This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Don't use locale definitions unless within scope
[perl5.git] / pod / perllocale.pod
CommitLineData
66cbab2c
KW
1=encoding utf8
2
5f05dabc
PP
3=head1 NAME
4
b0c42ed9 5perllocale - Perl locale handling (internationalization and localization)
5f05dabc
PP
6
7=head1 DESCRIPTION
8
66cbab2c
KW
9In the beginning there was ASCII, the "American Standard Code for
10Information Interchange", which works quite well for Americans with
11their English alphabet and dollar-denominated currency. But it doesn't
12work so well even for other English speakers, who may use different
13currencies, such as the pound sterling (as the symbol for that currency
14is not in ASCII); and it's hopelessly inadequate for many of the
15thousands of the world's other languages.
16
17To address these deficiencies, the concept of locales was invented
18(formally the ISO C, XPG4, POSIX 1.c "locale system"). And applications
19were and are being written that use the locale mechanism. The process of
20making such an application take account of its users' preferences in
21these kinds of matters is called B<internationalization> (often
22abbreviated as B<i18n>); telling such an application about a particular
23set of preferences is known as B<localization> (B<l10n>).
24
39332f68 25Perl has been extended to support the locale system. This
66cbab2c
KW
26is controlled per application by using one pragma, one function call,
27and several environment variables.
28
29Unfortunately, there are quite a few deficiencies with the design (and
30often, the implementations) of locales, and their use for character sets
31has mostly been supplanted by Unicode (see L<perlunitut> for an
32introduction to that, and keep on reading here for how Unicode interacts
33with locales in Perl).
34
7ee2ae1e 35Perl continues to support the old locale system, and starting in v5.16,
66cbab2c
KW
36provides a hybrid way to use the Unicode character set, along with the
37other portions of locales that may not be so problematic.
38(Unicode is also creating C<CLDR>, the "Common Locale Data Repository",
39L<http://cldr.unicode.org/> which includes more types of information than
40are available in the POSIX locale system. At the time of this writing,
41there was no CPAN module that provides access to this XML-encoded data.
42However, many of its locales have the POSIX-only data extracted, and are
43available at L<http://unicode.org/Public/cldr/latest/>.)
44
45=head1 WHAT IS A LOCALE
46
47A locale is a set of data that describes various aspects of how various
48communities in the world categorize their world. These categories are
49broken down into the following types (some of which include a brief
50note here):
51
52=over
53
54=item Category LC_NUMERIC: Numeric formatting
55
56This indicates how numbers should be formatted for human readability,
57for example the character used as the decimal point.
58
59=item Category LC_MONETARY: Formatting of monetary amounts
60
61=for comment
62The nbsp below makes this look better
63
64E<160>
65
66=item Category LC_TIME: Date/Time formatting
67
68=for comment
69The nbsp below makes this look better
70
71E<160>
72
73=item Category LC_MESSAGES: Error and other messages
74
75This for the most part is beyond the scope of Perl
76
77=item Category LC_COLLATE: Collation
78
76073c88 79This indicates the ordering of letters for comparison and sorting.
66cbab2c
KW
80In Latin alphabets, for example, "b", generally follows "a".
81
82=item Category LC_CTYPE: Character Types
83
84This indicates, for example if a character is an uppercase letter.
85
86=back
87
88More details on the categories are given below in L</LOCALE CATEGORIES>.
89
90Together, these categories go a long way towards being able to customize
91a single program to run in many different locations. But there are
92deficiencies, so keep reading.
5f05dabc
PP
93
94=head1 PREPARING TO USE LOCALES
95
66cbab2c
KW
96Perl will not use locales unless specifically requested to (see L</NOTES> below
97for the partial exception of C<write()>). But even if there is such a
98request, B<all> of the following must be true for it to work properly:
5f05dabc
PP
99
100=over 4
101
102=item *
103
104B<Your operating system must support the locale system>. If it does,
39332f68 105you should find that the C<setlocale()> function is a documented part of
5f05dabc
PP
106its C library.
107
108=item *
109
5a964f20 110B<Definitions for locales that you use must be installed>. You, or
14280422
DD
111your system administrator, must make sure that this is the case. The
112available locales, the location in which they are kept, and the manner
5a964f20
TC
113in which they are installed all vary from system to system. Some systems
114provide only a few, hard-wired locales and do not allow more to be
115added. Others allow you to add "canned" locales provided by the system
116supplier. Still others allow you or the system administrator to define
14280422 117and add arbitrary locales. (You may have to ask your supplier to
5a964f20 118provide canned locales that are not delivered with your operating
14280422 119system.) Read your system documentation for further illumination.
5f05dabc
PP
120
121=item *
122
123B<Perl must believe that the locale system is supported>. If it does,
124C<perl -V:d_setlocale> will say that the value for C<d_setlocale> is
125C<define>.
126
127=back
128
129If you want a Perl application to process and present your data
130according to a particular locale, the application code should include
2ae324a7 131the S<C<use locale>> pragma (see L<The use locale pragma>) where
5f05dabc
PP
132appropriate, and B<at least one> of the following must be true:
133
134=over 4
135
c052850d 136=item 1
5f05dabc 137
66cbab2c 138B<The locale-determining environment variables (see L</"ENVIRONMENT">)
5a964f20 139must be correctly set up> at the time the application is started, either
ef3087ec 140by yourself or by whomever set up your system account; or
5f05dabc 141
c052850d 142=item 2
5f05dabc 143
14280422
DD
144B<The application must set its own locale> using the method described in
145L<The setlocale function>.
5f05dabc
PP
146
147=back
148
149=head1 USING LOCALES
150
151=head2 The use locale pragma
152
14280422 153By default, Perl ignores the current locale. The S<C<use locale>>
66cbab2c 154pragma tells Perl to use the current locale for some operations.
7ee2ae1e 155Starting in v5.16, there is an optional parameter to this pragma:
66cbab2c
KW
156
157 use locale ':not_characters';
158
159This parameter allows better mixing of locales and Unicode, and is
160described fully in L</Unicode and UTF-8>, but briefly, it tells Perl to
161not use the character portions of the locale definition, that is
162the C<LC_CTYPE> and C<LC_COLLATE> categories. Instead it will use the
163native (extended by Unicode) character set. When using this parameter,
164you are responsible for getting the external character set translated
165into the native/Unicode one (which it already will be if it is one of
166the increasingly popular UTF-8 locales). There are convenient ways of
167doing this, as described in L</Unicode and UTF-8>.
c052850d
KW
168
169The current locale is set at execution time by
170L<setlocale()|/The setlocale function> described below. If that function
171hasn't yet been called in the course of the program's execution, the
66cbab2c 172current locale is that which was determined by the L</"ENVIRONMENT"> in
c052850d
KW
173effect at the start of the program, except that
174C<L<LC_NUMERIC|/Category LC_NUMERIC: Numeric Formatting>> is always
175initialized to the C locale (mentioned under L<Finding locales>).
70709c68
KW
176If there is no valid environment, the current locale is undefined. It
177is likely, but not necessarily, the "C" locale.
c052850d
KW
178
179The operations that are affected by locale are:
5f05dabc
PP
180
181=over 4
182
66cbab2c
KW
183=item B<Under C<use locale ':not_characters';>>
184
185=over 4
186
187=item *
188
39332f68 189B<Format declarations> (C<format()>) use C<LC_NUMERIC>
66cbab2c
KW
190
191=item *
192
39332f68 193B<The POSIX date formatting function> (C<strftime()>) uses C<LC_TIME>.
