This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlunicode.pod: Elaborate unicode bug for POSIX
[perl5.git] / pod / perlunicode.pod
CommitLineData
393fec97
GS
1=head1 NAME
2
3perlunicode - Unicode support in Perl
4
5=head1 DESCRIPTION
6
0a1f2d14 7=head2 Important Caveats
21bad921 8
376d9008 9Unicode support is an extensive requirement. While Perl does not
c349b1b9
JH
10implement the Unicode standard or the accompanying technical reports
11from cover to cover, Perl does support many Unicode features.
21bad921 12
2575c402 13People who want to learn to use Unicode in Perl, should probably read
9d1c51c1 14the L<Perl Unicode tutorial, perlunitut|perlunitut>, before reading
e4911a48 15this reference document.
2575c402 16
9d1c51c1
KW
17Also, the use of Unicode may present security issues that aren't obvious.
18Read L<Unicode Security Considerations|http://www.unicode.org/reports/tr36>.
19
13a2d996 20=over 4
21bad921 21
fae2c0fb 22=item Input and Output Layers
21bad921 23
376d9008 24Perl knows when a filehandle uses Perl's internal Unicode encodings
1bfb14c4
JH
25(UTF-8, or UTF-EBCDIC if in EBCDIC) if the filehandle is opened with
26the ":utf8" layer. Other encodings can be converted to Perl's
27encoding on input or from Perl's encoding on output by use of the
28":encoding(...)" layer. See L<open>.
c349b1b9 29
2575c402 30To indicate that Perl source itself is in UTF-8, use C<use utf8;>.
21bad921
GS
31
32=item Regular Expressions
33
c349b1b9 34The regular expression compiler produces polymorphic opcodes. That is,
376d9008 35the pattern adapts to the data and automatically switches to the Unicode
2575c402 36character scheme when presented with data that is internally encoded in
ac036724 37UTF-8, or instead uses a traditional byte scheme when presented with
2575c402 38byte data.
21bad921 39
ad0029c4 40=item C<use utf8> still needed to enable UTF-8/UTF-EBCDIC in scripts
21bad921 41
376d9008
JB
42As a compatibility measure, the C<use utf8> pragma must be explicitly
43included to enable recognition of UTF-8 in the Perl scripts themselves
1bfb14c4
JH
44(in string or regular expression literals, or in identifier names) on
45ASCII-based machines or to recognize UTF-EBCDIC on EBCDIC-based
376d9008 46machines. B<These are the only times when an explicit C<use utf8>
8f8cf39c 47is needed.> See L<utf8>.
21bad921 48
7aa207d6
JH
49=item BOM-marked scripts and UTF-16 scripts autodetected
50
51If a Perl script begins marked with the Unicode BOM (UTF-16LE, UTF16-BE,
52or UTF-8), or if the script looks like non-BOM-marked UTF-16 of either
53endianness, Perl will correctly read in the script as Unicode.
54(BOMless UTF-8 cannot be effectively recognized or differentiated from
55ISO 8859-1 or other eight-bit encodings.)
56
990e18f7
AT
57=item C<use encoding> needed to upgrade non-Latin-1 byte strings
58
38a44b82 59By default, there is a fundamental asymmetry in Perl's Unicode model:
990e18f7
AT
60implicit upgrading from byte strings to Unicode strings assumes that
61they were encoded in I<ISO 8859-1 (Latin-1)>, but Unicode strings are
62downgraded with UTF-8 encoding. This happens because the first 256
51f494cc 63codepoints in Unicode happens to agree with Latin-1.
990e18f7 64
990e18f7
AT
65See L</"Byte and Character Semantics"> for more details.
66
21bad921
GS
67=back
68
376d9008 69=head2 Byte and Character Semantics
393fec97 70
376d9008 71Beginning with version 5.6, Perl uses logically-wide characters to
3e4dbfed 72represent strings internally.
393fec97 73
376d9008
JB
74In future, Perl-level operations will be expected to work with
75characters rather than bytes.
393fec97 76
376d9008 77However, as an interim compatibility measure, Perl aims to
75daf61c
JH
78provide a safe migration path from byte semantics to character
79semantics for programs. For operations where Perl can unambiguously
376d9008 80decide that the input data are characters, Perl switches to
75daf61c
JH
81character semantics. For operations where this determination cannot
82be made without additional information from the user, Perl decides in
376d9008 83favor of compatibility and chooses to use byte semantics.
8cbd9a7a 84
51f494cc 85Under byte semantics, when C<use locale> is in effect, Perl uses the
e1b711da
KW
86semantics associated with the current locale. Absent a C<use locale>, and
87absent a C<use feature 'unicode_strings'> pragma, Perl currently uses US-ASCII
88(or Basic Latin in Unicode terminology) byte semantics, meaning that characters
89whose ordinal numbers are in the range 128 - 255 are undefined except for their
90ordinal numbers. This means that none have case (upper and lower), nor are any
91a member of character classes, like C<[:alpha:]> or C<\w>. (But all do belong
92to the C<\W> class or the Perl regular expression extension C<[:^alpha:]>.)
2bbc8d55 93
8cbd9a7a 94This behavior preserves compatibility with earlier versions of Perl,
376d9008 95which allowed byte semantics in Perl operations only if
e1b711da 96none of the program's inputs were marked as being a source of Unicode
8cbd9a7a
GS
97character data. Such data may come from filehandles, from calls to
98external programs, from information provided by the system (such as %ENV),
21bad921 99or from literals and constants in the source text.
8cbd9a7a 100
376d9008
JB
101The C<bytes> pragma will always, regardless of platform, force byte
102semantics in a particular lexical scope. See L<bytes>.
8cbd9a7a 103
e1b711da 104The C<use feature 'unicode_strings'> pragma is intended to always, regardless
9d1c51c1
KW
105of platform, force character (Unicode) semantics in a particular lexical scope.
106In release 5.12, it is partially implemented, applying only to case changes.
e1b711da
KW
107See L</The "Unicode Bug"> below.
108
8cbd9a7a 109The C<utf8> pragma is primarily a compatibility device that enables
75daf61c 110recognition of UTF-(8|EBCDIC) in literals encountered by the parser.
376d9008
JB
111Note that this pragma is only required while Perl defaults to byte
112semantics; when character semantics become the default, this pragma
113may become a no-op. See L<utf8>.
114
115Unless explicitly stated, Perl operators use character semantics
116for Unicode data and byte semantics for non-Unicode data.
117The decision to use character semantics is made transparently. If
118input data comes from a Unicode source--for example, if a character
fae2c0fb 119encoding layer is added to a filehandle or a literal Unicode
376d9008
JB
120string constant appears in a program--character semantics apply.
121Otherwise, byte semantics are in effect. The C<bytes> pragma should
e1b711da
KW
122be used to force byte semantics on Unicode data, and the C<use feature
123'unicode_strings'> pragma to force Unicode semantics on byte data (though in
1245.12 it isn't fully implemented).
376d9008
JB
125
126If strings operating under byte semantics and strings with Unicode
51f494cc 127character data are concatenated, the new string will have
d9b01026
KW
128character semantics. This can cause surprises: See L</BUGS>, below.
129You can choose to be warned when this happens. See L<encoding::warnings>.
7dedd01f 130
feda178f 131Under character semantics, many operations that formerly operated on
376d9008 132bytes now operate on characters. A character in Perl is
feda178f 133logically just a number ranging from 0 to 2**31 or so. Larger
376d9008
JB
134characters may encode into longer sequences of bytes internally, but
135this internal detail is mostly hidden for Perl code.
136See L<perluniintro> for more.
393fec97 137
376d9008 138=head2 Effects of Character Semantics
393fec97
GS
139
140Character semantics have the following effects:
141
142=over 4
143
144=item *
145
376d9008 146Strings--including hash keys--and regular expression patterns may
574c8022 147contain characters that have an ordinal value larger than 255.
393fec97 148
2575c402
JW
149If you use a Unicode editor to edit your program, Unicode characters may
150occur directly within the literal strings in UTF-8 encoding, or UTF-16.
151(The former requires a BOM or C<use utf8>, the latter requires a BOM.)
3e4dbfed 152
195e542a
KW
153Unicode characters can also be added to a string by using the C<\N{U+...}>
154notation. The Unicode code for the desired character, in hexadecimal,
155should be placed in the braces, after the C<U>. For instance, a smiley face is
6f335b04
KW
156C<\N{U+263A}>.
157
195e542a
KW
158Alternatively, you can use the C<\x{...}> notation for characters 0x100 and
159above. For characters below 0x100 you may get byte semantics instead of
6f335b04 160character semantics; see L</The "Unicode Bug">. On EBCDIC machines there is
195e542a 161the additional problem that the value for such characters gives the EBCDIC
6f335b04 162character rather than the Unicode one.
3e4dbfed
JF
163
164Additionally, if you
574c8022 165
3e4dbfed 166 use charnames ':full';
574c8022 167
1bfb14c4
JH
168you can use the C<\N{...}> notation and put the official Unicode
169character name within the braces, such as C<\N{WHITE SMILING FACE}>.
6f335b04 170See L<charnames>.
376d9008 171
393fec97
GS
172=item *
173
574c8022
JH
174If an appropriate L<encoding> is specified, identifiers within the
175Perl script may contain Unicode alphanumeric characters, including
376d9008
JB
176ideographs. Perl does not currently attempt to canonicalize variable
177names.
393fec97 178
393fec97
GS
179=item *
180
1bfb14c4 181Regular expressions match characters instead of bytes. "." matches
2575c402 182a character instead of a byte.
393fec97 183
393fec97
GS
184=item *
185
9d1c51c1 186Bracketed character classes in regular expressions match characters instead of
376d9008 187bytes and match against the character properties specified in the
1bfb14c4 188Unicode properties database. C<\w> can be used to match a Japanese
75daf61c 189ideograph, for instance.
393fec97 190
393fec97
GS
191=item *
192
9d1c51c1
KW
193Named Unicode properties, scripts, and block ranges may be used (like bracketed
194character classes) by using the C<\p{}> "matches property" construct and
822502e5 195the C<\P{}> negation, "doesn't match property".
2575c402 196See L</"Unicode Character Properties"> for more details.
822502e5
ST
197
198You can define your own character properties and use them
199in the regular expression with the C<\p{}> or C<\P{}> construct.
822502e5
ST
200See L</"User-Defined Character Properties"> for more details.
201
202=item *
203
9f815e24
KW
204The special pattern C<\X> matches a logical character, an "extended grapheme
205cluster" in Standardese. In Unicode what appears to the user to be a single
51f494cc
KW
206character, for example an accented C<G>, may in fact be composed of a sequence
207of characters, in this case a C<G> followed by an accent character. C<\X>
208will match the entire sequence.
