Commit | Line | Data |
393fec97 |
1 | =head1 NAME |
2 | |
3 | perlunicode - Unicode support in Perl |
4 | |
5 | =head1 DESCRIPTION |
6 | |
21bad921 |
7 | =head2 Important Caveat |
8 | |
393fec97 |
9 | WARNING: The implementation of Unicode support in Perl is incomplete. |
21bad921 |
10 | |
11 | The following areas need further work. |
12 | |
13a2d996 |
13 | =over 4 |
21bad921 |
14 | |
15 | =item Input and Output Disciplines |
16 | |
17 | There is currently no easy way to mark data read from a file or other |
18 | external source as being utf8. This will be one of the major areas of |
49cb94c6 |
19 | focus in the near future. Unfortunately it is unlikely that the Perl |
20 | 5.6 and earlier will ever gain this capability. |
21bad921 |
21 | |
22 | =item Regular Expressions |
23 | |
24 | The existing regular expression compiler does not produce polymorphic |
25 | opcodes. This means that the determination on whether to match Unicode |
26 | characters is made when the pattern is compiled, based on whether the |
27 | pattern contains Unicode characters, and not when the matching happens |
28 | at run time. This needs to be changed to adaptively match Unicode if |
29 | the string to be matched is Unicode. |
30 | |
31 | =item C<use utf8> still needed to enable a few features |
32 | |
33 | The C<utf8> pragma implements the tables used for Unicode support. These |
34 | tables are automatically loaded on demand, so the C<utf8> pragma need not |
35 | normally be used. |
36 | |
37 | However, as a compatibility measure, this pragma must be explicitly used |
38 | to enable recognition of UTF-8 encoded literals and identifiers in the |
39 | source text. |
40 | |
41 | =back |
42 | |
43 | =head2 Byte and Character semantics |
393fec97 |
44 | |
45 | Beginning with version 5.6, Perl uses logically wide characters to |
46 | represent strings internally. This internal representation of strings |
47 | uses the UTF-8 encoding. |
48 | |
21bad921 |
49 | In future, Perl-level operations can be expected to work with characters |
393fec97 |
50 | rather than bytes, in general. |
51 | |
8cbd9a7a |
52 | However, as strictly an interim compatibility measure, Perl v5.6 aims to |
53 | provide a safe migration path from byte semantics to character semantics |
54 | for programs. For operations where Perl can unambiguously decide that the |
55 | input data is characters, Perl now switches to character semantics. |
56 | For operations where this determination cannot be made without additional |
57 | information from the user, Perl decides in favor of compatibility, and |
58 | chooses to use byte semantics. |
59 | |
60 | This behavior preserves compatibility with earlier versions of Perl, |
61 | which allowed byte semantics in Perl operations, but only as long as |
62 | none of the program's inputs are marked as being as source of Unicode |
63 | character data. Such data may come from filehandles, from calls to |
64 | external programs, from information provided by the system (such as %ENV), |
21bad921 |
65 | or from literals and constants in the source text. |
8cbd9a7a |
66 | |
46487f74 |
67 | If the C<-C> command line switch is used, (or the ${^WIDE_SYSTEM_CALLS} |
68 | global flag is set to C<1>), all system calls will use the |
3969a896 |
69 | corresponding wide character APIs. This is currently only implemented |
49cb94c6 |
70 | on Windows as other platforms do not have a unified way of handling |
71 | wide character APIs. |
8cbd9a7a |
72 | |
8058d7ab |
73 | Regardless of the above, the C<bytes> pragma can always be used to force |
74 | byte semantics in a particular lexical scope. See L<bytes>. |
8cbd9a7a |
75 | |
76 | The C<utf8> pragma is primarily a compatibility device that enables |
21bad921 |
77 | recognition of UTF-8 in literals encountered by the parser. It may also |
78 | be used for enabling some of the more experimental Unicode support features. |
8cbd9a7a |
79 | Note that this pragma is only required until a future version of Perl |
80 | in which character semantics will become the default. This pragma may |