66cbab2c
KW
194
195=back
196
197=for comment
198The nbsp below makes this look better
199
200E<160>
201
202=item B<Under just plain C<use locale;>>
203
204The above operations are affected, as well as the following:
205
206=over 4
207
5f05dabc
PP
208=item *
209
14280422 210B<The comparison operators> (C<lt>, C<le>, C<cmp>, C<ge>, and C<gt>) and
39332f68
KW
211the POSIX string collation functions C<strcoll()> and C<strxfrm()> use
212C<LC_COLLATE>. C<sort()> is also affected if used without an
5a964f20 213explicit comparison function, because it uses C<cmp> by default.
14280422 214
5a964f20 215B<Note:> C<eq> and C<ne> are unaffected by locale: they always
de108802 216perform a char-by-char comparison of their scalar operands. What's
14280422
DD
217more, if C<cmp> finds that its operands are equal according to the
218collation sequence specified by the current locale, it goes on to
de108802
RGS
219perform a char-by-char comparison, and only returns I<0> (equal) if the
220operands are char-for-char identical. If you really want to know whether
5a964f20 221two strings--which C<eq> and C<cmp> may consider different--are equal
14280422
DD
222as far as collation in the locale is concerned, see the discussion in
223L<Category LC_COLLATE: Collation>.
5f05dabc
PP
224
225=item *
226
39332f68
KW
227B<Regular expressions and case-modification functions> (C<uc()>, C<lc()>,
228C<ucfirst()>, and C<lcfirst()>) use C<LC_CTYPE>
5f05dabc 229
5f05dabc
PP
230=back
231
66cbab2c 232=back
5f05dabc 233
5a964f20 234The default behavior is restored with the S<C<no locale>> pragma, or
ef3087ec 235upon reaching the end of the block enclosing C<use locale>.
66cbab2c
KW
236Note that C<use locale> and C<use locale ':not_characters'> may be
237nested, and that what is in effect within an inner scope will revert to
238the outer scope's rules at the end of the inner scope.
5f05dabc 239
5a964f20 240The string result of any operation that uses locale
14280422
DD
241information is tainted, as it is possible for a locale to be
242untrustworthy. See L<"SECURITY">.
5f05dabc
PP
243
244=head2 The setlocale function
245
14280422 246You can switch locales as often as you wish at run time with the
39332f68 247C<POSIX::setlocale()> function:
5f05dabc 248
5f05dabc
PP
249 # Import locale-handling tool set from POSIX module.
250 # This example uses: setlocale -- the function call
251 # LC_CTYPE -- explained below
252 use POSIX qw(locale_h);
253
14280422 254 # query and save the old locale
5f05dabc
PP
255 $old_locale = setlocale(LC_CTYPE);
256
257 setlocale(LC_CTYPE, "fr_CA.ISO8859-1");
258 # LC_CTYPE now in locale "French, Canada, codeset ISO 8859-1"
259
260 setlocale(LC_CTYPE, "");
261 # LC_CTYPE now reset to default defined by LC_ALL/LC_CTYPE/LANG
262 # environment variables. See below for documentation.
263
264 # restore the old locale
265 setlocale(LC_CTYPE, $old_locale);
266
39332f68 267The first argument of C<setlocale()> gives the B<category>, the second the
14280422
DD
268B<locale>. The category tells in what aspect of data processing you
269want to apply locale-specific rules. Category names are discussed in
66cbab2c 270L</LOCALE CATEGORIES> and L</"ENVIRONMENT">. The locale is the name of a
14280422
DD
271collection of customization information corresponding to a particular
272combination of language, country or territory, and codeset. Read on for
273hints on the naming of locales: not all systems name locales as in the
274example.
275
39332f68 276If no second argument is provided and the category is something other
502a173a
JH
277than LC_ALL, the function returns a string naming the current locale
278for the category. You can use this value as the second argument in a
39332f68 279subsequent call to C<setlocale()>.
502a173a
JH
280
281If no second argument is provided and the category is LC_ALL, the
282result is implementation-dependent. It may be a string of
c052850d 283concatenated locale names (separator also implementation-dependent)
39332f68 284or a single locale name. Please consult your L<setlocale(3)> man page for
502a173a
JH
285details.
286
287If a second argument is given and it corresponds to a valid locale,
288the locale for the category is set to that value, and the function
289returns the now-current locale value. You can then use this in yet
39332f68 290another call to C<setlocale()>. (In some implementations, the return
502a173a
JH
291value may sometimes differ from the value you gave as the second
292argument--think of it as an alias for the value you gave.)
5f05dabc
PP
293
294As the example shows, if the second argument is an empty string, the
295category's locale is returned to the default specified by the
296corresponding environment variables. Generally, this results in a
5a964f20 297return to the default that was in force when Perl started up: changes
54310121 298to the environment made by the application after startup may or may not
5a964f20 299be noticed, depending on your system's C library.
5f05dabc 300
14280422 301If the second argument does not correspond to a valid locale, the locale
39332f68 302for the category is not changed, and the function returns C<undef>.
5f05dabc 303
66cbab2c
KW
304Note that Perl ignores the current C<LC_CTYPE> and C<LC_COLLATE> locales
305within the scope of a C<use locale ':not_characters'>.
306
39332f68 307For further information about the categories, consult L<setlocale(3)>.
3e6e419a
JH
308
309=head2 Finding locales
310
39332f68 311For locales available in your system, consult also L<setlocale(3)> to
5a964f20
TC
312see whether it leads to the list of available locales (search for the
313I<SEE ALSO> section). If that fails, try the following command lines:
5f05dabc
PP
314
315 locale -a
316
317 nlsinfo
318
319 ls /usr/lib/nls/loc
320
321 ls /usr/lib/locale
322
323 ls /usr/lib/nls
324
b478f28d
JH
325 ls /usr/share/locale
326
5f05dabc
PP
327and see whether they list something resembling these
328
2bdf8add 329 en_US.ISO8859-1 de_DE.ISO8859-1 ru_RU.ISO8859-5
502a173a 330 en_US.iso88591 de_DE.iso88591 ru_RU.iso88595
2bdf8add 331 en_US de_DE ru_RU
14280422 332 en de ru
2bdf8add
JH
333 english german russian
334 english.iso88591 german.iso88591 russian.iso88595
502a173a 335 english.roman8 russian.koi8r
5f05dabc 336
39332f68 337Sadly, even though the calling interface for C<setlocale()> has been
528d65ad 338standardized, names of locales and the directories where the
5a964f20 339configuration resides have not been. The basic form of the name is
528d65ad
JH
340I<language_territory>B<.>I<codeset>, but the latter parts after
341I<language> are not always present. The I<language> and I<country>
342are usually from the standards B<ISO 3166> and B<ISO 639>, the
343two-letter abbreviations for the countries and the languages of the
344world, respectively. The I<codeset> part often mentions some B<ISO
3458859> character set, the Latin codesets. For example, C<ISO 8859-1>
346is the so-called "Western European codeset" that can be used to encode
347most Western European languages adequately. Again, there are several
348ways to write even the name of that one standard. Lamentably.
5f05dabc 349
14280422
DD
350Two special locales are worth particular mention: "C" and "POSIX".
351Currently these are effectively the same locale: the difference is
5a964f20
TC
352mainly that the first one is defined by the C standard, the second by
353the POSIX standard. They define the B<default locale> in which
14280422 354every program starts in the absence of locale information in its
5a964f20 355environment. (The I<default> default locale, if you will.) Its language
39332f68
KW
356is (American) English and its character codeset ASCII or, rarely, a
357superset thereof (such as the "DEC Multinational Character Set
358(DEC-MCS)"). B<Warning>. The C locale delivered by some vendors
359may not actually exactly match what the C standard calls for. So
360beware.
5f05dabc 361
14280422
DD
362B<NOTE>: Not all systems have the "POSIX" locale (not all systems are
363POSIX-conformant), so use "C" when you need explicitly to specify this
364default locale.