822502e5
ST
209
210=item *
211
212The C<tr///> operator translates characters instead of bytes. Note
213that the C<tr///CU> functionality has been removed. For similar
214functionality see pack('U0', ...) and pack('C0', ...).
215
216=item *
217
218Case translation operators use the Unicode case translation tables
219when character input is provided. Note that C<uc()>, or C<\U> in
220interpolated strings, translates to uppercase, while C<ucfirst>,
221or C<\u> in interpolated strings, translates to titlecase in languages
e1b711da
KW
222that make the distinction (which is equivalent to uppercase in languages
223without the distinction).
822502e5
ST
224
225=item *
226
227Most operators that deal with positions or lengths in a string will
228automatically switch to using character positions, including
229C<chop()>, C<chomp()>, C<substr()>, C<pos()>, C<index()>, C<rindex()>,
230C<sprintf()>, C<write()>, and C<length()>. An operator that
51f494cc
KW
231specifically does not switch is C<vec()>. Operators that really don't
232care include operators that treat strings as a bucket of bits such as
822502e5
ST
233C<sort()>, and operators dealing with filenames.
234
235=item *
236
51f494cc 237The C<pack()>/C<unpack()> letter C<C> does I<not> change, since it is often
822502e5
ST
238used for byte-oriented formats. Again, think C<char> in the C language.
239
240There is a new C<U> specifier that converts between Unicode characters
241and code points. There is also a C<W> specifier that is the equivalent of
242C<chr>/C<ord> and properly handles character values even if they are above 255.
243
244=item *
245
246The C<chr()> and C<ord()> functions work on characters, similar to
247C<pack("W")> and C<unpack("W")>, I<not> C<pack("C")> and
248C<unpack("C")>. C<pack("C")> and C<unpack("C")> are methods for
249emulating byte-oriented C<chr()> and C<ord()> on Unicode strings.
250While these methods reveal the internal encoding of Unicode strings,
251that is not something one normally needs to care about at all.
252
253=item *
254
255The bit string operators, C<& | ^ ~>, can operate on character data.
256However, for backward compatibility, such as when using bit string
257operations when characters are all less than 256 in ordinal value, one
258should not use C<~> (the bit complement) with characters of both
259values less than 256 and values greater than 256. Most importantly,
260DeMorgan's laws (C<~($x|$y) eq ~$x&~$y> and C<~($x&$y) eq ~$x|~$y>)
261will not hold. The reason for this mathematical I<faux pas> is that
262the complement cannot return B<both> the 8-bit (byte-wide) bit
263complement B<and> the full character-wide bit complement.
264
265=item *
266
9d1c51c1
KW
267You can define your own mappings to be used in C<lc()>,
268C<lcfirst()>, C<uc()>, and C<ucfirst()> (or their double-quoted string inlined
269versions such as C<\U>).
822502e5
ST
270See L</"User-Defined Case Mappings"> for more details.
271
272=back
273
274=over 4
275
276=item *
277
278And finally, C<scalar reverse()> reverses by character rather than by byte.
279
280=back
281
282=head2 Unicode Character Properties
283
51f494cc 284Most Unicode character properties are accessible by using regular expressions.
9d1c51c1
KW
285They are used (like bracketed character classes) by using the C<\p{}> "matches
286property" construct and the C<\P{}> negation, "doesn't match property".
287
288Note that the only time that Perl considers a sequence of individual code
289points as a single logical character is in the C<\X> construct, already
290mentioned above. Therefore "character" in this discussion means a single
291Unicode code point.
51f494cc 292
9d1c51c1 293For instance, C<\p{Uppercase}> matches any single character with the Unicode
51f494cc
KW
294"Uppercase" property, while C<\p{L}> matches any character with a
295General_Category of "L" (letter) property. Brackets are not
9d1c51c1 296required for single letter property names, so C<\p{L}> is equivalent to C<\pL>.
51f494cc 297
9d1c51c1
KW
298More formally, C<\p{Uppercase}> matches any single character whose Unicode
299Uppercase property value is True, and C<\P{Uppercase}> matches any character
300whose Uppercase property value is False, and they could have been written as
301C<\p{Uppercase=True}> and C<\p{Uppercase=False}>, respectively.
51f494cc
KW
302
303This formality is needed when properties are not binary, that is if they can
304take on more values than just True and False. For example, the Bidi_Class (see
305L</"Bidirectional Character Types"> below), can take on a number of different
306values, such as Left, Right, Whitespace, and others. To match these, one needs
e1b711da 307to specify the property name (Bidi_Class), and the value being matched against
9d1c51c1 308(Left, Right, etc.). This is done, as in the examples above, by having the
9f815e24 309two components separated by an equal sign (or interchangeably, a colon), like
51f494cc
KW
310C<\p{Bidi_Class: Left}>.
311
312All Unicode-defined character properties may be written in these compound forms
313of C<\p{property=value}> or C<\p{property:value}>, but Perl provides some
314additional properties that are written only in the single form, as well as
315single-form short-cuts for all binary properties and certain others described
316below, in which you may omit the property name and the equals or colon
317separator.
318
319Most Unicode character properties have at least two synonyms (or aliases if you
320prefer), a short one that is easier to type, and a longer one which is more
321descriptive and hence it is easier to understand what it means. Thus the "L"
322and "Letter" above are equivalent and can be used interchangeably. Likewise,
323"Upper" is a synonym for "Uppercase", and we could have written
324C<\p{Uppercase}> equivalently as C<\p{Upper}>. Also, there are typically
325various synonyms for the values the property can be. For binary properties,
326"True" has 3 synonyms: "T", "Yes", and "Y"; and "False has correspondingly "F",
327"No", and "N". But be careful. A short form of a value for one property may
e1b711da 328not mean the same thing as the same short form for another. Thus, for the
51f494cc
KW
329General_Category property, "L" means "Letter", but for the Bidi_Class property,
330"L" means "Left". A complete list of properties and synonyms is in
331L<perluniprops>.
332
333Upper/lower case differences in the property names and values are irrelevant,
334thus C<\p{Upper}> means the same thing as C<\p{upper}> or even C<\p{UpPeR}>.
335Similarly, you can add or subtract underscores anywhere in the middle of a
336word, so that these are also equivalent to C<\p{U_p_p_e_r}>. And white space
337is irrelevant adjacent to non-word characters, such as the braces and the equals
338or colon separators so C<\p{ Upper }> and C<\p{ Upper_case : Y }> are
339equivalent to these as well. In fact, in most cases, white space and even
340hyphens can be added or deleted anywhere. So even C<\p{ Up-per case = Yes}> is
341equivalent. All this is called "loose-matching" by Unicode. The few places
342where stricter matching is employed is in the middle of numbers, and the Perl
343extension properties that begin or end with an underscore. Stricter matching
344cares about white space (except adjacent to the non-word characters) and
345hyphens, and non-interior underscores.
4193bef7 346
376d9008
JB
347You can also use negation in both C<\p{}> and C<\P{}> by introducing a caret
348(^) between the first brace and the property name: C<\p{^Tamil}> is
eb0cc9e3 349equal to C<\P{Tamil}>.
4193bef7 350
51f494cc 351=head3 B<General_Category>
14bb0a9a 352
51f494cc
KW
353Every Unicode character is assigned a general category, which is the "most
354usual categorization of a character" (from
355L<http://www.unicode.org/reports/tr44>).
822502e5 356
9f815e24 357The compound way of writing these is like C<\p{General_Category=Number}>
51f494cc
KW
358(short, C<\p{gc:n}>). But Perl furnishes shortcuts in which everything up
359through the equal or colon separator is omitted. So you can instead just write
360C<\pN>.
822502e5 361
51f494cc 362Here are the short and long forms of the General Category properties:
393fec97 363
d73e5302
JH
364 Short Long
365
366 L Letter
51f494cc
KW
367 LC, L& Cased_Letter (that is: [\p{Ll}\p{Lu}\p{Lt}])
368 Lu Uppercase_Letter
369 Ll Lowercase_Letter
370 Lt Titlecase_Letter
371 Lm Modifier_Letter
372 Lo Other_Letter
d73e5302
JH
373
374 M Mark
51f494cc
KW
375 Mn Nonspacing_Mark
376 Mc Spacing_Mark
377 Me Enclosing_Mark
d73e5302
JH
378
379 N Number
51f494cc
KW
380 Nd Decimal_Number (also Digit)
381 Nl Letter_Number
382 No Other_Number
383
384 P Punctuation (also Punct)
385 Pc Connector_Punctuation
386 Pd Dash_Punctuation
387 Ps Open_Punctuation
388 Pe Close_Punctuation
389 Pi Initial_Punctuation
d73e5302 390 (may behave like Ps or Pe depending on usage)
51f494cc 391 Pf Final_Punctuation
d73e5302 392 (may behave like Ps or Pe depending on usage)
51f494cc 393 Po Other_Punctuation
d73e5302
JH
394
395 S Symbol
51f494cc
KW
396 Sm Math_Symbol
397 Sc Currency_Symbol
398 Sk Modifier_Symbol
399 So Other_Symbol
d73e5302
JH
400
401 Z Separator
51f494cc
KW
402 Zs Space_Separator
403 Zl Line_Separator
404 Zp Paragraph_Separator
d73e5302
JH
405
406 C Other
d88362ca 407 Cc Control (also Cntrl)
e150c829 408 Cf Format
eb0cc9e3 409 Cs Surrogate (not usable)
51f494cc 410 Co Private_Use
e150c829 411 Cn Unassigned
1ac13f9a 412
376d9008 413Single-letter properties match all characters in any of the
3e4dbfed 414two-letter sub-properties starting with the same letter.
9d1c51c1 415C<LC> and C<L&> are special cases, which are both aliases for the set consisting of everything matched by C<Ll>, C<Lu>, and C<Lt>.
32293815 416
eb0cc9e3 417Because Perl hides the need for the user to understand the internal
1bfb14c4
JH
418representation of Unicode characters, there is no need to implement
419the somewhat messy concept of surrogates. C<Cs> is therefore not
eb0cc9e3 420supported.
d73e5302 421
51f494cc 422=head3 B<Bidirectional Character Types>
822502e5 423
9d1c51c1
KW
424Because scripts differ in their directionality (Hebrew is
425written right to left, for example) Unicode supplies these properties in
51f494cc 426the Bidi_Class class:
32293815 427
eb0cc9e3 428 Property Meaning
92e830a9 429
12ac2576
JP
430 L Left-to-Right
431 LRE Left-to-Right Embedding
432 LRO Left-to-Right Override
433 R Right-to-Left
51f494cc 434 AL Arabic Letter
12ac2576
JP
435 RLE Right-to-Left Embedding
436 RLO Right-to-Left Override
437 PDF Pop Directional Format
438 EN European Number
51f494cc
KW
439 ES European Separator
440 ET European Terminator
12ac2576 441 AN Arabic Number
51f494cc 442 CS Common Separator
12ac2576
JP
443 NSM Non-Spacing Mark
444 BN Boundary Neutral
445 B Paragraph Separator
446 S Segment Separator
447 WS Whitespace
448 ON Other Neutrals
449
51f494cc
KW
450This property is always written in the compound form.
