This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
README*: remove deprecated L<"section"> and L<section> syntax
[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
31f05a37
KW
30often, the implementations) of locales. Unicode was invented (see
31L<perlunitut> for an introduction to that) in part to address these
32design deficiencies, and nowadays, there is a series of "UTF-8
33locales", based on Unicode. These are locales whose character set is
34Unicode, encoded in UTF-8. Starting in v5.20, Perl fully supports
9accf6df 35UTF-8 locales, except for sorting and string comparisons like C<lt> and
a4a439fb
KW
36C<ge>. Starting in v5.26, Perl can handle these reasonably as well,
37depending on the platform's implementation. However, for earlier
38releases or for better control, use L<Unicode::Collate> . Perl continues to
39support the old non UTF-8 locales as well. There are currently no UTF-8
40locales for EBCDIC platforms.
31f05a37 41
66cbab2c
KW
42(Unicode is also creating C<CLDR>, the "Common Locale Data Repository",
43L<http://cldr.unicode.org/> which includes more types of information than
44are available in the POSIX locale system. At the time of this writing,
45there was no CPAN module that provides access to this XML-encoded data.
e784ce68
KW
46However, it is possible to compute the POSIX locale data from them, and
47earlier CLDR versions had these already extracted for you as UTF-8 locales
48L<http://unicode.org/Public/cldr/2.0.1/>.)
66cbab2c
KW
49
50=head1 WHAT IS A LOCALE
51
52A locale is a set of data that describes various aspects of how various
53communities in the world categorize their world. These categories are
54broken down into the following types (some of which include a brief
55note here):
56
57=over
58
cb88b78e 59=item Category C<LC_NUMERIC>: Numeric formatting
66cbab2c
KW
60
61This indicates how numbers should be formatted for human readability,
62for example the character used as the decimal point.
63
cb88b78e 64=item Category C<LC_MONETARY>: Formatting of monetary amounts
66cbab2c
KW
65
66=for comment
ebc3223b 67The nbsp below makes this look better (though not great)
66cbab2c
KW
68
69E<160>
70
cb88b78e 71=item Category C<LC_TIME>: Date/Time formatting
66cbab2c
KW
72
73=for comment
ebc3223b 74The nbsp below makes this look better (though not great)
66cbab2c
KW
75
76E<160>
77
cb88b78e 78=item Category C<LC_MESSAGES>: Error and other messages
66cbab2c 79
2619d284 80This is used by Perl itself only for accessing operating system error
03c702c5 81messages via L<$!|perlvar/$ERRNO> and L<$^E|perlvar/$EXTENDED_OS_ERROR>.
66cbab2c 82
cb88b78e 83=item Category C<LC_COLLATE>: Collation
66cbab2c 84
76073c88 85This indicates the ordering of letters for comparison and sorting.
66cbab2c
KW
86In Latin alphabets, for example, "b", generally follows "a".
87
cb88b78e 88=item Category C<LC_CTYPE>: Character Types
66cbab2c
KW
89
90This indicates, for example if a character is an uppercase letter.
91
2619d284
KW
92=item Other categories
93
94Some platforms have other categories, dealing with such things as
95measurement units and paper sizes. None of these are used directly by
96Perl, but outside operations that Perl interacts with may use
d6ded950 97these. See L</Not within the scope of "use locale"> below.
2619d284 98
66cbab2c
KW
99=back
100
2619d284
KW
101More details on the categories used by Perl are given below in L</LOCALE
102CATEGORIES>.
66cbab2c
KW
103
104Together, these categories go a long way towards being able to customize
105a single program to run in many different locations. But there are
106deficiencies, so keep reading.
5f05dabc
PP
107
108=head1 PREPARING TO USE LOCALES
109
c5c88224
KW
110Perl itself (outside the L<POSIX> module) will not use locales unless
111specifically requested to (but
2619d284
KW
112again note that Perl may interact with code that does use them). Even
113if there is such a request, B<all> of the following must be true
b960a36e 114for it to work properly:
5f05dabc
PP
115
116=over 4
117
118=item *
119
120B<Your operating system must support the locale system>. If it does,
39332f68 121you should find that the C<setlocale()> function is a documented part of
5f05dabc
PP
122its C library.
123
124=item *
125
5a964f20 126B<Definitions for locales that you use must be installed>. You, or
14280422
DD
127your system administrator, must make sure that this is the case. The
128available locales, the location in which they are kept, and the manner
5a964f20
TC
129in which they are installed all vary from system to system. Some systems
130provide only a few, hard-wired locales and do not allow more to be
131added. Others allow you to add "canned" locales provided by the system
132supplier. Still others allow you or the system administrator to define
14280422 133and add arbitrary locales. (You may have to ask your supplier to
5a964f20 134provide canned locales that are not delivered with your operating
14280422 135system.) Read your system documentation for further illumination.