5f05dabc 365
3e6e419a
JH
366=head2 LOCALE PROBLEMS
367
5a964f20 368You may encounter the following warning message at Perl startup:
3e6e419a
JH
369
370 perl: warning: Setting locale failed.
371 perl: warning: Please check that your locale settings:
372 LC_ALL = "En_US",
373 LANG = (unset)
374 are supported and installed on your system.
375 perl: warning: Falling back to the standard locale ("C").
376
5a964f20
TC
377This means that your locale settings had LC_ALL set to "En_US" and
378LANG exists but has no value. Perl tried to believe you but could not.
379Instead, Perl gave up and fell back to the "C" locale, the default locale
380that is supposed to work no matter what. This usually means your locale
381settings were wrong, they mention locales your system has never heard
382of, or the locale installation in your system has problems (for example,
383some system files are broken or missing). There are quick and temporary
384fixes to these problems, as well as more thorough and lasting fixes.
3e6e419a
JH
385
386=head2 Temporarily fixing locale problems
387
5a964f20 388The two quickest fixes are either to render Perl silent about any
3e6e419a
JH
389locale inconsistencies or to run Perl under the default locale "C".
390
391Perl's moaning about locale problems can be silenced by setting the
900bd440
JH
392environment variable PERL_BADLANG to a zero value, for example "0".
393This method really just sweeps the problem under the carpet: you tell
394Perl to shut up even when Perl sees that something is wrong. Do not
395be surprised if later something locale-dependent misbehaves.
3e6e419a
JH
396
397Perl can be run under the "C" locale by setting the environment
5a964f20
TC
398variable LC_ALL to "C". This method is perhaps a bit more civilized
399than the PERL_BADLANG approach, but setting LC_ALL (or
400other locale variables) may affect other programs as well, not just
401Perl. In particular, external programs run from within Perl will see
3e6e419a 402these changes. If you make the new settings permanent (read on), all
f979aebc 403programs you run see the changes. See L<"ENVIRONMENT"> for
5a964f20 404the full list of relevant environment variables and L<USING LOCALES>
e05ffc7d 405for their effects in Perl. Effects in other programs are
5a964f20 406easily deducible. For example, the variable LC_COLLATE may well affect
b432a672 407your B<sort> program (or whatever the program that arranges "records"
3e6e419a
JH
408alphabetically in your system is called).
409
5a964f20
TC
410You can test out changing these variables temporarily, and if the
411new settings seem to help, put those settings into your shell startup
412files. Consult your local documentation for the exact details. For in
413Bourne-like shells (B<sh>, B<ksh>, B<bash>, B<zsh>):
3e6e419a
JH
414
415 LC_ALL=en_US.ISO8859-1
416 export LC_ALL
417
5a964f20
TC
418This assumes that we saw the locale "en_US.ISO8859-1" using the commands
419discussed above. We decided to try that instead of the above faulty
420locale "En_US"--and in Cshish shells (B<csh>, B<tcsh>)
3e6e419a
JH
421
422 setenv LC_ALL en_US.ISO8859-1
c47ff5f1 423
c406981e
JH
424or if you have the "env" application you can do in any shell
425
426 env LC_ALL=en_US.ISO8859-1 perl ...
427
5a964f20 428If you do not know what shell you have, consult your local
3e6e419a
JH
429helpdesk or the equivalent.
430
431=head2 Permanently fixing locale problems
432
5a964f20
TC
433The slower but superior fixes are when you may be able to yourself
434fix the misconfiguration of your own environment variables. The
3e6e419a
JH
435mis(sing)configuration of the whole system's locales usually requires
436the help of your friendly system administrator.
437
5a964f20
TC
438First, see earlier in this document about L<Finding locales>. That tells
439how to find which locales are really supported--and more importantly,
440installed--on your system. In our example error message, environment
441variables affecting the locale are listed in the order of decreasing
442importance (and unset variables do not matter). Therefore, having
443LC_ALL set to "En_US" must have been the bad choice, as shown by the
444error message. First try fixing locale settings listed first.
3e6e419a 445
5a964f20
TC
446Second, if using the listed commands you see something B<exactly>
447(prefix matches do not count and case usually counts) like "En_US"
448without the quotes, then you should be okay because you are using a
449locale name that should be installed and available in your system.
4a4eefd0 450In this case, see L<Permanently fixing your system's locale configuration>.
3e6e419a 451
4a4eefd0 452=head2 Permanently fixing your system's locale configuration
3e6e419a 453
5a964f20 454This is when you see something like:
3e6e419a
JH
455
456 perl: warning: Please check that your locale settings:
457 LC_ALL = "En_US",
458 LANG = (unset)
459 are supported and installed on your system.
460
461but then cannot see that "En_US" listed by the above-mentioned
5a964f20
TC
462commands. You may see things like "en_US.ISO8859-1", but that isn't
463the same. In this case, try running under a locale
464that you can list and which somehow matches what you tried. The
3e6e419a 465rules for matching locale names are a bit vague because
e05ffc7d 466standardization is weak in this area. See again the
13a2d996 467L<Finding locales> about general rules.
3e6e419a 468
b687b08b 469=head2 Fixing system locale configuration
3e6e419a 470
5a964f20
TC
471Contact a system administrator (preferably your own) and report the exact
472error message you get, and ask them to read this same documentation you
473are now reading. They should be able to check whether there is something
474wrong with the locale configuration of the system. The L<Finding locales>
475section is unfortunately a bit vague about the exact commands and places
476because these things are not that standardized.
3e6e419a 477
5f05dabc
PP
478=head2 The localeconv function
479
39332f68 480The C<POSIX::localeconv()> function allows you to get particulars of the
14280422
DD
481locale-dependent numeric formatting information specified by the current
482C<LC_NUMERIC> and C<LC_MONETARY> locales. (If you just want the name of
39332f68 483the current locale for a particular category, use C<POSIX::setlocale()>
5a964f20 484with a single parameter--see L<The setlocale function>.)
5f05dabc
PP
485
486 use POSIX qw(locale_h);
5f05dabc
PP
487
488 # Get a reference to a hash of locale-dependent info
489 $locale_values = localeconv();
490
491 # Output sorted list of the values
492 for (sort keys %$locale_values) {
14280422 493 printf "%-20s = %s\n", $_, $locale_values->{$_}
5f05dabc
PP
494 }
495
39332f68 496C<localeconv()> takes no arguments, and returns B<a reference to> a hash.
5a964f20 497The keys of this hash are variable names for formatting, such as
502a173a 498C<decimal_point> and C<thousands_sep>. The values are the
cea6626f 499corresponding, er, values. See L<POSIX/localeconv> for a longer
502a173a
JH
500example listing the categories an implementation might be expected to
501provide; some provide more and others fewer. You don't need an
39332f68 502explicit C<use locale>, because C<localeconv()> always observes the
502a173a 503current locale.
5f05dabc 504
5a964f20
TC
505Here's a simple-minded example program that rewrites its command-line
506parameters as integers correctly formatted in the current locale:
5f05dabc 507
ef3087ec
KW
508 use POSIX qw(locale_h);
509
510 # Get some of locale's numeric formatting parameters
511 my ($thousands_sep, $grouping) =
512 @{localeconv()}{'thousands_sep', 'grouping'};
513
514 # Apply defaults if values are missing
515 $thousands_sep = ',' unless $thousands_sep;
516
517 # grouping and mon_grouping are packed lists
518 # of small integers (characters) telling the
519 # grouping (thousand_seps and mon_thousand_seps
520 # being the group dividers) of numbers and
521 # monetary quantities. The integers' meanings:
522 # 255 means no more grouping, 0 means repeat
523 # the previous grouping, 1-254 means use that
524 # as the current grouping. Grouping goes from
525 # right to left (low to high digits). In the
526 # below we cheat slightly by never using anything
527 # else than the first grouping (whatever that is).