451For example, C<\p{Bidi_Class:R}> matches characters that are normally
eb0cc9e3
JH
452written right to left.
453
51f494cc
KW
454=head3 B<Scripts>
455
e1b711da
KW
456The world's languages are written in a number of scripts. This sentence
457(unless you're reading it in translation) is written in Latin, while Russian is
458written in Cyrllic, and Greek is written in, well, Greek; Japanese mainly in
459Hiragana or Katakana. There are many more.
51f494cc
KW
460
461The Unicode Script property gives what script a given character is in,
9d1c51c1
KW
462and the property can be specified with the compound form like
463C<\p{Script=Hebrew}> (short: C<\p{sc=hebr}>). Perl furnishes shortcuts for all
464script names. You can omit everything up through the equals (or colon), and
465simply write C<\p{Latin}> or C<\P{Cyrillic}>.
51f494cc
KW
466
467A complete list of scripts and their shortcuts is in L<perluniprops>.
468
51f494cc 469=head3 B<Use of "Is" Prefix>
822502e5 470
1bfb14c4 471For backward compatibility (with Perl 5.6), all properties mentioned
51f494cc
KW
472so far may have C<Is> or C<Is_> prepended to their name, so C<\P{Is_Lu}>, for
473example, is equal to C<\P{Lu}>, and C<\p{IsScript:Arabic}> is equal to
474C<\p{Arabic}>.
eb0cc9e3 475
51f494cc 476=head3 B<Blocks>
2796c109 477
1bfb14c4
JH
478In addition to B<scripts>, Unicode also defines B<blocks> of
479characters. The difference between scripts and blocks is that the
480concept of scripts is closer to natural languages, while the concept
51f494cc 481of blocks is more of an artificial grouping based on groups of Unicode
9f815e24 482characters with consecutive ordinal values. For example, the "Basic Latin"
51f494cc 483block is all characters whose ordinals are between 0 and 127, inclusive, in
9f815e24
KW
484other words, the ASCII characters. The "Latin" script contains some letters
485from this block as well as several more, like "Latin-1 Supplement",
9d1c51c1 486"Latin Extended-A", etc., but it does not contain all the characters from
51f494cc
KW
487those blocks. It does not, for example, contain digits, because digits are
488shared across many scripts. Digits and similar groups, like punctuation, are in
489the script called C<Common>. There is also a script called C<Inherited> for
490characters that modify other characters, and inherit the script value of the
491controlling character.
492
493For more about scripts versus blocks, see UAX#24 "Unicode Script Property":
494L<http://www.unicode.org/reports/tr24>
495
496The Script property is likely to be the one you want to use when processing
497natural language; the Block property may be useful in working with the nuts and
498bolts of Unicode.
499
500Block names are matched in the compound form, like C<\p{Block: Arrows}> or
501C<\p{Blk=Hebrew}>. Unlike most other properties only a few block names have a
502Unicode-defined short name. But Perl does provide a (slight) shortcut: You
503can say, for example C<\p{In_Arrows}> or C<\p{In_Hebrew}>. For backwards
504compatibility, the C<In> prefix may be omitted if there is no naming conflict
505with a script or any other property, and you can even use an C<Is> prefix
506instead in those cases. But it is not a good idea to do this, for a couple
507reasons:
508
509=over 4
510
511=item 1
512
513It is confusing. There are many naming conflicts, and you may forget some.
9f815e24 514For example, C<\p{Hebrew}> means the I<script> Hebrew, and NOT the I<block>
51f494cc
KW
515Hebrew. But would you remember that 6 months from now?
516
517=item 2
518
519It is unstable. A new version of Unicode may pre-empt the current meaning by
520creating a property with the same name. There was a time in very early Unicode
9f815e24 521releases when C<\p{Hebrew}> would have matched the I<block> Hebrew; now it
51f494cc 522doesn't.
32293815 523
393fec97
GS
524=back
525
51f494cc
KW
526Some people just prefer to always use C<\p{Block: foo}> and C<\p{Script: bar}>
527instead of the shortcuts, for clarity, and because they can't remember the
528difference between 'In' and 'Is' anyway (or aren't confident that those who
529eventually will read their code will know).
530
531A complete list of blocks and their shortcuts is in L<perluniprops>.
532
9f815e24
KW
533=head3 B<Other Properties>
534
535There are many more properties than the very basic ones described here.
536A complete list is in L<perluniprops>.
537
538Unicode defines all its properties in the compound form, so all single-form
539properties are Perl extensions. A number of these are just synonyms for the
540Unicode ones, but some are genunine extensions, including a couple that are in
541the compound form. And quite a few of these are actually recommended by Unicode
542(in L<http://www.unicode.org/reports/tr18>).
543
544This section gives some details on all the extensions that aren't synonyms for
545compound-form Unicode properties (for those, you'll have to refer to the
546L<Unicode Standard|http://www.unicode.org/reports/tr44>.
547
548=over
549
550=item B<C<\p{All}>>
551
552This matches any of the 1_114_112 Unicode code points. It is a synonym for
553C<\p{Any}>.
554
555=item B<C<\p{Alnum}>>
556
557This matches any C<\p{Alphabetic}> or C<\p{Decimal_Number}> character.
558
559=item B<C<\p{Any}>>
560
561This matches any of the 1_114_112 Unicode code points. It is a synonym for
562C<\p{All}>.
563
564=item B<C<\p{Assigned}>>
565
566This matches any assigned code point; that is, any code point whose general
567category is not Unassigned (or equivalently, not Cn).
568
569=item B<C<\p{Blank}>>
570
571This is the same as C<\h> and C<\p{HorizSpace}>: A character that changes the
572spacing horizontally.
573
574=item B<C<\p{Decomposition_Type: Non_Canonical}>> (Short: C<\p{Dt=NonCanon}>)
575
576Matches a character that has a non-canonical decomposition.
577
578To understand the use of this rarely used property=value combination, it is
579necessary to know some basics about decomposition.
580Consider a character, say H. It could appear with various marks around it,
581such as an acute accent, or a circumflex, or various hooks, circles, arrows,
9d1c51c1 582I<etc.>, above, below, to one side and/or the other, etc. There are many
9f815e24
KW
583possibilities among the world's languages. The number of combinations is
584astronomical, and if there were a character for each combination, it would
585soon exhaust Unicode's more than a million possible characters. So Unicode
586took a different approach: there is a character for the base H, and a
587character for each of the possible marks, and they can be combined variously
588to get a final logical character. So a logical character--what appears to be a
589single character--can be a sequence of more than one individual characters.
590This is called an "extended grapheme cluster". (Perl furnishes the C<\X>
591construct to match such sequences.)
592
593But Unicode's intent is to unify the existing character set standards and
594practices, and a number of pre-existing standards have single characters that
595mean the same thing as some of these combinations. An example is ISO-8859-1,
596which has quite a few of these in the Latin-1 range, an example being "LATIN
597CAPITAL LETTER E WITH ACUTE". Because this character was in this pre-existing
598standard, Unicode added it to its repertoire. But this character is considered
599by Unicode to be equivalent to the sequence consisting of first the character
600"LATIN CAPITAL LETTER E", then the character "COMBINING ACUTE ACCENT".
601
602"LATIN CAPITAL LETTER E WITH ACUTE" is called a "pre-composed" character, and
603the equivalence with the sequence is called canonical equivalence. All
604pre-composed characters are said to have a decomposition (into the equivalent
605sequence) and the decomposition type is also called canonical.
606
607However, many more characters have a different type of decomposition, a
608"compatible" or "non-canonical" decomposition. The sequences that form these
609decompositions are not considered canonically equivalent to the pre-composed
610character. An example, again in the Latin-1 range, is the "SUPERSCRIPT ONE".
611It is kind of like a regular digit 1, but not exactly; its decomposition
612into the digit 1 is called a "compatible" decomposition, specifically a
613"super" decomposition. There are several such compatibility
614decompositions (see L<http://www.unicode.org/reports/tr44>), including one
615called "compat" which means some miscellaneous type of decomposition
616that doesn't fit into the decomposition categories that Unicode has chosen.
617
618Note that most Unicode characters don't have a decomposition, so their
619decomposition type is "None".
620
621Perl has added the C<Non_Canonical> type, for your convenience, to mean any of
622the compatibility decompositions.
623
624=item B<C<\p{Graph}>>
625
626Matches any character that is graphic. Theoretically, this means a character
627that on a printer would cause ink to be used.
628
629=item B<C<\p{HorizSpace}>>
630
631This is the same as C<\h> and C<\p{Blank}>: A character that changes the
632spacing horizontally.
633
634=item B<C<\p{In=*}>>
635
636This is a synonym for C<\p{Present_In=*}>
637
638=item B<C<\p{PerlSpace}>>
639
640This is the same as C<\s>, restricted to ASCII, namely C<S<[ \f\n\r\t]>>.
641
642Mnemonic: Perl's (original) space
643
644=item B<C<\p{PerlWord}>>
645
646This is the same as C<\w>, restricted to ASCII, namely C<[A-Za-z0-9_]>
647
648Mnemonic: Perl's (original) word.
649
650=item B<C<\p{PosixAlnum}>>
651
652This matches any alphanumeric character in the ASCII range, namely
653C<[A-Za-z0-9]>.
654
655=item B<C<\p{PosixAlpha}>>
656
657This matches any alphabetic character in the ASCII range, namely C<[A-Za-z]>.
658
659=item B<C<\p{PosixBlank}>>
660
661This matches any blank character in the ASCII range, namely C<S<[ \t]>>.
662
663=item B<C<\p{PosixCntrl}>>
664
665This matches any control character in the ASCII range, namely C<[\x00-\x1F\x7F]>
666
667=item B<C<\p{PosixDigit}>>
668
669This matches any digit character in the ASCII range, namely C<[0-9]>.
670
671=item B<C<\p{PosixGraph}>>
672
673This matches any graphical character in the ASCII range, namely C<[\x21-\x7E]>.
674
675=item B<C<\p{PosixLower}>>
676
677This matches any lowercase character in the ASCII range, namely C<[a-z]>.