81 | then become a no-op. See L<utf8>. |
82 | |
83 | Unless mentioned otherwise, Perl operators will use character semantics |
84 | when they are dealing with Unicode data, and byte semantics otherwise. |
85 | Thus, character semantics for these operations apply transparently; if |
86 | the input data came from a Unicode source (for example, by adding a |
87 | character encoding discipline to the filehandle whence it came, or a |
88 | literal UTF-8 string constant in the program), character semantics |
89 | apply; otherwise, byte semantics are in effect. To force byte semantics |
8058d7ab |
90 | on Unicode data, the C<bytes> pragma should be used. |
393fec97 |
91 | |
92 | Under character semantics, many operations that formerly operated on |
93 | bytes change to operating on characters. For ASCII data this makes |
94 | no difference, because UTF-8 stores ASCII in single bytes, but for |
21bad921 |
95 | any character greater than C<chr(127)>, the character may be stored in |
393fec97 |
96 | a sequence of two or more bytes, all of which have the high bit set. |
97 | But by and large, the user need not worry about this, because Perl |
98 | hides it from the user. A character in Perl is logically just a number |
99 | ranging from 0 to 2**32 or so. Larger characters encode to longer |
100 | sequences of bytes internally, but again, this is just an internal |
101 | detail which is hidden at the Perl level. |
102 | |
8cbd9a7a |
103 | =head2 Effects of character semantics |
393fec97 |
104 | |
105 | Character semantics have the following effects: |
106 | |
107 | =over 4 |
108 | |
109 | =item * |
110 | |
111 | Strings and patterns may contain characters that have an ordinal value |
21bad921 |
112 | larger than 255. |
393fec97 |
113 | |
114 | Presuming you use a Unicode editor to edit your program, such characters |
115 | will typically occur directly within the literal strings as UTF-8 |
116 | characters, but you can also specify a particular character with an |
117 | extension of the C<\x> notation. UTF-8 characters are specified by |
118 | putting the hexadecimal code within curlies after the C<\x>. For instance, |
4375e838 |
119 | a Unicode smiley face is C<\x{263A}>. |
393fec97 |
120 | |
121 | =item * |
122 | |
123 | Identifiers within the Perl script may contain Unicode alphanumeric |
124 | characters, including ideographs. (You are currently on your own when |
125 | it comes to using the canonical forms of characters--Perl doesn't (yet) |
126 | attempt to canonicalize variable names for you.) |
127 | |
393fec97 |
128 | =item * |
129 | |
130 | Regular expressions match characters instead of bytes. For instance, |
131 | "." matches a character instead of a byte. (However, the C<\C> pattern |
49cb94c6 |
132 | is available to force a match a single byte ("C<char>" in C, hence C<\C>).) |
393fec97 |
133 | |
393fec97 |
134 | =item * |
135 | |
136 | Character classes in regular expressions match characters instead of |
137 | bytes, and match against the character properties specified in the |
138 | Unicode properties database. So C<\w> can be used to match an ideograph, |
139 | for instance. |
140 | |
393fec97 |
141 | =item * |
142 | |
143 | Named Unicode properties and block ranges make be used as character |
144 | classes via the new C<\p{}> (matches property) and C<\P{}> (doesn't |
145 | match property) constructs. For instance, C<\p{Lu}> matches any |
146 | character with the Unicode uppercase property, while C<\p{M}> matches |
147 | any mark character. Single letter properties may omit the brackets, so |
148 | that can be written C<\pM> also. Many predefined character classes are |
149 | available, such as C<\p{IsMirrored}> and C<\p{InTibetan}>. |
150 | |
393fec97 |
151 | =item * |
152 | |
153 | The special pattern C<\X> match matches any extended Unicode sequence |
154 | (a "combining character sequence" in Standardese), where the first |
155 | character is a base character and subsequent characters are mark |
156 | characters that apply to the base character. It is equivalent to |
157 | C<(?:\PM\pM*)>. |