528 if ($grouping) {
529 @grouping = unpack("C*", $grouping);
530 } else {
531 @grouping = (3);
532 }
533
534 # Format command line params for current locale
535 for (@ARGV) {
536 $_ = int; # Chop non-integer part
537 1 while
538 s/(\d)(\d{$grouping[0]}($|$thousands_sep))/$1$thousands_sep$2/;
539 print "$_";
540 }
541 print "\n";
5f05dabc 542
74c76037 543=head2 I18N::Langinfo
4bbcc6e8
JH
544
545Another interface for querying locale-dependent information is the
39332f68 546C<I18N::Langinfo::langinfo()> function, available at least in Unix-like
4bbcc6e8
JH
547systems and VMS.
548
39332f68
KW
549The following example will import the C<langinfo()> function itself and
550three constants to be used as arguments to C<langinfo()>: a constant for
74c76037
JH
551the abbreviated first day of the week (the numbering starts from
552Sunday = 1) and two more constants for the affirmative and negative
553answers for a yes/no question in the current locale.
4bbcc6e8 554
74c76037 555 use I18N::Langinfo qw(langinfo ABDAY_1 YESSTR NOSTR);
4bbcc6e8 556
ef3087ec
KW
557 my ($abday_1, $yesstr, $nostr)
558 = map { langinfo } qw(ABDAY_1 YESSTR NOSTR);
4bbcc6e8 559
74c76037 560 print "$abday_1? [$yesstr/$nostr] ";
4bbcc6e8 561
74c76037
JH
562In other words, in the "C" (or English) locale the above will probably
563print something like:
564
e05ffc7d 565 Sun? [yes/no]
4bbcc6e8
JH
566
567See L<I18N::Langinfo> for more information.
568
5f05dabc
PP
569=head1 LOCALE CATEGORIES
570
5a964f20
TC
571The following subsections describe basic locale categories. Beyond these,
572some combination categories allow manipulation of more than one
573basic category at a time. See L<"ENVIRONMENT"> for a discussion of these.
5f05dabc
PP
574
575=head2 Category LC_COLLATE: Collation
576
66cbab2c
KW
577In the scope of S<C<use locale>> (but not a
578C<use locale ':not_characters'>), Perl looks to the C<LC_COLLATE>
5a964f20 579environment variable to determine the application's notions on collation
b4ffc3db
TC
580(ordering) of characters. For example, "b" follows "a" in Latin
581alphabets, but where do "E<aacute>" and "E<aring>" belong? And while
f87fa335 582"color" follows "chocolate" in English, what about in traditional Spanish?
5f05dabc 583
60f0fa02
JH
584The following collations all make sense and you may meet any of them
585if you "use locale".
586
587 A B C D E a b c d e
35316ca3 588 A a B b C c D d E e
60f0fa02
JH
589 a A b B c C d D e E
590 a b c d e A B C D E
591
f1cbbd6e 592Here is a code snippet to tell what "word"
5a964f20 593characters are in the current locale, in that locale's order:
5f05dabc
PP
594
595 use locale;
35316ca3 596 print +(sort grep /\w/, map { chr } 0..255), "\n";
5f05dabc 597
14280422
DD
598Compare this with the characters that you see and their order if you
599state explicitly that the locale should be ignored:
5f05dabc
PP
600
601 no locale;
35316ca3 602 print +(sort grep /\w/, map { chr } 0..255), "\n";
5f05dabc
PP
603
604This machine-native collation (which is what you get unless S<C<use
605locale>> has appeared earlier in the same block) must be used for
606sorting raw binary data, whereas the locale-dependent collation of the
b0c42ed9 607first example is useful for natural text.
5f05dabc 608
14280422
DD
609As noted in L<USING LOCALES>, C<cmp> compares according to the current
610collation locale when C<use locale> is in effect, but falls back to a
de108802 611char-by-char comparison for strings that the locale says are equal. You
39332f68 612can use C<POSIX::strcoll()> if you don't want this fall-back:
14280422
DD
613
614 use POSIX qw(strcoll);
615 $equal_in_locale =
616 !strcoll("space and case ignored", "SpaceAndCaseIgnored");
617
39332f68 618C<$equal_in_locale> will be true if the collation locale specifies a
5a964f20 619dictionary-like ordering that ignores space characters completely and
9e3a2af8 620which folds case.
14280422 621
5a964f20 622If you have a single string that you want to check for "equality in
14280422 623locale" against several others, you might think you could gain a little
39332f68 624efficiency by using C<POSIX::strxfrm()> in conjunction with C<eq>:
14280422
DD
625
626 use POSIX qw(strxfrm);
627 $xfrm_string = strxfrm("Mixed-case string");
628 print "locale collation ignores spaces\n"
629 if $xfrm_string eq strxfrm("Mixed-casestring");
630 print "locale collation ignores hyphens\n"
631 if $xfrm_string eq strxfrm("Mixedcase string");
632 print "locale collation ignores case\n"
633 if $xfrm_string eq strxfrm("mixed-case string");
634
39332f68 635C<strxfrm()> takes a string and maps it into a transformed string for use
de108802 636in char-by-char comparisons against other transformed strings during
14280422 637collation. "Under the hood", locale-affected Perl comparison operators
39332f68
KW
638call C<strxfrm()> for both operands, then do a char-by-char
639comparison of the transformed strings. By calling C<strxfrm()> explicitly
14280422 640and using a non locale-affected comparison, the example attempts to save
5a964f20 641a couple of transformations. But in fact, it doesn't save anything: Perl
2ae324a7 642magic (see L<perlguts/Magic Variables>) creates the transformed version of a
5a964f20 643string the first time it's needed in a comparison, then keeps this version around
14280422 644in case it's needed again. An example rewritten the easy way with
e38874e2 645C<cmp> runs just about as fast. It also copes with null characters
39332f68 646embedded in strings; if you call C<strxfrm()> directly, it treats the first
5a964f20
TC
647null it finds as a terminator. don't expect the transformed strings
648it produces to be portable across systems--or even from one revision
39332f68 649of your operating system to the next. In short, don't call C<strxfrm()>
e38874e2 650directly: let Perl do it for you.
14280422 651
5a964f20 652Note: C<use locale> isn't shown in some of these examples because it isn't
39332f68 653needed: C<strcoll()> and C<strxfrm()> exist only to generate locale-dependent
14280422 654results, and so always obey the current C<LC_COLLATE> locale.
5f05dabc
PP
655
656=head2 Category LC_CTYPE: Character Types
657
66cbab2c
KW
658In the scope of S<C<use locale>> (but not a
659C<use locale ':not_characters'>), Perl obeys the C<LC_CTYPE> locale
14280422
DD
660setting. This controls the application's notion of which characters are
661alphabetic. This affects Perl's C<\w> regular expression metanotation,
f1cbbd6e
GS
662which stands for alphanumeric characters--that is, alphabetic,
663numeric, and including other special characters such as the underscore or
664hyphen. (Consult L<perlre> for more information about
14280422 665regular expressions.) Thanks to C<LC_CTYPE>, depending on your locale
b4ffc3db
TC
666setting, characters like "E<aelig>", "E<eth>", "E<szlig>", and
667"E<oslash>" may be understood as C<\w> characters.
5f05dabc 668
2c268ad5 669The C<LC_CTYPE> locale also provides the map used in transliterating
68dc0745 670characters between lower and uppercase. This affects the case-mapping
39332f68 671functions--C<fc()>, C<lc()>, C<lcfirst()>, C<uc()>, and C<ucfirst()>; case-mapping
b9cc4f69
KW
672interpolation with C<\F>, C<\l>, C<\L>, C<\u>, or C<\U> in double-quoted
673strings and C<s///> substitutions; and case-independent regular expression
e38874e2
DD
674pattern matching using the C<i> modifier.