678
679=item B<C<\p{PosixPrint}>>
680
681This matches any printable character in the ASCII range, namely C<[\x20-\x7E]>.
682These are the graphical characters plus SPACE.
683
684=item B<C<\p{PosixPunct}>>
685
686This matches any punctuation character in the ASCII range, namely
687C<[\x21-\x2F\x3A-\x40\x5B-\x60\x7B-\x7E]>. These are the
688graphical characters that aren't word characters. Note that the Posix standard
689includes in its definition of punctuation, those characters that Unicode calls
690"symbols."
691
692=item B<C<\p{PosixSpace}>>
693
694This matches any space character in the ASCII range, namely
695C<S<[ \f\n\r\t\x0B]>> (the last being a vertical tab).
696
697=item B<C<\p{PosixUpper}>>
698
699This matches any uppercase character in the ASCII range, namely C<[A-Z]>.
700
701=item B<C<\p{Present_In: *}>> (Short: C<\p{In=*}>)
702
703This property is used when you need to know in what Unicode version(s) a
704character is.
705
706The "*" above stands for some two digit Unicode version number, such as
707C<1.1> or C<4.0>; or the "*" can also be C<Unassigned>. This property will
708match the code points whose final disposition has been settled as of the
709Unicode release given by the version number; C<\p{Present_In: Unassigned}>
710will match those code points whose meaning has yet to be assigned.
711
712For example, C<U+0041> "LATIN CAPITAL LETTER A" was present in the very first
713Unicode release available, which is C<1.1>, so this property is true for all
714valid "*" versions. On the other hand, C<U+1EFF> was not assigned until version
7155.1 when it became "LATIN SMALL LETTER Y WITH LOOP", so the only "*" that
716would match it are 5.1, 5.2, and later.
717
718Unicode furnishes the C<Age> property from which this is derived. The problem
719with Age is that a strict interpretation of it (which Perl takes) has it
720matching the precise release a code point's meaning is introduced in. Thus
721C<U+0041> would match only 1.1; and C<U+1EFF> only 5.1. This is not usually what
722you want.
723
724Some non-Perl implementations of the Age property may change its meaning to be
725the same as the Perl Present_In property; just be aware of that.
726
727Another confusion with both these properties is that the definition is not
728that the code point has been assigned, but that the meaning of the code point
729has been determined. This is because 66 code points will always be
730unassigned, and, so the Age for them is the Unicode version the decision to
731make them so was made in. For example, C<U+FDD0> is to be permanently
732unassigned to a character, and the decision to do that was made in version 3.1,
733so C<\p{Age=3.1}> matches this character and C<\p{Present_In: 3.1}> and up
734matches as well.
735
736=item B<C<\p{Print}>>
737
ae5b72c8 738This matches any character that is graphical or blank, except controls.
9f815e24
KW
739
740=item B<C<\p{SpacePerl}>>
741
742This is the same as C<\s>, including beyond ASCII.
743
4d4acfba
KW
744Mnemonic: Space, as modified by Perl. (It doesn't include the vertical tab
745which both the Posix standard and Unicode consider to be space.)
9f815e24
KW
746
747=item B<C<\p{VertSpace}>>
748
749This is the same as C<\v>: A character that changes the spacing vertically.
750
751=item B<C<\p{Word}>>
752
753This is the same as C<\w>, including beyond ASCII.
754
755=back
756
376d9008 757=head2 User-Defined Character Properties
491fd90a 758
51f494cc
KW
759You can define your own binary character properties by defining subroutines
760whose names begin with "In" or "Is". The subroutines can be defined in any
761package. The user-defined properties can be used in the regular expression
762C<\p> and C<\P> constructs; if you are using a user-defined property from a
763package other than the one you are in, you must specify its package in the
764C<\p> or C<\P> construct.
bac0b425 765
51f494cc 766 # assuming property Is_Foreign defined in Lang::
bac0b425
JP
767 package main; # property package name required
768 if ($txt =~ /\p{Lang::IsForeign}+/) { ... }
769
770 package Lang; # property package name not required
771 if ($txt =~ /\p{IsForeign}+/) { ... }
772
773
774Note that the effect is compile-time and immutable once defined.
491fd90a 775
376d9008
JB
776The subroutines must return a specially-formatted string, with one
777or more newline-separated lines. Each line must be one of the following:
491fd90a
JH
778
779=over 4
780
781=item *
782
510254c9
A
783A single hexadecimal number denoting a Unicode code point to include.
784
785=item *
786
99a6b1f0 787Two hexadecimal numbers separated by horizontal whitespace (space or
376d9008 788tabular characters) denoting a range of Unicode code points to include.
491fd90a
JH
789
790=item *
791
376d9008 792Something to include, prefixed by "+": a built-in character
bac0b425
JP
793property (prefixed by "utf8::") or a user-defined character property,
794to represent all the characters in that property; two hexadecimal code
795points for a range; or a single hexadecimal code point.
491fd90a
JH
796
797=item *
798
376d9008 799Something to exclude, prefixed by "-": an existing character
bac0b425
JP
800property (prefixed by "utf8::") or a user-defined character property,
801to represent all the characters in that property; two hexadecimal code
802points for a range; or a single hexadecimal code point.
491fd90a
JH
803
804=item *
805
376d9008 806Something to negate, prefixed "!": an existing character
bac0b425
JP
807property (prefixed by "utf8::") or a user-defined character property,
808to represent all the characters in that property; two hexadecimal code
809points for a range; or a single hexadecimal code point.
810
811=item *
812
813Something to intersect with, prefixed by "&": an existing character
814property (prefixed by "utf8::") or a user-defined character property,
815for all the characters except the characters in the property; two
816hexadecimal code points for a range; or a single hexadecimal code point.
491fd90a
JH
817
818=back
819
820For example, to define a property that covers both the Japanese
821syllabaries (hiragana and katakana), you can define
822
823 sub InKana {
d88362ca 824 return <<END;
d5822f25
A
825 3040\t309F
826 30A0\t30FF
491fd90a
JH
827 END
828 }
829
d5822f25
A
830Imagine that the here-doc end marker is at the beginning of the line.
831Now you can use C<\p{InKana}> and C<\P{InKana}>.
491fd90a
JH
832
833You could also have used the existing block property names:
834
835 sub InKana {
d88362ca 836 return <<'END';
491fd90a
JH
837 +utf8::InHiragana
838 +utf8::InKatakana
839 END
840 }
841
842Suppose you wanted to match only the allocated characters,
d5822f25 843not the raw block ranges: in other words, you want to remove
491fd90a
JH
844the non-characters:
845
846 sub InKana {
d88362ca 847 return <<'END';
491fd90a
JH
848 +utf8::InHiragana
849 +utf8::InKatakana
850 -utf8::IsCn
851 END
852 }
853
854The negation is useful for defining (surprise!) negated classes.
855
856 sub InNotKana {
d88362ca 857 return <<'END';
491fd90a
JH
858 !utf8::InHiragana
859 -utf8::InKatakana
860 +utf8::IsCn
861 END
862 }
863
bac0b425
JP
864Intersection is useful for getting the common characters matched by
865two (or more) classes.
866
867 sub InFooAndBar {
868 return <<'END';
869 +main::Foo
870 &main::Bar
871 END
872 }
873
ac036724 874It's important to remember not to use "&" for the first set; that
bac0b425
JP
875would be intersecting with nothing (resulting in an empty set).
876
822502e5
ST
877=head2 User-Defined Case Mappings
878
d5cd9e7b
KW
879You can also define your own mappings to be used in C<lc()>,
880C<lcfirst()>, C<uc()>, and C<ucfirst()> (or their string-inlined versions,
881C<\L>, C<\l>, C<\U>, and C<\u>).
822502e5 882The principle is similar to that of user-defined character
51f494cc 883properties: to define subroutines
d5cd9e7b
KW
884with names C<ToLower> (for C<lc()> and C<lcfirst()>); C<ToTitle> (for
885C<ucfirst()>); and C<ToUpper> (for C<uc()>).
3a2263fe 886
51f494cc 887The string returned by the subroutines needs to be two hexadecimal numbers
e1b711da
KW
888separated by two tabulators: the two numbers being, respectively, the source
889code point and the destination code point. For example:
3a2263fe
RGS
890
891 sub ToUpper {
d88362ca 892 return <<END;
51f494cc 893 0061\t\t0041
3a2263fe
RGS
894 END
895 }
896
d5cd9e7b 897defines a mapping for C<uc()> (and C<\U>) that causes only the character "a"
51f494cc 898to be mapped to "A"; all other characters will remain unchanged.
3a2263fe 899
51f494cc
KW
900(For serious hackers only) The above means you have to furnish a complete
901mapping; you can't just override a couple of characters and leave the rest
71648f9a 902unchanged. You can find all the official mappings in the directory
d5cd9e7b
KW
903C<$Config{privlib}>F</unicore/To/>. The mapping data is returned as the
904here-document. The C<utf8::ToSpecI<Foo>> hashes in those files are special
905exception mappings derived from
71648f9a 906C<$Config{privlib}>F</unicore/SpecialCasing.txt>. (The "Digit" and
9f815e24 907"Fold" mappings that one can see in the directory are not directly
d5cd9e7b 908user-accessible, one can use either the L<Unicode::UCD> module, or just match
71648f9a
KW
909case-insensitively, which is what uses the "Fold" mapping. Neither are user
910overridable.)
3a2263fe 911
71648f9a
KW
912If you have many mappings to change, you can take the official mapping data,
913change by hand the affected code points, and place the whole thing into your
914subroutine. But this will only be valid on Perls that use the same Unicode
915version. Another option would be to have your subroutine read the official
916mapping file(s) and overwrite the affected code points.
3a2263fe 917
71648f9a
KW
918If you have only a few mappings to change, starting in 5.14 you can use the
919following trick, here illustrated for Turkish.
920
921 use Config;
922
923 sub ToUpper {
924 my $official = do "$Config{privlib}/unicore/To/Upper.pl";
925 $utf8::ToSpecUpper{'i'} =
926 "\N{LATIN CAPITAL LETTER I WITH DOT ABOVE}";
927 return $official;
928 }
929
930This takes the official mappings and overrides just one, for "LATIN SMALL
931LETTER I". The keys to the hash must be in UTF-8 (or on EBCDIC platforms,
932UTF-EBCDIC), as illustrated by the inverse function.
933
934 sub ToLower {
935 my $official = do $lower;
936 $utf8::ToSpecLower{"\xc4\xb0"} = "i";
937 return $official;
938 }
939
940This example is for an ASCII platform, and C<\xc4\xb0> is the UTF-8 string that
941represents C<\N{LATIN CAPITAL LETTER I WITH DOT ABOVE}>, C<U+0130>.