5f05dabc
PP
136
137=item *
138
139B<Perl must believe that the locale system is supported>. If it does,
140C<perl -V:d_setlocale> will say that the value for C<d_setlocale> is
141C<define>.
142
143=back
144
145If you want a Perl application to process and present your data
146according to a particular locale, the application code should include
d6ded950 147the S<C<use locale>> pragma (see L<The "use locale" pragma>) where
5f05dabc
PP
148appropriate, and B<at least one> of the following must be true:
149
150=over 4
151
c052850d 152=item 1
5f05dabc 153
66cbab2c 154B<The locale-determining environment variables (see L</"ENVIRONMENT">)
5a964f20 155must be correctly set up> at the time the application is started, either
ef3087ec 156by yourself or by whomever set up your system account; or
5f05dabc 157
c052850d 158=item 2
5f05dabc 159
14280422
DD
160B<The application must set its own locale> using the method described in
161L<The setlocale function>.
5f05dabc
PP
162
163=back
164
165=head1 USING LOCALES
166
d6ded950 167=head2 The C<"use locale"> pragma
5f05dabc 168
fc82b82e
KW
169WARNING! Do NOT use this pragma in scripts that have multiple
170L<threads|threads> active. The locale is not local to a single thread.
171Another thread may change the locale at any time, which could cause at a
172minimum that a given thread is operating in a locale it isn't expecting
173to be in. On some platforms, segfaults can also occur. The locale
174change need not be explicit; some operations cause perl to change the
175locale itself. You are vulnerable simply by having done a C<"use
176locale">.
177
c5c88224
KW
178By default, Perl itself (outside the L<POSIX> module)
179ignores the current locale. The S<C<use locale>>
66cbab2c 180pragma tells Perl to use the current locale for some operations.
d6ded950
KW
181Starting in v5.16, there are optional parameters to this pragma,
182described below, which restrict which operations are affected by it.
c052850d
KW
183
184The current locale is set at execution time by
185L<setlocale()|/The setlocale function> described below. If that function
186hasn't yet been called in the course of the program's execution, the
66cbab2c 187current locale is that which was determined by the L</"ENVIRONMENT"> in
ebc3223b 188effect at the start of the program.
dfcc8045 189If there is no valid environment, the current locale is whatever the
65ebb059
KW
190system default has been set to. On POSIX systems, it is likely, but
191not necessarily, the "C" locale. On Windows, the default is set via the
192computer's S<C<Control Panel-E<gt>Regional and Language Options>> (or its
193current equivalent).
c052850d
KW
194
195The operations that are affected by locale are:
5f05dabc
PP
196
197=over 4
198
d6ded950 199=item B<Not within the scope of C<"use locale">>
b960a36e 200
d6ded950
KW
201Only certain operations originating outside Perl should be affected, as
202follows:
1d2ab946
KW
203
204=over 4
205
206=item *
2619d284 207
663d437a 208The current locale is used when going outside of Perl with
2619d284
KW
209operations like L<system()|perlfunc/system LIST> or
210L<qxE<sol>E<sol>|perlop/qxE<sol>STRINGE<sol>>, if those operations are
211locale-sensitive.
212
1d2ab946
KW
213=item *
214
2619d284
KW
215Also Perl gives access to various C library functions through the
216L<POSIX> module. Some of those functions are always affected by the
217current locale. For example, C<POSIX::strftime()> uses C<LC_TIME>;
218C<POSIX::strtod()> uses C<LC_NUMERIC>; C<POSIX::strcoll()> and
9accf6df 219C<POSIX::strxfrm()> use C<LC_COLLATE>. All such functions
2619d284 220will behave according to the current underlying locale, even if that
1d2ab946 221locale isn't exposed to Perl space.
2619d284 222
1d2ab946
KW
223=item *
224
52686f2a
KW
225XS modules for all categories but C<LC_NUMERIC> get the underlying
226locale, and hence any C library functions they call will use that
8a384d3a 227underlying locale. For more discussion, see L<perlxs/CAVEATS>.
1d2ab946
KW
228
229=back
230
c5c88224
KW
231Note that all C programs (including the perl interpreter, which is
232written in C) always have an underlying locale. That locale is the "C"
233locale unless changed by a call to L<setlocale()|/The setlocale
234function>. When Perl starts up, it changes the underlying locale to the
235one which is indicated by the L</ENVIRONMENT>. When using the L<POSIX>
236module or writing XS code, it is important to keep in mind that the
237underlying locale may be something other than "C", even if the program
238hasn't explicitly changed it.