675
5a964f20 676Finally, C<LC_CTYPE> affects the POSIX character-class test
39332f68 677functions--C<isalpha()>, C<islower()>, and so on. For example, if you move
5a964f20 678from the "C" locale to a 7-bit Scandinavian one, you may find--possibly
39332f68 679to your surprise--that "|" moves from the C<ispunct()> class to C<isalpha()>.
ef3087ec
KW
680Unfortunately, this creates big problems for regular expressions. "|" still
681means alternation even though it matches C<\w>.
5f05dabc 682
5d63e270
KW
683Note that there are quite a few things that are unaffected by the
684current locale. All the escape sequences for particular characters,
685C<\n> for example, always mean the platform's native one. This means,
686for example, that C<\N> in regular expressions (every character
687but new-line) work on the platform character set.
688
14280422
DD
689B<Note:> A broken or malicious C<LC_CTYPE> locale definition may result
690in clearly ineligible characters being considered to be alphanumeric by
e199995e 691your application. For strict matching of (mundane) ASCII letters and
5a964f20 692digits--for example, in command strings--locale-aware applications
e199995e 693should use C<\w> with the C</a> regular expression modifier. See L<"SECURITY">.
5f05dabc
PP
694
695=head2 Category LC_NUMERIC: Numeric Formatting
696
39332f68 697After a proper C<POSIX::setlocale()> call, Perl obeys the C<LC_NUMERIC>
2095dafa 698locale information, which controls an application's idea of how numbers
39332f68
KW
699should be formatted for human readability by the C<printf()>, C<sprintf()>, and
700C<write()> functions. String-to-numeric conversion by the C<POSIX::strtod()>
5a964f20 701function is also affected. In most implementations the only effect is to
b4ffc3db 702change the character used for the decimal point--perhaps from "." to ",".
5a964f20 703These functions aren't aware of such niceties as thousands separation and
2095dafa 704so on. (See L<The localeconv function> if you care about these things.)
5a964f20 705
3cf03d68 706Output produced by print() is also affected by the current locale: it
3cf03d68
JH
707corresponds to what you'd get from printf() in the "C" locale. The
708same is true for Perl's internal conversions between numeric and
709string formats:
5f05dabc 710
2095dafa 711 use POSIX qw(strtod setlocale LC_NUMERIC);
68e8f474 712 use locale;
2095dafa
RGS
713
714 setlocale LC_NUMERIC, "";
14280422 715
5f05dabc
PP
716 $n = 5/2; # Assign numeric 2.5 to $n
717
35316ca3 718 $a = " $n"; # Locale-dependent conversion to string
5f05dabc 719
35316ca3 720 print "half five is $n\n"; # Locale-dependent output
5f05dabc
PP
721
722 printf "half five is %g\n", $n; # Locale-dependent output
723
14280422
DD
724 print "DECIMAL POINT IS COMMA\n"
725 if $n == (strtod("2,5"))[0]; # Locale-dependent conversion
5f05dabc 726
4bbcc6e8
JH
727See also L<I18N::Langinfo> and C<RADIXCHAR>.
728
5f05dabc
PP
729=head2 Category LC_MONETARY: Formatting of monetary amounts
730
e199995e 731The C standard defines the C<LC_MONETARY> category, but not a function
5a964f20 732that is affected by its contents. (Those with experience of standards
b0c42ed9 733committees will recognize that the working group decided to punt on the
14280422 734issue.) Consequently, Perl takes no notice of it. If you really want
e05ffc7d
KW
735to use C<LC_MONETARY>, you can query its contents--see
736L<The localeconv function>--and use the information that it returns in your
737application's own formatting of currency amounts. However, you may well
738find that the information, voluminous and complex though it may be, still
739does not quite meet your requirements: currency formatting is a hard nut
13a2d996 740to crack.
5f05dabc 741
4bbcc6e8
JH
742See also L<I18N::Langinfo> and C<CRNCYSTR>.
743
5f05dabc
PP
744=head2 LC_TIME
745
39332f68 746Output produced by C<POSIX::strftime()>, which builds a formatted
5f05dabc
PP
747human-readable date/time string, is affected by the current C<LC_TIME>
748locale. Thus, in a French locale, the output produced by the C<%B>
749format element (full month name) for the first month of the year would
5a964f20 750be "janvier". Here's how to get a list of long month names in the
5f05dabc
PP
751current locale:
752
753 use POSIX qw(strftime);
14280422
DD
754 for (0..11) {
755 $long_month_name[$_] =
756 strftime("%B", 0, 0, 0, 1, $_, 96);
5f05dabc
PP
757 }
758
5a964f20 759Note: C<use locale> isn't needed in this example: as a function that
39332f68 760exists only to generate locale-dependent results, C<strftime()> always
14280422 761obeys the current C<LC_TIME> locale.
5f05dabc 762
4bbcc6e8 763See also L<I18N::Langinfo> and C<ABDAY_1>..C<ABDAY_7>, C<DAY_1>..C<DAY_7>,
2a2bf5f4 764C<ABMON_1>..C<ABMON_12>, and C<ABMON_1>..C<ABMON_12>.
4bbcc6e8 765
5f05dabc
PP
766=head2 Other categories
767
5a964f20
TC
768The remaining locale category, C<LC_MESSAGES> (possibly supplemented
769by others in particular implementations) is not currently used by
98a6f11e 770Perl--except possibly to affect the behavior of library functions
771called by extensions outside the standard Perl distribution and by the
772operating system and its utilities. Note especially that the string
773value of C<$!> and the error messages given by external utilities may
774be changed by C<LC_MESSAGES>. If you want to have portable error
265f5c4a 775codes, use C<%!>. See L<Errno>.
14280422
DD
776
777=head1 SECURITY
778
5a964f20 779Although the main discussion of Perl security issues can be found in
14280422
DD
780L<perlsec>, a discussion of Perl's locale handling would be incomplete
781if it did not draw your attention to locale-dependent security issues.
5a964f20
TC
782Locales--particularly on systems that allow unprivileged users to
783build their own locales--are untrustworthy. A malicious (or just plain
14280422
DD
784broken) locale can make a locale-aware application give unexpected
785results. Here are a few possibilities:
786
787=over 4
788
789=item *
790
791Regular expression checks for safe file names or mail addresses using
5a964f20 792C<\w> may be spoofed by an C<LC_CTYPE> locale that claims that
14280422
DD
793characters such as "E<gt>" and "|" are alphanumeric.
794
795=item *
796
e38874e2
DD
797String interpolation with case-mapping, as in, say, C<$dest =
798"C:\U$name.$ext">, may produce dangerous results if a bogus LC_CTYPE
799case-mapping table is in effect.
800
801=item *
802
14280422
DD
803A sneaky C<LC_COLLATE> locale could result in the names of students with
804"D" grades appearing ahead of those with "A"s.
805
806=item *
807
5a964f20 808An application that takes the trouble to use information in
14280422 809C<LC_MONETARY> may format debits as if they were credits and vice versa
5a964f20 810if that locale has been subverted. Or it might make payments in US
14280422
DD
811dollars instead of Hong Kong dollars.
812
813=item *
814
39332f68 815The date and day names in dates formatted by C<strftime()> could be
14280422 816manipulated to advantage by a malicious user able to subvert the
5a964f20 817C<LC_DATE> locale. ("Look--it says I wasn't in the building on
14280422
DD
818Sunday.")
819
820=back
821
822Such dangers are not peculiar to the locale system: any aspect of an
5a964f20 823application's environment which may be modified maliciously presents
14280422 824similar challenges. Similarly, they are not specific to Perl: any
5a964f20 825programming language that allows you to write programs that take
14280422
DD
826account of their environment exposes you to these issues.
