1769 When the database contains only UTF-8, a wrapper function or method is
1770 a convenient way to replace all your fetchrow_array and
1771 fetchrow_hashref calls. A wrapper function will also make it easier to
1772 adapt to future enhancements in your database driver. Note that at the
1773 time of this writing (October 2002), the DBI has no standardized way
1774 to deal with UTF-8 data. Please check the documentation to verify if
1778 # $what is one of fetchrow_{array,hashref}
1779 my($self, $sth, $what) = @_;
1785 my @arr = $sth->$what;
1787 defined && /[^\000-\177]/ && Encode::_utf8_on($_);
1791 my $ret = $sth->$what;
1793 for my $k (keys %$ret) {
1796 && Encode::_utf8_on($_) for $ret->{$k};
1800 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
1810 A large scalar that you know can only contain ASCII
1812 Scalars that contain only ASCII and are marked as UTF-8 are sometimes
1813 a drag to your program. If you recognize such a situation, just remove
1816 utf8::downgrade($val) if $] > 5.007;
1822 L<perlunitut>, L<perluniintro>, L<perluniprops>, L<Encode>, L<open>, L<utf8>, L<bytes>,
1823 L<perlretut>, L<perlvar/"${^UNICODE}">
1824 L<http://www.unicode.org/reports/tr44>).