239
ebc3223b
KW
240=for comment
241The nbsp below makes this look better (though not great)
242
243E<160>
244
245=item B<Lingering effects of C<S<use locale>>>
1d2ab946
KW
246
247Certain Perl operations that are set-up within the scope of a
d6ded950 248C<use locale> retain that effect even outside the scope.
b960a36e
KW
249These include:
250
251=over 4
252
253=item *
254
255The output format of a L<write()|perlfunc/write> is determined by an
256earlier format declaration (L<perlfunc/format>), so whether or not the
257output is affected by locale is determined by if the C<format()> is
d6ded950 258within the scope of a C<use locale>, not whether the C<write()>
b960a36e
KW
259is.
260
261=item *
262
263Regular expression patterns can be compiled using
33be4c61 264L<qrE<sol>E<sol>|perlop/qrE<sol>STRINGE<sol>msixpodualn> with actual
b960a36e
KW
265matching deferred to later. Again, it is whether or not the compilation
266was done within the scope of C<use locale> that determines the match
267behavior, not if the matches are done within such a scope or not.
268
269=back
270
ebc3223b
KW
271=for comment
272The nbsp below makes this look better (though not great)
273
d6ded950 274
ebc3223b
KW
275E<160>
276
d6ded950 277=item B<Under C<"use locale";>>
66cbab2c
KW
278
279=over 4
280
281=item *
282
d6ded950 283All the above operations
b960a36e
KW
284
285=item *
286
287B<Format declarations> (L<perlfunc/format>) and hence any subsequent
288C<write()>s use C<LC_NUMERIC>.
66cbab2c
KW
289
290=item *
291
b960a36e
KW
292B<stringification and output> use C<LC_NUMERIC>.
293These include the results of
294C<print()>,
295C<printf()>,
296C<say()>,
297and
298C<sprintf()>.
66cbab2c 299
b960a36e
KW
300=item *
301
302B<The comparison operators> (C<lt>, C<le>, C<cmp>, C<ge>, and C<gt>) use
39332f68 303C<LC_COLLATE>. C<sort()> is also affected if used without an
5a964f20 304explicit comparison function, because it uses C<cmp> by default.
14280422 305
5a964f20 306B<Note:> C<eq> and C<ne> are unaffected by locale: they always
de108802 307perform a char-by-char comparison of their scalar operands. What's
14280422
DD
308more, if C<cmp> finds that its operands are equal according to the
309collation sequence specified by the current locale, it goes on to
de108802
RGS
310perform a char-by-char comparison, and only returns I<0> (equal) if the
311operands are char-for-char identical. If you really want to know whether
5a964f20 312two strings--which C<eq> and C<cmp> may consider different--are equal
14280422 313as far as collation in the locale is concerned, see the discussion in
cb88b78e 314L<Category C<LC_COLLATE>: Collation>.
5f05dabc
PP
315
316=item *
317
39332f68
KW
318B<Regular expressions and case-modification functions> (C<uc()>, C<lc()>,
319C<ucfirst()>, and C<lcfirst()>) use C<LC_CTYPE>
5f05dabc 320
2c6ee1a7
KW
321=item *
322
dbf3c4d7
KW
323B<The variables L<$!|perlvar/$ERRNO>> (and its synonyms C<$ERRNO> and
324C<$OS_ERROR>) B<and L<$^E|perlvar/$EXTENDED_OS_ERROR>> (and its synonym
2c6ee1a7
KW
325C<$EXTENDED_OS_ERROR>) when used as strings use C<LC_MESSAGES>.
326
5f05dabc
PP
327=back
328
66cbab2c 329=back
5f05dabc 330
5a964f20 331The default behavior is restored with the S<C<no locale>> pragma, or
ef3087ec 332upon reaching the end of the block enclosing C<use locale>.
3cd61afa 333Note that C<use locale> calls may be
66cbab2c
KW
334nested, and that what is in effect within an inner scope will revert to
335the outer scope's rules at the end of the inner scope.
5f05dabc 336
5a964f20 337The string result of any operation that uses locale
14280422
DD
338information is tainted, as it is possible for a locale to be
339untrustworthy. See L<"SECURITY">.
5f05dabc 340
d6ded950
KW
341Starting in Perl v5.16 in a very limited way, and more generally in
342v5.22, you can restrict which category or categories are enabled by this
343particular instance of the pragma by adding parameters to it. For
344example,
345
346 use locale qw(:ctype :numeric);
347
348enables locale awareness within its scope of only those operations
349(listed above) that are affected by C<LC_CTYPE> and C<LC_NUMERIC>.