827
5a964f20
TC
828Perl cannot protect you from all possibilities shown in the
829examples--there is no substitute for your own vigilance--but, when
14280422 830C<use locale> is in effect, Perl uses the tainting mechanism (see
5a964f20 831L<perlsec>) to mark string results that become locale-dependent, and
14280422 832which may be untrustworthy in consequence. Here is a summary of the
5a964f20 833tainting behavior of operators and functions that may be affected by
14280422
DD
834the locale:
835
836=over 4
837
551e1d92
RB
838=item *
839
840B<Comparison operators> (C<lt>, C<le>, C<ge>, C<gt> and C<cmp>):
14280422
DD
841
842Scalar true/false (or less/equal/greater) result is never tainted.
843
551e1d92
RB
844=item *
845
846B<Case-mapping interpolation> (with C<\l>, C<\L>, C<\u> or C<\U>)
e38874e2
DD
847
848Result string containing interpolated material is tainted if
66cbab2c 849C<use locale> (but not S<C<use locale ':not_characters'>>) is in effect.
e38874e2 850
551e1d92
RB
851=item *
852
853B<Matching operator> (C<m//>):
14280422
DD
854
855Scalar true/false result never tainted.
856
5a964f20 857Subpatterns, either delivered as a list-context result or as $1 etc.
66cbab2c
KW
858are tainted if C<use locale> (but not S<C<use locale ':not_characters'>>)
859is in effect, and the subpattern regular
e38874e2 860expression contains C<\w> (to match an alphanumeric character), C<\W>
6b0ac556
OK
861(non-alphanumeric character), C<\s> (whitespace character), or C<\S>
862(non whitespace character). The matched-pattern variable, $&, $`
e38874e2
DD
863(pre-match), $' (post-match), and $+ (last match) are also tainted if
864C<use locale> is in effect and the regular expression contains C<\w>,
865C<\W>, C<\s>, or C<\S>.
14280422 866
551e1d92
RB
867=item *
868
869B<Substitution operator> (C<s///>):
14280422 870
e38874e2 871Has the same behavior as the match operator. Also, the left
66cbab2c
KW
872operand of C<=~> becomes tainted when C<use locale>
873(but not S<C<use locale ':not_characters'>>) is in effect if modified as
874a result of a substitution based on a regular
e38874e2 875expression match involving C<\w>, C<\W>, C<\s>, or C<\S>; or of
7b8d334a 876case-mapping with C<\l>, C<\L>,C<\u> or C<\U>.
14280422 877
551e1d92
RB
878=item *
879
39332f68 880B<Output formatting functions> (C<printf()> and C<write()>):
14280422 881
3cf03d68
JH
882Results are never tainted because otherwise even output from print,
883for example C<print(1/7)>, should be tainted if C<use locale> is in
884effect.
14280422 885
551e1d92
RB
886=item *
887
39332f68 888B<Case-mapping functions> (C<lc()>, C<lcfirst()>, C<uc()>, C<ucfirst()>):
14280422 889
66cbab2c
KW
890Results are tainted if C<use locale> (but not
891S<C<use locale ':not_characters'>>) is in effect.
14280422 892
551e1d92
RB
893=item *
894
39332f68
KW
895B<POSIX locale-dependent functions> (C<localeconv()>, C<strcoll()>,
896C<strftime()>, C<strxfrm()>):
14280422
DD
897
898Results are never tainted.
899
551e1d92
RB
900=item *
901
39332f68
KW
902B<POSIX character class tests> (C<isalnum()>, C<isalpha()>, C<isdigit()>,
903C<isgraph()>, C<islower()>, C<isprint()>, C<ispunct()>, C<isspace()>, C<isupper()>,
904C<isxdigit()>):
14280422
DD
905
906True/false results are never tainted.
907
908=back
909
910Three examples illustrate locale-dependent tainting.
911The first program, which ignores its locale, won't run: a value taken
54310121 912directly from the command line may not be used to name an output file
14280422
DD
913when taint checks are enabled.
914
915 #/usr/local/bin/perl -T
916 # Run with taint checking
917
54310121 918 # Command line sanity check omitted...
14280422
DD
919 $tainted_output_file = shift;
920
921 open(F, ">$tainted_output_file")
3183d96c 922 or warn "Open of $tainted_output_file failed: $!\n";
14280422
DD
923
924The program can be made to run by "laundering" the tainted value through
5a964f20
TC
925a regular expression: the second example--which still ignores locale
926information--runs, creating the file named on its command line
14280422
DD
927if it can.
928
929 #/usr/local/bin/perl -T
930
931 $tainted_output_file = shift;
932 $tainted_output_file =~ m%[\w/]+%;
933 $untainted_output_file = $&;
934
935 open(F, ">$untainted_output_file")
936 or warn "Open of $untainted_output_file failed: $!\n";
937
5a964f20 938Compare this with a similar but locale-aware program:
14280422
DD
939
940 #/usr/local/bin/perl -T
941
942 $tainted_output_file = shift;
943 use locale;
944 $tainted_output_file =~ m%[\w/]+%;
945 $localized_output_file = $&;
946
947 open(F, ">$localized_output_file")
948 or warn "Open of $localized_output_file failed: $!\n";
949
950This third program fails to run because $& is tainted: it is the result
5a964f20 951of a match involving C<\w> while C<use locale> is in effect.
5f05dabc
PP
952
953=head1 ENVIRONMENT
954
955=over 12
956
957=item PERL_BADLANG
958
14280422 959A string that can suppress Perl's warning about failed locale settings
54310121 960at startup. Failure can occur if the locale support in the operating
5a964f20 961system is lacking (broken) in some way--or if you mistyped the name of
900bd440
JH
962a locale when you set up your environment. If this environment
963variable is absent, or has a value that does not evaluate to integer
964zero--that is, "0" or ""-- Perl will complain about locale setting
965failures.
5f05dabc 966
14280422
DD
967B<NOTE>: PERL_BADLANG only gives you a way to hide the warning message.
968The message tells about some problem in your system's locale support,
969and you should investigate what the problem is.
5f05dabc
PP
970
971=back
972
973The following environment variables are not specific to Perl: They are
39332f68 974part of the standardized (ISO C, XPG4, POSIX 1.c) C<setlocale()> method
14280422 975for controlling an application's opinion on data.
5f05dabc
PP
976
977=over 12
978
979=item LC_ALL
980
5a964f20 981C<LC_ALL> is the "override-all" locale environment variable. If
5f05dabc
PP
982set, it overrides all the rest of the locale environment variables.
983
528d65ad
JH
984=item LANGUAGE
985
986B<NOTE>: C<LANGUAGE> is a GNU extension, it affects you only if you
987are using the GNU libc. This is the case if you are using e.g. Linux.
e1020413 988If you are using "commercial" Unixes you are most probably I<not>
22b6f60d
JH
989using GNU libc and you can ignore C<LANGUAGE>.
990
991However, in the case you are using C<LANGUAGE>: it affects the
992language of informational, warning, and error messages output by
993commands (in other words, it's like C<LC_MESSAGES>) but it has higher
96090e4f 994priority than C<LC_ALL>. Moreover, it's not a single value but
22b6f60d
JH
995instead a "path" (":"-separated list) of I<languages> (not locales).
996See the GNU C<gettext> library documentation for more information.
528d65ad 997
5f05dabc
PP
998=item LC_CTYPE
999
1000In the absence of C<LC_ALL>, C<LC_CTYPE> chooses the character type
1001locale. In the absence of both C<LC_ALL> and C<LC_CTYPE>, C<LANG>
1002chooses the character type locale.
1003
1004=item LC_COLLATE
1005
14280422
DD
1006In the absence of C<LC_ALL>, C<LC_COLLATE> chooses the collation
1007(sorting) locale. In the absence of both C<LC_ALL> and C<LC_COLLATE>,
1008C<LANG> chooses the collation locale.