158 | |
393fec97 |
159 | =item * |
160 | |
383e7cdd |
161 | The C<tr///> operator translates characters instead of bytes. Note |
162 | that the C<tr///CU> functionality has been removed, as the interface |
163 | was a mistake. For similar functionality see pack('U0', ...) and |
164 | pack('C0', ...). |
393fec97 |
165 | |
393fec97 |
166 | =item * |
167 | |
168 | Case translation operators use the Unicode case translation tables |
169 | when provided character input. Note that C<uc()> translates to |
170 | uppercase, while C<ucfirst> translates to titlecase (for languages |
171 | that make the distinction). Naturally the corresponding backslash |
172 | sequences have the same semantics. |
173 | |
174 | =item * |
175 | |
176 | Most operators that deal with positions or lengths in the string will |
177 | automatically switch to using character positions, including C<chop()>, |
178 | C<substr()>, C<pos()>, C<index()>, C<rindex()>, C<sprintf()>, |
179 | C<write()>, and C<length()>. Operators that specifically don't switch |
180 | include C<vec()>, C<pack()>, and C<unpack()>. Operators that really |
181 | don't care include C<chomp()>, as well as any other operator that |
182 | treats a string as a bucket of bits, such as C<sort()>, and the |
183 | operators dealing with filenames. |
184 | |
185 | =item * |
186 | |
187 | The C<pack()>/C<unpack()> letters "C<c>" and "C<C>" do I<not> change, |
188 | since they're often used for byte-oriented formats. (Again, think |
189 | "C<char>" in the C language.) However, there is a new "C<U>" specifier |
190 | that will convert between UTF-8 characters and integers. (It works |
191 | outside of the utf8 pragma too.) |
192 | |
193 | =item * |
194 | |
195 | The C<chr()> and C<ord()> functions work on characters. This is like |
196 | C<pack("U")> and C<unpack("U")>, not like C<pack("C")> and |
197 | C<unpack("C")>. In fact, the latter are how you now emulate |
198 | byte-oriented C<chr()> and C<ord()> under utf8. |
199 | |
200 | =item * |
201 | |
a1ca4561 |
202 | The bit string operators C<& | ^ ~> can operate on character data. |
203 | However, for backward compatibility reasons (bit string operations |
204 | when the characters all are less than 256 in ordinal value) one cannot |
205 | mix C<~> (the bit complement) and characters both less than 256 and |
206 | equal or greater than 256. Most importantly, the DeMorgan's laws |
207 | (C<~($x|$y) eq ~$x&~$y>, C<~($x&$y) eq ~$x|~$y>) won't hold. |
208 | Another way to look at this is that the complement cannot return |
209 | B<both> the 8-bit (byte) wide bit complement, and the full character |
210 | wide bit complement. |
211 | |
212 | =item * |
213 | |
393fec97 |
214 | And finally, C<scalar reverse()> reverses by character rather than by byte. |
215 | |
216 | =back |
217 | |
8cbd9a7a |
218 | =head2 Character encodings for input and output |
219 | |
49cb94c6 |
220 | This feature is in the process of getting implemented. |
221 | |
222 | (For Perl 5.6 and earlier the support is unlikely to get integrated |
223 | to the core language and some external module will be required.) |
8cbd9a7a |
224 | |
393fec97 |
225 | =head1 CAVEATS |
226 | |
227 | As of yet, there is no method for automatically coercing input and |
228 | output to some encoding other than UTF-8. This is planned in the near |
229 | future, however. |
230 | |
8cbd9a7a |
231 | Whether an arbitrary piece of data will be treated as "characters" or |
232 | "bytes" by internal operations cannot be divined at the current time. |
393fec97 |
233 | |
234 | Use of locales with utf8 may lead to odd results. Currently there is |
235 | some attempt to apply 8-bit locale info to characters in the range |
236 | 0..255, but this is demonstrably incorrect for locales that use |
237 | characters above that range (when mapped into Unicode). It will also |
238 | tend to run slower. Avoidance of locales is strongly encouraged. |
239 | |
240 | =head1 SEE ALSO |
241 | |
8058d7ab |
242 | L<bytes>, L<utf8>, L<perlvar/"${^WIDE_SYSTEM_CALLS}"> |
393fec97 |
243 | |
244 | =cut |