942
943(The trick illustrated here does work in earlier releases, but only if all the
944characters you want to override have ordinal values of 256 or higher.)
945
946The mappings are in effect only for the package they are defined in, and only
947on scalars that have been marked as having Unicode characters, for example by
948using C<utf8::upgrade()>. You can get around the latter restriction in the
949scope of a C<S<use subs>>:
950
951 use subs qw(uc ucfirst lc lcfirst);
952
953 sub uc($) {
954 my $string = shift;
955 utf8::upgrade($string);
956 return CORE::uc($string);
957 }
958
959 sub lc($) {
960 my $string = shift;
961 utf8::upgrade($string);
962
963 # Unless an I is before a dot_above, it turns into a dotless i.
964 $string =~
965 s/I (?! [^\p{ccc=0}\p{ccc=Above}]* \x{0307} )/\x{131}/gx;
966
967 # But when the I is followed by a dot_above, remove the
968 # dot_above so the end result will be i.
969 $string =~ s/I ([^\p{ccc=0}\p{ccc=Above}]* ) \x{0307}/i$1/gx;
970 return CORE::lc($string);
971 }
972
973These examples (also for Turkish) make sure the input is in UTF-8, and then
974call the corresponding official function, which will use the C<ToUpper()> and
975C<ToLower()> functions you have defined in the package. The C<lc()> example
976shows how you can add context-dependent casing. (For Turkish, there other
977required functions: C<ucfirst>, C<lcfirst>, and C<ToTitle>. These are very
978similar to the ones given above.)
3a2263fe 979
376d9008 980=head2 Character Encodings for Input and Output
8cbd9a7a 981
7221edc9 982See L<Encode>.
8cbd9a7a 983
c29a771d 984=head2 Unicode Regular Expression Support Level
776f8809 985
376d9008
JB
986The following list of Unicode support for regular expressions describes
987all the features currently supported. The references to "Level N"
8158862b
ST
988and the section numbers refer to the Unicode Technical Standard #18,
989"Unicode Regular Expressions", version 11, in May 2005.
776f8809
JH
990
991=over 4
992
993=item *
994
995Level 1 - Basic Unicode Support
996
d88362ca
KW
997 RL1.1 Hex Notation - done [1]
998 RL1.2 Properties - done [2][3]
999 RL1.2a Compatibility Properties - done [4]
1000 RL1.3 Subtraction and Intersection - MISSING [5]
1001 RL1.4 Simple Word Boundaries - done [6]
1002 RL1.5 Simple Loose Matches - done [7]
1003 RL1.6 Line Boundaries - MISSING [8]
1004 RL1.7 Supplementary Code Points - done [9]
8158862b
ST
1005
1006 [1] \x{...}
1007 [2] \p{...} \P{...}
d88362ca
KW
1008 [3] supports not only minimal list, but all Unicode character
1009 properties (see L</Unicode Character Properties>)
8158862b
ST
1010 [4] \d \D \s \S \w \W \X [:prop:] [:^prop:]
1011 [5] can use regular expression look-ahead [a] or
d88362ca
KW
1012 user-defined character properties [b] to emulate set
1013 operations
8158862b 1014 [6] \b \B
d88362ca
KW
1015 [7] note that Perl does Full case-folding in matching (but with
1016 bugs), not Simple: for example U+1F88 is equivalent to
1017 U+1F00 U+03B9, not with 1F80. This difference matters
1018 mainly for certain Greek capital letters with certain
1019 modifiers: the Full case-folding decomposes the letter,
1020 while the Simple case-folding would map it to a single
1021 character.
1022 [8] should do ^ and $ also on U+000B (\v in C), FF (\f), CR
1023 (\r), CRLF (\r\n), NEL (U+0085), LS (U+2028), and PS
1024 (U+2029); should also affect <>, $., and script line
1025 numbers; should not split lines within CRLF [c] (i.e. there
1026 is no empty line between \r and \n)
1027 [9] UTF-8/UTF-EBDDIC used in perl allows not only U+10000 to
1028 U+10FFFF but also beyond U+10FFFF [d]
7207e29d 1029
237bad5b 1030[a] You can mimic class subtraction using lookahead.
8158862b 1031For example, what UTS#18 might write as
29bdacb8 1032
dbe420b4
JH
1033 [{Greek}-[{UNASSIGNED}]]
1034
1035in Perl can be written as:
1036
1d81abf3
JH
1037 (?!\p{Unassigned})\p{InGreekAndCoptic}
1038 (?=\p{Assigned})\p{InGreekAndCoptic}
dbe420b4
JH
1039
1040But in this particular example, you probably really want
1041
1bfb14c4 1042 \p{GreekAndCoptic}
dbe420b4
JH
1043
1044which will match assigned characters known to be part of the Greek script.
29bdacb8 1045
5ca1ac52 1046Also see the Unicode::Regex::Set module, it does implement the full
8158862b
ST
1047UTS#18 grouping, intersection, union, and removal (subtraction) syntax.
1048
1049[b] '+' for union, '-' for removal (set-difference), '&' for intersection
1050(see L</"User-Defined Character Properties">)
1051
1052[c] Try the C<:crlf> layer (see L<PerlIO>).
5ca1ac52 1053
c670e63a
KW
1054[d] U+FFFF will currently generate a warning message if 'utf8' warnings are
1055 enabled
237bad5b 1056
776f8809
JH
1057=item *
1058
1059Level 2 - Extended Unicode Support
1060
8158862b 1061 RL2.1 Canonical Equivalents - MISSING [10][11]
c670e63a 1062 RL2.2 Default Grapheme Clusters - MISSING [12]
8158862b
ST
1063 RL2.3 Default Word Boundaries - MISSING [14]
1064 RL2.4 Default Loose Matches - MISSING [15]
1065 RL2.5 Name Properties - MISSING [16]
1066 RL2.6 Wildcard Properties - MISSING
1067
1068 [10] see UAX#15 "Unicode Normalization Forms"
1069 [11] have Unicode::Normalize but not integrated to regexes
e1b711da 1070 [12] have \X but we don't have a "Grapheme Cluster Mode"
8158862b
ST
1071 [14] see UAX#29, Word Boundaries
1072 [15] see UAX#21 "Case Mappings"
5bd59e57 1073 [16] missing loose match [e]
8158862b
ST
1074
1075[e] C<\N{...}> allows namespaces (see L<charnames>).
776f8809
JH
1076
1077=item *
1078
8158862b
ST
1079Level 3 - Tailored Support
1080
1081 RL3.1 Tailored Punctuation - MISSING
1082 RL3.2 Tailored Grapheme Clusters - MISSING [17][18]
1083 RL3.3 Tailored Word Boundaries - MISSING
1084 RL3.4 Tailored Loose Matches - MISSING
1085 RL3.5 Tailored Ranges - MISSING
1086 RL3.6 Context Matching - MISSING [19]
1087 RL3.7 Incremental Matches - MISSING
1088 ( RL3.8 Unicode Set Sharing )
1089 RL3.9 Possible Match Sets - MISSING
1090 RL3.10 Folded Matching - MISSING [20]
1091 RL3.11 Submatchers - MISSING
1092
1093 [17] see UAX#10 "Unicode Collation Algorithms"
1094 [18] have Unicode::Collate but not integrated to regexes
d88362ca
KW
1095 [19] have (?<=x) and (?=x), but look-aheads or look-behinds
1096 should see outside of the target substring
1097 [20] need insensitive matching for linguistic features other
1098 than case; for example, hiragana to katakana, wide and
1099 narrow, simplified Han to traditional Han (see UTR#30
1100 "Character Foldings")
776f8809
JH
1101
1102=back
1103
c349b1b9
JH
1104=head2 Unicode Encodings
1105
376d9008
JB
1106Unicode characters are assigned to I<code points>, which are abstract
1107numbers. To use these numbers, various encodings are needed.
c349b1b9
JH
1108
1109=over 4
1110
c29a771d 1111=item *
5cb3728c
RB
1112
1113UTF-8
c349b1b9 1114
3e4dbfed 1115UTF-8 is a variable-length (1 to 6 bytes, current character allocations
376d9008
JB
1116require 4 bytes), byte-order independent encoding. For ASCII (and we
1117really do mean 7-bit ASCII, not another 8-bit encoding), UTF-8 is
1118transparent.
c349b1b9 1119
8c007b5a 1120The following table is from Unicode 3.2.
05632f9a 1121
d88362ca 1122 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
05632f9a 1123
d88362ca 1124 U+0000..U+007F 00..7F
e1b711da 1125 U+0080..U+07FF * C2..DF 80..BF
d88362ca 1126 U+0800..U+0FFF E0 * A0..BF 80..BF
ec90690f
ST
1127 U+1000..U+CFFF E1..EC 80..BF 80..BF
1128 U+D000..U+D7FF ED 80..9F 80..BF
e1b711da 1129 U+D800..U+DFFF +++++++ utf16 surrogates, not legal utf8 +++++++
ec90690f 1130 U+E000..U+FFFF EE..EF 80..BF 80..BF
d88362ca
KW
1131 U+10000..U+3FFFF F0 * 90..BF 80..BF 80..BF
1132 U+40000..U+FFFFF F1..F3 80..BF 80..BF 80..BF
1133 U+100000..U+10FFFF F4 80..8F 80..BF 80..BF
e1b711da
KW
1134
1135Note the gaps before several of the byte entries above marked by '*'. These are
1136caused by legal UTF-8 avoiding non-shortest encodings: it is technically
1137possible to UTF-8-encode a single code point in different ways, but that is
1138explicitly forbidden, and the shortest possible encoding should always be used
1139(and that is what Perl does).
37361303 1140
376d9008 1141Another way to look at it is via bits:
05632f9a
JH
1142
1143 Code Points 1st Byte 2nd Byte 3rd Byte 4th Byte
1144
1145 0aaaaaaa 0aaaaaaa
1146 00000bbbbbaaaaaa 110bbbbb 10aaaaaa
1147 ccccbbbbbbaaaaaa 1110cccc 10bbbbbb 10aaaaaa
1148 00000dddccccccbbbbbbaaaaaa 11110ddd 10cccccc 10bbbbbb 10aaaaaa
1149
9f815e24 1150As you can see, the continuation bytes all begin with "10", and the
e1b711da 1151leading bits of the start byte tell how many bytes there are in the
05632f9a
JH
1152encoded character.