5f05dabc
PP
1009
1010=item LC_MONETARY
1011
14280422
DD
1012In the absence of C<LC_ALL>, C<LC_MONETARY> chooses the monetary
1013formatting locale. In the absence of both C<LC_ALL> and C<LC_MONETARY>,
1014C<LANG> chooses the monetary formatting locale.
5f05dabc
PP
1015
1016=item LC_NUMERIC
1017
1018In the absence of C<LC_ALL>, C<LC_NUMERIC> chooses the numeric format
1019locale. In the absence of both C<LC_ALL> and C<LC_NUMERIC>, C<LANG>
1020chooses the numeric format.
1021
1022=item LC_TIME
1023
14280422
DD
1024In the absence of C<LC_ALL>, C<LC_TIME> chooses the date and time
1025formatting locale. In the absence of both C<LC_ALL> and C<LC_TIME>,
1026C<LANG> chooses the date and time formatting locale.
5f05dabc
PP
1027
1028=item LANG
1029
14280422
DD
1030C<LANG> is the "catch-all" locale environment variable. If it is set, it
1031is used as the last resort after the overall C<LC_ALL> and the
5f05dabc
PP
1032category-specific C<LC_...>.
1033
1034=back
1035
7e4353e9
RGS
1036=head2 Examples
1037
1038The LC_NUMERIC controls the numeric output:
1039
ef3087ec
KW
1040 use locale;
1041 use POSIX qw(locale_h); # Imports setlocale() and the LC_ constants.
1042 setlocale(LC_NUMERIC, "fr_FR") or die "Pardon";
1043 printf "%g\n", 1.23; # If the "fr_FR" succeeded, probably shows 1,23.
7e4353e9 1044
39332f68 1045and also how strings are parsed by C<POSIX::strtod()> as numbers:
7e4353e9 1046
ef3087ec
KW
1047 use locale;
1048 use POSIX qw(locale_h strtod);
1049 setlocale(LC_NUMERIC, "de_DE") or die "Entschuldigung";
1050 my $x = strtod("2,34") + 5;
1051 print $x, "\n"; # Probably shows 7,34.
7e4353e9 1052
5f05dabc
PP
1053=head1 NOTES
1054
1055=head2 Backward compatibility
1056
b0c42ed9 1057Versions of Perl prior to 5.004 B<mostly> ignored locale information,
5a964f20
TC
1058generally behaving as if something similar to the C<"C"> locale were
1059always in force, even if the program environment suggested otherwise
1060(see L<The setlocale function>). By default, Perl still behaves this
1061way for backward compatibility. If you want a Perl application to pay
1062attention to locale information, you B<must> use the S<C<use locale>>
062ca197
KW
1063pragma (see L<The use locale pragma>) or, in the unlikely event
1064that you want to do so for just pattern matching, the
70709c68
KW
1065C</l> regular expression modifier (see L<perlre/Character set
1066modifiers>) to instruct it to do so.
b0c42ed9
JH
1067
1068Versions of Perl from 5.002 to 5.003 did use the C<LC_CTYPE>
5a964f20
TC
1069information if available; that is, C<\w> did understand what
1070were the letters according to the locale environment variables.
b0c42ed9
JH
1071The problem was that the user had no control over the feature:
1072if the C library supported locales, Perl used them.
1073
1074=head2 I18N:Collate obsolete
1075
5a964f20 1076In versions of Perl prior to 5.004, per-locale collation was possible
b0c42ed9
JH
1077using the C<I18N::Collate> library module. This module is now mildly
1078obsolete and should be avoided in new applications. The C<LC_COLLATE>
1079functionality is now integrated into the Perl core language: One can
1080use locale-specific scalar data completely normally with C<use locale>,
1081so there is no longer any need to juggle with the scalar references of
1082C<I18N::Collate>.
5f05dabc 1083
14280422 1084=head2 Sort speed and memory use impacts
5f05dabc
PP
1085
1086Comparing and sorting by locale is usually slower than the default
14280422
DD
1087sorting; slow-downs of two to four times have been observed. It will
1088also consume more memory: once a Perl scalar variable has participated
1089in any string comparison or sorting operation obeying the locale
1090collation rules, it will take 3-15 times more memory than before. (The
1091exact multiplier depends on the string's contents, the operating system
1092and the locale.) These downsides are dictated more by the operating
1093system's implementation of the locale system than by Perl.
5f05dabc 1094
39332f68 1095=head2 C<write()> and C<LC_NUMERIC>
e38874e2 1096
903eb63f
NT
1097If a program's environment specifies an LC_NUMERIC locale and C<use
1098locale> is in effect when the format is declared, the locale is used
1099to specify the decimal point character in formatted output. Formatted
39332f68 1100output cannot be controlled by C<use locale> at the time when C<write()>
903eb63f 1101is called.
e38874e2 1102
5f05dabc
PP
1103=head2 Freely available locale definitions
1104
66cbab2c
KW
1105The Unicode CLDR project extracts the POSIX portion of many of its
1106locales, available at
1107
1108 http://unicode.org/Public/cldr/latest/
1109
08d7a6b2
LB
1110There is a large collection of locale definitions at:
1111
1112 http://std.dkuug.dk/i18n/WG15-collection/locales/
1113
1114You should be aware that it is
14280422 1115unsupported, and is not claimed to be fit for any purpose. If your
5a964f20 1116system allows installation of arbitrary locales, you may find the
14280422
DD
1117definitions useful as they are, or as a basis for the development of
1118your own locales.
5f05dabc 1119
14280422 1120=head2 I18n and l10n
5f05dabc 1121
b0c42ed9
JH
1122"Internationalization" is often abbreviated as B<i18n> because its first
1123and last letters are separated by eighteen others. (You may guess why
1124the internalin ... internaliti ... i18n tends to get abbreviated.) In
1125the same way, "localization" is often abbreviated to B<l10n>.
14280422
DD
1126
1127=head2 An imperfect standard
1128
1129Internationalization, as defined in the C and POSIX standards, can be
1130criticized as incomplete, ungainly, and having too large a granularity.
1131(Locales apply to a whole process, when it would arguably be more useful
1132to have them apply to a single thread, window group, or whatever.) They
1133also have a tendency, like standards groups, to divide the world into
1134nations, when we all know that the world can equally well be divided
e199995e 1135into bankers, bikers, gamers, and so on.
5f05dabc 1136
b310b053
JH
1137=head1 Unicode and UTF-8
1138
7ee2ae1e
KW
1139The support of Unicode is new starting from Perl version v5.6, and more fully
1140implemented in version v5.8 and later. See L<perluniintro>. It is
66cbab2c 1141strongly recommended that when combining Unicode and locale (starting in
7ee2ae1e 1142v5.16), you use
66cbab2c
KW
1143
1144 use locale ':not_characters';
1145
1146When this form of the pragma is used, only the non-character portions of
1147locales are used by Perl, for example C<LC_NUMERIC>. Perl assumes that
1148you have translated all the characters it is to operate on into Unicode
1149(actually the platform's native character set (ASCII or EBCDIC) plus
1150Unicode). For data in files, this can conveniently be done by also
1151specifying
1152
1153 use open ':locale';
1154
1155This pragma arranges for all inputs from files to be translated into
1156Unicode from the current locale as specified in the environment (see
1157L</ENVIRONMENT>), and all outputs to files to be translated back
1158into the locale. (See L<open>). On a per-filehandle basis, you can
1159instead use the L<PerlIO::locale> module, or the L<Encode::Locale>
1160module, both available from CPAN. The latter module also has methods to
1161ease the handling of C<ARGV> and environment variables, and can be used
1162on individual strings. Also, if you know that all your locales will be
1163UTF-8, as many are these days, you can use the L<B<-C>|perlrun/-C>
1164command line switch.
1165
1166This form of the pragma allows essentially seamless handling of locales
1167with Unicode. The collation order will be Unicode's. It is strongly
1168recommended that when you need to order and sort strings that you use
1169the standard module L<Unicode::Collate> which gives much better results
1170in many instances than you can get with the old-style locale handling.
1171
7ee2ae1e 1172For pre-v5.16 Perls, or if you use the locale pragma without the
66cbab2c
KW
1173C<:not_characters> parameter, Perl tries to work with both Unicode and
1174locales--but there are problems.
1175
1176Perl does not handle multi-byte locales in this case, such as have been
1177used for various
dc4bfc4b
KW
1178Asian languages, such as Big5 or Shift JIS. However, the increasingly
1179common multi-byte UTF-8 locales, if properly implemented, may work
1180reasonably well (depending on your C library implementation) in this
1181form of the locale pragma, simply because both
1182they and Perl store characters that take up multiple bytes the same way.
1183However, some, if not most, C library implementations may not process
1184the characters in the upper half of the Latin-1 range (128 - 255)
1185properly under LC_CTYPE. To see if a character is a particular type
1186under a locale, Perl uses the functions like C<isalnum()>. Your C
1187library may not work for UTF-8 locales with those functions, instead
1188only working under the newer wide library functions like C<iswalnum()>.
e199995e
KW
1189
1190Perl generally takes the tack to use locale rules on code points that can fit
66cbab2c
KW
1191in a single byte, and Unicode rules for those that can't (though this
1192isn't uniformly applied, see the note at the end of this section). This
1193prevents many problems in locales that aren't UTF-8. Suppose the locale
1194is ISO8859-7, Greek. The character at 0xD7 there is a capital Chi. But
1195in the ISO8859-1 locale, Latin1, it is a multiplication sign. The POSIX
1196regular expression character class C<[[:alpha:]]> will magically match
11970xD7 in the Greek locale but not in the Latin one.
e199995e
KW
1198
1199However, there are places where this breaks down. Certain constructs are
b4ffc3db
TC
1200for Unicode only, such as C<\p{Alpha}>. They assume that 0xD7 always has its
1201Unicode meaning (or the equivalent on EBCDIC platforms). Since Latin1 is a
1202subset of Unicode and 0xD7 is the multiplication sign in both Latin1 and
1203Unicode, C<\p{Alpha}> will never match it, regardless of locale. A similar
1204issue occurs with C<\N{...}>. It is therefore a bad idea to use C<\p{}> or
66cbab2c
KW
1205C<\N{}> under plain C<use locale>--I<unless> you can guarantee that the
1206locale will be a ISO8859-1. Use POSIX character classes instead.
1207
1208Another problem with this approach is that operations that cross the
1209single byte/multiple byte boundary are not well-defined, and so are
1210disallowed. (This boundary is between the codepoints at 255/256.).
1211For example, lower casing LATIN CAPITAL LETTER Y WITH DIAERESIS (U+0178)
1212should return LATIN SMALL LETTER Y WITH DIAERESIS (U+00FF). But in the
1213Greek locale, for example, there is no character at 0xFF, and Perl
1214has no way of knowing what the character at 0xFF is really supposed to
1215represent. Thus it disallows the operation. In this mode, the
1216lowercase of U+0178 is itself.
1217
1218The same problems ensue if you enable automatic UTF-8-ification of your
e199995e 1219standard file handles, default C<open()> layer, and C<@ARGV> on non-ISO8859-1,
b4ffc3db
TC
1220non-UTF-8 locales (by using either the B<-C> command line switch or the
1221C<PERL_UNICODE> environment variable; see L<perlrun>).
1222Things are read in as UTF-8, which would normally imply a Unicode
1223interpretation, but the presence of a locale causes them to be interpreted
1224in that locale instead. For example, a 0xD7 code point in the Unicode
1225input, which should mean the multiplication sign, won't be interpreted by
66cbab2c 1226Perl that way under the Greek locale. This is not a problem
b4ffc3db 1227I<provided> you make certain that all locales will always and only be either
66cbab2c 1228an ISO8859-1, or, if you don't have a deficient C library, a UTF-8 locale.
b4ffc3db
TC
1229
1230Vendor locales are notoriously buggy, and it is difficult for Perl to test
1231its locale-handling code because this interacts with code that Perl has no
1232control over; therefore the locale-handling code in Perl may be buggy as
66cbab2c
KW
1233well. (However, the Unicode-supplied locales should be better, and
1234there is a feed back mechanism to correct any problems. See
1235L</Freely available locale definitions>.)
1236
7ee2ae1e 1237If you have Perl v5.16, the problems mentioned above go away if you use
66cbab2c 1238the C<:not_characters> parameter to the locale pragma (except for vendor
7ee2ae1e 1239bugs in the non-character portions). If you don't have v5.16, and you
66cbab2c
KW
1240I<do> have locales that work, using them may be worthwhile for certain
1241specific purposes, as long as you keep in mind the gotchas already
1242mentioned. For example, if the collation for your locales works, it
1243runs faster under locales than under L<Unicode::Collate>; and you gain
1244access to such things as the local currency symbol and the names of the
7ee2ae1e 1245months and days of the week. (But to hammer home the point, in v5.16,
66cbab2c
KW
1246you get this access without the downsides of locales by using the
1247C<:not_characters> form of the pragma.)
1248
1249Note: The policy of using locale rules for code points that can fit in a
1250byte, and Unicode rules for those that can't is not uniformly applied.
7ee2ae1e 1251Pre-v5.12, it was somewhat haphazard; in v5.12 it was applied fairly
66cbab2c 1252consistently to regular expression matching except for bracketed
7ee2ae1e
KW
1253character classes; in v5.14 it was extended to all regex matches; and in
1254v5.16 to the casing operations such as C<"\L"> and C<uc()>. For
66cbab2c
KW
1255collation, in all releases, the system's C<strxfrm()> function is called,
1256and whatever it does is what you get.
b310b053 1257
5f05dabc
PP
1258=head1 BUGS
1259
1260=head2 Broken systems
1261
5a964f20 1262In certain systems, the operating system's locale support
2bdf8add 1263is broken and cannot be fixed or used by Perl. Such deficiencies can
b4ffc3db 1264and will result in mysterious hangs and/or Perl core dumps when
2bdf8add 1265C<use locale> is in effect. When confronted with such a system,
7f2de2d2 1266please report in excruciating detail to <F<perlbug@perl.org>>, and
b4ffc3db 1267also contact your vendor: bug fixes may exist for these problems
2bdf8add
JH
1268in your operating system. Sometimes such bug fixes are called an
1269operating system upgrade.
5f05dabc
PP
1270
1271=head1 SEE ALSO
1272
b310b053
JH
1273L<I18N::Langinfo>, L<perluniintro>, L<perlunicode>, L<open>,
1274L<POSIX/isalnum>, L<POSIX/isalpha>,
4bbcc6e8
JH
1275L<POSIX/isdigit>, L<POSIX/isgraph>, L<POSIX/islower>,
1276L<POSIX/isprint>, L<POSIX/ispunct>, L<POSIX/isspace>,
1277L<POSIX/isupper>, L<POSIX/isxdigit>, L<POSIX/localeconv>,
1278L<POSIX/setlocale>, L<POSIX/strcoll>, L<POSIX/strftime>,
1279L<POSIX/strtod>, L<POSIX/strxfrm>.
5f05dabc
PP
1280
1281=head1 HISTORY
1282
b0c42ed9 1283Jarkko Hietaniemi's original F<perli18n.pod> heavily hacked by Dominic
5a964f20 1284Dunlop, assisted by the perl5-porters. Prose worked over a bit by
c052850d 1285Tom Christiansen, and updated by Perl 5 porters.