1153
c29a771d 1154=item *
5cb3728c
RB
1155
1156UTF-EBCDIC
dbe420b4 1157
376d9008 1158Like UTF-8 but EBCDIC-safe, in the way that UTF-8 is ASCII-safe.
dbe420b4 1159
c29a771d 1160=item *
5cb3728c 1161
1e54db1a 1162UTF-16, UTF-16BE, UTF-16LE, Surrogates, and BOMs (Byte Order Marks)
c349b1b9 1163
1bfb14c4
JH
1164The followings items are mostly for reference and general Unicode
1165knowledge, Perl doesn't use these constructs internally.
dbe420b4 1166
c349b1b9 1167UTF-16 is a 2 or 4 byte encoding. The Unicode code points
1bfb14c4
JH
1168C<U+0000..U+FFFF> are stored in a single 16-bit unit, and the code
1169points C<U+10000..U+10FFFF> in two 16-bit units. The latter case is
c349b1b9
JH
1170using I<surrogates>, the first 16-bit unit being the I<high
1171surrogate>, and the second being the I<low surrogate>.
1172
376d9008 1173Surrogates are code points set aside to encode the C<U+10000..U+10FFFF>
c349b1b9 1174range of Unicode code points in pairs of 16-bit units. The I<high
9f815e24 1175surrogates> are the range C<U+D800..U+DBFF> and the I<low surrogates>
376d9008 1176are the range C<U+DC00..U+DFFF>. The surrogate encoding is
c349b1b9 1177
d88362ca
KW
1178 $hi = ($uni - 0x10000) / 0x400 + 0xD800;
1179 $lo = ($uni - 0x10000) % 0x400 + 0xDC00;
c349b1b9
JH
1180
1181and the decoding is
1182
d88362ca 1183 $uni = 0x10000 + ($hi - 0xD800) * 0x400 + ($lo - 0xDC00);
c349b1b9 1184
feda178f 1185If you try to generate surrogates (for example by using chr()), you
e1b711da 1186will get a warning, if warnings are turned on, because those code
376d9008 1187points are not valid for a Unicode character.
9466bab6 1188
376d9008 1189Because of the 16-bitness, UTF-16 is byte-order dependent. UTF-16
c349b1b9 1190itself can be used for in-memory computations, but if storage or
376d9008
JB
1191transfer is required either UTF-16BE (big-endian) or UTF-16LE
1192(little-endian) encodings must be chosen.
c349b1b9
JH
1193
1194This introduces another problem: what if you just know that your data
376d9008
JB
1195is UTF-16, but you don't know which endianness? Byte Order Marks, or
1196BOMs, are a solution to this. A special character has been reserved
86bbd6d1 1197in Unicode to function as a byte order marker: the character with the
376d9008 1198code point C<U+FEFF> is the BOM.
042da322 1199
c349b1b9 1200The trick is that if you read a BOM, you will know the byte order,
376d9008
JB
1201since if it was written on a big-endian platform, you will read the
1202bytes C<0xFE 0xFF>, but if it was written on a little-endian platform,
1203you will read the bytes C<0xFF 0xFE>. (And if the originating platform
1204was writing in UTF-8, you will read the bytes C<0xEF 0xBB 0xBF>.)
042da322 1205
86bbd6d1 1206The way this trick works is that the character with the code point
376d9008
JB
1207C<U+FFFE> is guaranteed not to be a valid Unicode character, so the
1208sequence of bytes C<0xFF 0xFE> is unambiguously "BOM, represented in
1bfb14c4 1209little-endian format" and cannot be C<U+FFFE>, represented in big-endian
e1b711da
KW
1210format". (Actually, C<U+FFFE> is legal for use by your program, even for
1211input/output, but better not use it if you need a BOM. But it is "illegal for
1212interchange", so that an unsuspecting program won't get confused.)
c349b1b9 1213
c29a771d 1214=item *
5cb3728c 1215
1e54db1a 1216UTF-32, UTF-32BE, UTF-32LE
c349b1b9
JH
1217
1218The UTF-32 family is pretty much like the UTF-16 family, expect that
042da322 1219the units are 32-bit, and therefore the surrogate scheme is not
376d9008
JB
1220needed. The BOM signatures will be C<0x00 0x00 0xFE 0xFF> for BE and
1221C<0xFF 0xFE 0x00 0x00> for LE.
c349b1b9 1222
c29a771d 1223=item *
5cb3728c
RB
1224
1225UCS-2, UCS-4
c349b1b9 1226
86bbd6d1 1227Encodings defined by the ISO 10646 standard. UCS-2 is a 16-bit
376d9008 1228encoding. Unlike UTF-16, UCS-2 is not extensible beyond C<U+FFFF>,
339cfa0e
JH
1229because it does not use surrogates. UCS-4 is a 32-bit encoding,
1230functionally identical to UTF-32.
c349b1b9 1231
c29a771d 1232=item *
5cb3728c
RB
1233
1234UTF-7
c349b1b9 1235
376d9008
JB
1236A seven-bit safe (non-eight-bit) encoding, which is useful if the
1237transport or storage is not eight-bit safe. Defined by RFC 2152.
c349b1b9 1238
95a1a48b
JH
1239=back
1240
0d7c09bb
JH
1241=head2 Security Implications of Unicode
1242
e1b711da
KW
1243Read L<Unicode Security Considerations|http://www.unicode.org/reports/tr36>.
1244Also, note the following:
1245
0d7c09bb
JH
1246=over 4
1247
1248=item *
1249
1250Malformed UTF-8
bf0fa0b2
JH
1251
1252Unfortunately, the specification of UTF-8 leaves some room for
1253interpretation of how many bytes of encoded output one should generate
376d9008
JB
1254from one input Unicode character. Strictly speaking, the shortest
1255possible sequence of UTF-8 bytes should be generated,
1256because otherwise there is potential for an input buffer overflow at
feda178f 1257the receiving end of a UTF-8 connection. Perl always generates the
e1b711da 1258shortest length UTF-8, and with warnings on, Perl will warn about
376d9008
JB
1259non-shortest length UTF-8 along with other malformations, such as the
1260surrogates, which are not real Unicode code points.
bf0fa0b2 1261
0d7c09bb
JH
1262=item *
1263
1264Regular expressions behave slightly differently between byte data and
376d9008
JB
1265character (Unicode) data. For example, the "word character" character
1266class C<\w> will work differently depending on if data is eight-bit bytes
1267or Unicode.
0d7c09bb 1268
376d9008
JB
1269In the first case, the set of C<\w> characters is either small--the
1270default set of alphabetic characters, digits, and the "_"--or, if you
0d7c09bb
JH
1271are using a locale (see L<perllocale>), the C<\w> might contain a few
1272more letters according to your language and country.
1273
376d9008 1274In the second case, the C<\w> set of characters is much, much larger.
1bfb14c4
JH
1275Most importantly, even in the set of the first 256 characters, it will
1276probably match different characters: unlike most locales, which are
1277specific to a language and country pair, Unicode classifies all the
1278characters that are letters I<somewhere> as C<\w>. For example, your
1279locale might not think that LATIN SMALL LETTER ETH is a letter (unless
1280you happen to speak Icelandic), but Unicode does.
0d7c09bb 1281
376d9008 1282As discussed elsewhere, Perl has one foot (two hooves?) planted in
1bfb14c4
JH
1283each of two worlds: the old world of bytes and the new world of
1284characters, upgrading from bytes to characters when necessary.
376d9008
JB
1285If your legacy code does not explicitly use Unicode, no automatic
1286switch-over to characters should happen. Characters shouldn't get
1bfb14c4
JH
1287downgraded to bytes, either. It is possible to accidentally mix bytes
1288and characters, however (see L<perluniintro>), in which case C<\w> in
1289regular expressions might start behaving differently. Review your
1290code. Use warnings and the C<strict> pragma.
0d7c09bb
JH
1291
1292=back
1293
c349b1b9
JH
1294=head2 Unicode in Perl on EBCDIC
1295
376d9008
JB
1296The way Unicode is handled on EBCDIC platforms is still
1297experimental. On such platforms, references to UTF-8 encoding in this
1298document and elsewhere should be read as meaning the UTF-EBCDIC
1299specified in Unicode Technical Report 16, unless ASCII vs. EBCDIC issues
c349b1b9 1300are specifically discussed. There is no C<utfebcdic> pragma or
376d9008 1301":utfebcdic" layer; rather, "utf8" and ":utf8" are reused to mean
86bbd6d1
PN
1302the platform's "natural" 8-bit encoding of Unicode. See L<perlebcdic>
1303for more discussion of the issues.
c349b1b9 1304
b310b053
JH
1305=head2 Locales
1306
4616122b 1307Usually locale settings and Unicode do not affect each other, but
b310b053
JH
1308there are a couple of exceptions:
1309
1310=over 4
1311
1312=item *
1313
8aa8f774
JH
1314You can enable automatic UTF-8-ification of your standard file
1315handles, default C<open()> layer, and C<@ARGV> by using either
1316the C<-C> command line switch or the C<PERL_UNICODE> environment
1317variable, see L<perlrun> for the documentation of the C<-C> switch.
b310b053
JH
1318
1319=item *
1320
376d9008
JB
1321Perl tries really hard to work both with Unicode and the old
1322byte-oriented world. Most often this is nice, but sometimes Perl's
1323straddling of the proverbial fence causes problems.
b310b053
JH
1324
1325=back
1326
1aad1664
JH
1327=head2 When Unicode Does Not Happen
1328
1329While Perl does have extensive ways to input and output in Unicode,
1330and few other 'entry points' like the @ARGV which can be interpreted
1331as Unicode (UTF-8), there still are many places where Unicode (in some
1332encoding or another) could be given as arguments or received as
1333results, or both, but it is not.
1334
e1b711da
KW
1335The following are such interfaces. Also, see L</The "Unicode Bug">.
1336For all of these interfaces Perl
6cd4dd6c
JH
1337currently (as of 5.8.3) simply assumes byte strings both as arguments
1338and results, or UTF-8 strings if the C<encoding> pragma has been used.
1aad1664
JH
1339
1340One reason why Perl does not attempt to resolve the role of Unicode in
e1b711da 1341these cases is that the answers are highly dependent on the operating
1aad1664
JH
1342system and the file system(s). For example, whether filenames can be
1343in Unicode, and in exactly what kind of encoding, is not exactly a
1344portable concept. Similarly for the qx and system: how well will the
1345'command line interface' (and which of them?) handle Unicode?
1346
1347=over 4
1348
557a2462
RB
1349=item *
1350
51f494cc 1351chdir, chmod, chown, chroot, exec, link, lstat, mkdir,
1e8e8236 1352rename, rmdir, stat, symlink, truncate, unlink, utime, -X
557a2462
RB
1353
1354=item *
1355
1356%ENV
1357
1358=item *
1359
1360glob (aka the <*>)
1361
1362=item *
1aad1664 1363
557a2462 1364open, opendir, sysopen
1aad1664 1365
557a2462 1366=item *
1aad1664 1367
557a2462 1368qx (aka the backtick operator), system
1aad1664 1369
557a2462 1370=item *
1aad1664 1371
557a2462 1372readdir, readlink
1aad1664
JH
1373
1374=back
1375
e1b711da
KW
1376=head2 The "Unicode Bug"
1377
1378The term, the "Unicode bug" has been applied to an inconsistency with the
6f335b04 1379Unicode characters whose ordinals are in the Latin-1 Supplement block, that
e1b711da
KW
1380is, between 128 and 255. Without a locale specified, unlike all other
1381characters or code points, these characters have very different semantics in
1382byte semantics versus character semantics.
1383
1384In character semantics they are interpreted as Unicode code points, which means
1385they have the same semantics as Latin-1 (ISO-8859-1).
1386
1387In byte semantics, they are considered to be unassigned characters, meaning
1388that the only semantics they have is their ordinal numbers, and that they are
1389not members of various character classes. None are considered to match C<\w>
1390for example, but all match C<\W>. (On EBCDIC platforms, the behavior may
1391be different from this, depending on the underlying C language library
1392functions.)
1393
1394The behavior is known to have effects on these areas:
1395
1396=over 4
1397
1398=item *
1399
1400Changing the case of a scalar, that is, using C<uc()>, C<ucfirst()>, C<lc()>,
1401and C<lcfirst()>, or C<\L>, C<\U>, C<\u> and C<\l> in regular expression
1402substitutions.
1403
1404=item *
1405
1406Using caseless (C</i>) regular expression matching
1407
1408=item *
1409
630d17dc
KW
1410Matching a number of properties in regular expressions, namely C<\b>,
1411C<\B>, C<\s>, C<\S>, C<\w>, C<\W>, and all the Posix character classes
1412I<except> C<[[:ascii:]]>.
e1b711da
KW
1413
1414=item *
1415
1416User-defined case change mappings. You can create a C<ToUpper()> function, for
1417example, which overrides Perl's built-in case mappings. The scalar must be
1418encoded in utf8 for your function to actually be invoked.
1419
1420=back
1421
1422This behavior can lead to unexpected results in which a string's semantics
1423suddenly change if a code point above 255 is appended to or removed from it,
1424which changes the string's semantics from byte to character or vice versa. As
1425an example, consider the following program and its output:
1426
1427 $ perl -le'
1428 $s1 = "\xC2";
1429 $s2 = "\x{2660}";
1430 for ($s1, $s2, $s1.$s2) {
1431 print /\w/ || 0;
1432 }
1433 '
1434 0
1435 0
1436 1
1437
9f815e24 1438If there's no C<\w> in C<s1> or in C<s2>, why does their concatenation have one?
e1b711da
KW
1439
1440This anomaly stems from Perl's attempt to not disturb older programs that
1441didn't use Unicode, and hence had no semantics for characters outside of the
1442ASCII range (except in a locale), along with Perl's desire to add Unicode
1443support seamlessly. The result wasn't seamless: these characters were
1444orphaned.
1445
1446Work is being done to correct this, but only some of it was complete in time
1447for the 5.12 release. What has been finished is the important part of the case
1448changing component. Due to concerns, and some evidence, that older code might
1449have come to rely on the existing behavior, the new behavior must be explicitly
1450enabled by the feature C<unicode_strings> in the L<feature> pragma, even though
1451no new syntax is involved.
1452
1453See L<perlfunc/lc> for details on how this pragma works in combination with
1454various others for casing. Even though the pragma only affects casing
1455operations in the 5.12 release, it is planned to have it affect all the
1456problematic behaviors in later releases: you can't have one without them all.
1457
1458In the meantime, a workaround is to always call utf8::upgrade($string), or to
6f335b04
KW
1459use the standard module L<Encode>. Also, a scalar that has any characters
1460whose ordinal is above 0x100, or which were specified using either of the
1461C<\N{...}> notations will automatically have character semantics.
e1b711da 1462
1aad1664
JH
1463=head2 Forcing Unicode in Perl (Or Unforcing Unicode in Perl)
1464
e1b711da
KW
1465Sometimes (see L</"When Unicode Does Not Happen"> or L</The "Unicode Bug">)
1466there are situations where you simply need to force a byte
2bbc8d55
SP
1467string into UTF-8, or vice versa. The low-level calls
1468utf8::upgrade($bytestring) and utf8::downgrade($utf8string[, FAIL_OK]) are
1aad1664
JH
1469the answers.
1470
2bbc8d55
SP
1471Note that utf8::downgrade() can fail if the string contains characters
1472that don't fit into a byte.
1aad1664 1473
e1b711da
KW
1474Calling either function on a string that already is in the desired state is a
1475no-op.
1476
95a1a48b
JH
1477=head2 Using Unicode in XS
1478
3a2263fe
RGS
1479If you want to handle Perl Unicode in XS extensions, you may find the
1480following C APIs useful. See also L<perlguts/"Unicode Support"> for an
1481explanation about Unicode at the XS level, and L<perlapi> for the API
1482details.
95a1a48b
JH
1483
1484=over 4
1485
1486=item *
1487
1bfb14c4 1488C<DO_UTF8(sv)> returns true if the C<UTF8> flag is on and the bytes
2bbc8d55 1489pragma is not in effect. C<SvUTF8(sv)> returns true if the C<UTF8>
1bfb14c4
JH
1490flag is on; the bytes pragma is ignored. The C<UTF8> flag being on
1491does B<not> mean that there are any characters of code points greater
1492than 255 (or 127) in the scalar or that there are even any characters
1493in the scalar. What the C<UTF8> flag means is that the sequence of
1494octets in the representation of the scalar is the sequence of UTF-8
1495encoded code points of the characters of a string. The C<UTF8> flag
1496being off means that each octet in this representation encodes a
1497single character with code point 0..255 within the string. Perl's
1498Unicode model is not to use UTF-8 until it is absolutely necessary.
95a1a48b
JH
1499
1500=item *
1501
2bbc8d55 1502C<uvchr_to_utf8(buf, chr)> writes a Unicode character code point into
1bfb14c4 1503a buffer encoding the code point as UTF-8, and returns a pointer
2bbc8d55 1504pointing after the UTF-8 bytes. It works appropriately on EBCDIC machines.
95a1a48b
JH
1505
1506=item *
1507
2bbc8d55 1508C<utf8_to_uvchr(buf, lenp)> reads UTF-8 encoded bytes from a buffer and
376d9008 1509returns the Unicode character code point and, optionally, the length of
2bbc8d55 1510the UTF-8 byte sequence. It works appropriately on EBCDIC machines.
95a1a48b
JH
1511
1512=item *
1513
376d9008
JB
1514C<utf8_length(start, end)> returns the length of the UTF-8 encoded buffer
1515in characters. C<sv_len_utf8(sv)> returns the length of the UTF-8 encoded
95a1a48b
JH
1516scalar.
1517
1518=item *
1519
376d9008
JB
1520C<sv_utf8_upgrade(sv)> converts the string of the scalar to its UTF-8
1521encoded form. C<sv_utf8_downgrade(sv)> does the opposite, if
1522possible. C<sv_utf8_encode(sv)> is like sv_utf8_upgrade except that
1523it does not set the C<UTF8> flag. C<sv_utf8_decode()> does the
1524opposite of C<sv_utf8_encode()>. Note that none of these are to be
1525used as general-purpose encoding or decoding interfaces: C<use Encode>
1526for that. C<sv_utf8_upgrade()> is affected by the encoding pragma
1527but C<sv_utf8_downgrade()> is not (since the encoding pragma is
1528designed to be a one-way street).
95a1a48b
JH
1529
1530=item *
1531
376d9008 1532C<is_utf8_char(s)> returns true if the pointer points to a valid UTF-8
90f968e0 1533character.
95a1a48b
JH
1534
1535=item *
1536
376d9008 1537C<is_utf8_string(buf, len)> returns true if C<len> bytes of the buffer
95a1a48b
JH
1538are valid UTF-8.
1539
1540=item *
1541
376d9008
JB
1542C<UTF8SKIP(buf)> will return the number of bytes in the UTF-8 encoded
1543character in the buffer. C<UNISKIP(chr)> will return the number of bytes
1544required to UTF-8-encode the Unicode character code point. C<UTF8SKIP()>
90f968e0 1545is useful for example for iterating over the characters of a UTF-8
376d9008 1546encoded buffer; C<UNISKIP()> is useful, for example, in computing
90f968e0 1547the size required for a UTF-8 encoded buffer.
95a1a48b
JH
1548
1549=item *
1550
376d9008 1551C<utf8_distance(a, b)> will tell the distance in characters between the
95a1a48b
JH
1552two pointers pointing to the same UTF-8 encoded buffer.
1553
1554=item *
1555
2bbc8d55 1556C<utf8_hop(s, off)> will return a pointer to a UTF-8 encoded buffer
376d9008
JB
1557that is C<off> (positive or negative) Unicode characters displaced
1558from the UTF-8 buffer C<s>. Be careful not to overstep the buffer:
1559C<utf8_hop()> will merrily run off the end or the beginning of the
1560buffer if told to do so.
95a1a48b 1561
d2cc3551
JH
1562=item *
1563
376d9008
JB
1564C<pv_uni_display(dsv, spv, len, pvlim, flags)> and
1565C<sv_uni_display(dsv, ssv, pvlim, flags)> are useful for debugging the
1566output of Unicode strings and scalars. By default they are useful
1567only for debugging--they display B<all> characters as hexadecimal code
1bfb14c4
JH
1568points--but with the flags C<UNI_DISPLAY_ISPRINT>,
1569C<UNI_DISPLAY_BACKSLASH>, and C<UNI_DISPLAY_QQ> you can make the
1570output more readable.
d2cc3551
JH
1571
1572=item *
1573
2bbc8d55 1574C<ibcmp_utf8(s1, pe1, l1, u1, s2, pe2, l2, u2)> can be used to
376d9008
JB
1575compare two strings case-insensitively in Unicode. For case-sensitive
1576comparisons you can just use C<memEQ()> and C<memNE()> as usual.
d2cc3551 1577
c349b1b9
JH
1578=back
1579
95a1a48b
JH
1580For more information, see L<perlapi>, and F<utf8.c> and F<utf8.h>
1581in the Perl source code distribution.
1582
e1b711da
KW
1583=head2 Hacking Perl to work on earlier Unicode versions (for very serious hackers only)
1584
1585Perl by default comes with the latest supported Unicode version built in, but
1586you can change to use any earlier one.
1587
1588Download the files in the version of Unicode that you want from the Unicode web
1589site L<http://www.unicode.org>). These should replace the existing files in
1590C<\$Config{privlib}>/F<unicore>. (C<\%Config> is available from the Config
1591module.) Follow the instructions in F<README.perl> in that directory to change
1592some of their names, and then run F<make>.
1593
1594It is even possible to download them to a different directory, and then change
1595F<utf8_heavy.pl> in the directory C<\$Config{privlib}> to point to the new
1596directory, or maybe make a copy of that directory before making the change, and
1597using C<@INC> or the C<-I> run-time flag to switch between versions at will
1598(but because of caching, not in the middle of a process), but all this is
1599beyond the scope of these instructions.
1600
c29a771d
JH
1601=head1 BUGS
1602
376d9008 1603=head2 Interaction with Locales
7eabb34d 1604
376d9008
JB
1605Use of locales with Unicode data may lead to odd results. Currently,
1606Perl attempts to attach 8-bit locale info to characters in the range
16070..255, but this technique is demonstrably incorrect for locales that
1608use characters above that range when mapped into Unicode. Perl's
1609Unicode support will also tend to run slower. Use of locales with
1610Unicode is discouraged.
c29a771d 1611
9f815e24 1612=head2 Problems with characters in the Latin-1 Supplement range
2bbc8d55 1613
e1b711da
KW
1614See L</The "Unicode Bug">
1615
1616=head2 Problems with case-insensitive regular expression matching
1617
1618There are problems with case-insensitive matches, including those involving
1619character classes (enclosed in [square brackets]), characters whose fold
9f815e24
KW
1620is to multiple characters (such as the single character LATIN SMALL LIGATURE
1621FFL matches case-insensitively with the 3-character string C<ffl>), and
1622characters in the Latin-1 Supplement.
2bbc8d55 1623
376d9008 1624=head2 Interaction with Extensions
7eabb34d 1625
376d9008 1626When Perl exchanges data with an extension, the extension should be
2575c402 1627able to understand the UTF8 flag and act accordingly. If the
376d9008
JB
1628extension doesn't know about the flag, it's likely that the extension
1629will return incorrectly-flagged data.
7eabb34d
A
1630
1631So if you're working with Unicode data, consult the documentation of
1632every module you're using if there are any issues with Unicode data
1633exchange. If the documentation does not talk about Unicode at all,
a73d23f6 1634suspect the worst and probably look at the source to learn how the
376d9008 1635module is implemented. Modules written completely in Perl shouldn't
a73d23f6
RGS
1636cause problems. Modules that directly or indirectly access code written
1637in other programming languages are at risk.
7eabb34d 1638
376d9008 1639For affected functions, the simple strategy to avoid data corruption is
7eabb34d 1640to always make the encoding of the exchanged data explicit. Choose an
376d9008 1641encoding that you know the extension can handle. Convert arguments passed
7eabb34d
A
1642to the extensions to that encoding and convert results back from that
1643encoding. Write wrapper functions that do the conversions for you, so
1644you can later change the functions when the extension catches up.
1645
376d9008 1646To provide an example, let's say the popular Foo::Bar::escape_html
7eabb34d
A
1647function doesn't deal with Unicode data yet. The wrapper function
1648would convert the argument to raw UTF-8 and convert the result back to
376d9008 1649Perl's internal representation like so:
7eabb34d
A
1650
1651 sub my_escape_html ($) {
d88362ca
KW
1652 my($what) = shift;
1653 return unless defined $what;
1654 Encode::decode_utf8(Foo::Bar::escape_html(
1655 Encode::encode_utf8($what)));
7eabb34d
A
1656 }
1657
1658Sometimes, when the extension does not convert data but just stores
1659and retrieves them, you will be in a position to use the otherwise
1660dangerous Encode::_utf8_on() function. Let's say the popular
66b79f27 1661C<Foo::Bar> extension, written in C, provides a C<param> method that
7eabb34d
A
1662lets you store and retrieve data according to these prototypes:
1663
1664 $self->param($name, $value); # set a scalar
1665 $value = $self->param($name); # retrieve a scalar
1666
1667If it does not yet provide support for any encoding, one could write a
1668derived class with such a C<param> method:
1669
1670 sub param {
1671 my($self,$name,$value) = @_;
1672 utf8::upgrade($name); # make sure it is UTF-8 encoded
af55fc6a 1673 if (defined $value) {
7eabb34d
A
1674 utf8::upgrade($value); # make sure it is UTF-8 encoded
1675 return $self->SUPER::param($name,$value);
1676 } else {
1677 my $ret = $self->SUPER::param($name);
1678 Encode::_utf8_on($ret); # we know, it is UTF-8 encoded
1679 return $ret;
1680 }
1681 }
1682
a73d23f6
RGS
1683Some extensions provide filters on data entry/exit points, such as
1684DB_File::filter_store_key and family. Look out for such filters in
66b79f27 1685the documentation of your extensions, they can make the transition to
7eabb34d
A
1686Unicode data much easier.
1687
376d9008 1688=head2 Speed
7eabb34d 1689
c29a771d 1690Some functions are slower when working on UTF-8 encoded strings than
574c8022 1691on byte encoded strings. All functions that need to hop over
7c17141f
JH
1692characters such as length(), substr() or index(), or matching regular
1693expressions can work B<much> faster when the underlying data are
1694byte-encoded.
1695
1696In Perl 5.8.0 the slowness was often quite spectacular; in Perl 5.8.1
1697a caching scheme was introduced which will hopefully make the slowness
a104b433
JH
1698somewhat less spectacular, at least for some operations. In general,
1699operations with UTF-8 encoded strings are still slower. As an example,
1700the Unicode properties (character classes) like C<\p{Nd}> are known to
1701be quite a bit slower (5-20 times) than their simpler counterparts
1702like C<\d> (then again, there 268 Unicode characters matching C<Nd>
1703compared with the 10 ASCII characters matching C<d>).
666f95b9 1704
e1b711da
KW
1705=head2 Problems on EBCDIC platforms
1706
1707There are a number of known problems with Perl on EBCDIC platforms. If you
1708want to use Perl there, send email to perlbug@perl.org.
fe749c9a
KW
1709
1710In earlier versions, when byte and character data were concatenated,
1711the new string was sometimes created by
1712decoding the byte strings as I<ISO 8859-1 (Latin-1)>, even if the
1713old Unicode string used EBCDIC.
1714
1715If you find any of these, please report them as bugs.
1716
c8d992ba
A
1717=head2 Porting code from perl-5.6.X
1718
1719Perl 5.8 has a different Unicode model from 5.6. In 5.6 the programmer
1720was required to use the C<utf8> pragma to declare that a given scope
1721expected to deal with Unicode data and had to make sure that only
1722Unicode data were reaching that scope. If you have code that is
1723working with 5.6, you will need some of the following adjustments to
1724your code. The examples are written such that the code will continue
1725to work under 5.6, so you should be safe to try them out.
1726
1727=over 4
1728
1729=item *
1730
1731A filehandle that should read or write UTF-8
1732
1733 if ($] > 5.007) {
740d4bb2 1734 binmode $fh, ":encoding(utf8)";
c8d992ba
A
1735 }
1736
1737=item *
1738
1739A scalar that is going to be passed to some extension
1740
1741Be it Compress::Zlib, Apache::Request or any extension that has no
1742mention of Unicode in the manpage, you need to make sure that the
2575c402 1743UTF8 flag is stripped off. Note that at the time of this writing
c8d992ba
A
1744(October 2002) the mentioned modules are not UTF-8-aware. Please
1745check the documentation to verify if this is still true.
1746
1747 if ($] > 5.007) {
1748 require Encode;
1749 $val = Encode::encode_utf8($val); # make octets
1750 }
1751
1752=item *
1753
1754A scalar we got back from an extension
1755
1756If you believe the scalar comes back as UTF-8, you will most likely
2575c402 1757want the UTF8 flag restored:
c8d992ba
A
1758
1759 if ($] > 5.007) {
1760 require Encode;
1761 $val = Encode::decode_utf8($val);
1762 }
1763
1764=item *
1765
1766Same thing, if you are really sure it is UTF-8
1767
1768 if ($] > 5.007) {
1769 require Encode;
1770 Encode::_utf8_on($val);
1771 }
1772
1773=item *
1774
1775A wrapper for fetchrow_array and fetchrow_hashref
1776
1777When the database contains only UTF-8, a wrapper function or method is
1778a convenient way to replace all your fetchrow_array and
1779fetchrow_hashref calls. A wrapper function will also make it easier to
1780adapt to future enhancements in your database driver. Note that at the
1781time of this writing (October 2002), the DBI has no standardized way
1782to deal with UTF-8 data. Please check the documentation to verify if
1783that is still true.
1784
1785 sub fetchrow {
d88362ca
KW
1786 # $what is one of fetchrow_{array,hashref}
1787 my($self, $sth, $what) = @_;
c8d992ba
A
1788 if ($] < 5.007) {
1789 return $sth->$what;
1790 } else {
1791 require Encode;
1792 if (wantarray) {
1793 my @arr = $sth->$what;
1794 for (@arr) {
1795 defined && /[^\000-\177]/ && Encode::_utf8_on($_);
1796 }
1797 return @arr;
1798 } else {
1799 my $ret = $sth->$what;
1800 if (ref $ret) {
1801 for my $k (keys %$ret) {
d88362ca
KW
1802 defined
1803 && /[^\000-\177]/
1804 && Encode::_utf8_on($_) for $ret->{$k};
c8d992ba
A
1805 }
1806 return $ret;
1807 } else {
1808 defined && /[^\000-\177]/ && Encode::_utf8_on($_) for $ret;
1809 return $ret;
1810 }
1811 }
1812 }
1813 }
1814
1815
1816=item *
1817
1818A large scalar that you know can only contain ASCII
1819
1820Scalars that contain only ASCII and are marked as UTF-8 are sometimes
1821a drag to your program. If you recognize such a situation, just remove
2575c402 1822the UTF8 flag:
c8d992ba
A
1823
1824 utf8::downgrade($val) if $] > 5.007;
1825
1826=back
1827
393fec97
GS
1828=head1 SEE ALSO
1829
51f494cc 1830L<perlunitut>, L<perluniintro>, L<perluniprops>, L<Encode>, L<open>, L<utf8>, L<bytes>,
a05d7ebb 1831L<perlretut>, L<perlvar/"${^UNICODE}">
51f494cc 1832L<http://www.unicode.org/reports/tr44>).
393fec97
GS
1833
1834=cut