350
351The possible categories are: C<:collate>, C<:ctype>, C<:messages>,
352C<:monetary>, C<:numeric>, C<:time>, and the pseudo category
353C<:characters> (described below).
354
355Thus you can say
356
357 use locale ':messages';
358
359and only L<$!|perlvar/$ERRNO> and L<$^E|perlvar/$EXTENDED_OS_ERROR>
360will be locale aware. Everything else is unaffected.
361
362Since Perl doesn't currently do anything with the C<LC_MONETARY>
363category, specifying C<:monetary> does effectively nothing. Some
364systems have other categories, such as C<LC_PAPER_SIZE>, but Perl
365also doesn't know anything about them, and there is no way to specify
366them in this pragma's arguments.
367
368You can also easily say to use all categories but one, by either, for
369example,
370
371 use locale ':!ctype';
372 use locale ':not_ctype';
373
374both of which mean to enable locale awarness of all categories but
375C<LC_CTYPE>. Only one category argument may be specified in a
376S<C<use locale>> if it is of the negated form.
377
378Prior to v5.22 only one form of the pragma with arguments is available:
379
380 use locale ':not_characters';
381
382(and you have to say C<not_>; you can't use the bang C<!> form). This
383pseudo category is a shorthand for specifying both C<:collate> and
384C<:ctype>. Hence, in the negated form, it is nearly the same thing as
385saying
386
387 use locale qw(:messages :monetary :numeric :time);
388
389We use the term "nearly", because C<:not_characters> also turns on
390S<C<use feature 'unicode_strings'>> within its scope. This form is
391less useful in v5.20 and later, and is described fully in
392L</Unicode and UTF-8>, but briefly, it tells Perl to not use the
393character portions of the locale definition, that is the C<LC_CTYPE> and
394C<LC_COLLATE> categories. Instead it will use the native character set
395(extended by Unicode). When using this parameter, you are responsible
396for getting the external character set translated into the
397native/Unicode one (which it already will be if it is one of the
398increasingly popular UTF-8 locales). There are convenient ways of doing
399this, as described in L</Unicode and UTF-8>.
400
5f05dabc
PP
401=head2 The setlocale function
402
fc82b82e
KW
403WARNING! Do NOT use this function in a L<thread|threads>. The locale
404will change in all other threads at the same time, and should your
405thread get paused by the operating system, and another started, that
406thread will not have the locale it is expecting. On some platforms,
407there can be a race leading to segfaults if two threads call this
408function nearly simultaneously.
409
14280422 410You can switch locales as often as you wish at run time with the
39332f68 411C<POSIX::setlocale()> function:
5f05dabc 412
5f05dabc
PP
413 # Import locale-handling tool set from POSIX module.
414 # This example uses: setlocale -- the function call
415 # LC_CTYPE -- explained below
2619d284
KW
416 # (Showing the testing for success/failure of operations is
417 # omitted in these examples to avoid distracting from the main
ebc3223b 418 # point)
6ea81ccf 419
5f05dabc 420 use POSIX qw(locale_h);
dfcc8045
KW
421 use locale;
422 my $old_locale;
5f05dabc 423
14280422 424 # query and save the old locale
5f05dabc
PP
425 $old_locale = setlocale(LC_CTYPE);
426
427 setlocale(LC_CTYPE, "fr_CA.ISO8859-1");
428 # LC_CTYPE now in locale "French, Canada, codeset ISO 8859-1"
429
430 setlocale(LC_CTYPE, "");
65ebb059
KW
431 # LC_CTYPE now reset to the default defined by the
432 # LC_ALL/LC_CTYPE/LANG environment variables, or to the system
433 # default. See below for documentation.
5f05dabc
PP
434
435 # restore the old locale
436 setlocale(LC_CTYPE, $old_locale);
437
39332f68 438The first argument of C<setlocale()> gives the B<category>, the second the
14280422
DD
439B<locale>. The category tells in what aspect of data processing you
440want to apply locale-specific rules. Category names are discussed in
66cbab2c 441L</LOCALE CATEGORIES> and L</"ENVIRONMENT">. The locale is the name of a
14280422
DD
442collection of customization information corresponding to a particular
443combination of language, country or territory, and codeset. Read on for
444hints on the naming of locales: not all systems name locales as in the
445example.
446
39332f68 447If no second argument is provided and the category is something other
cb88b78e 448than C<LC_ALL>, the function returns a string naming the current locale
502a173a 449for the category. You can use this value as the second argument in a
f170b852
KW
450subsequent call to C<setlocale()>, B<but> on some platforms the string
451is opaque, not something that most people would be able to decipher as
452to what locale it means.
502a173a 453
cb88b78e 454If no second argument is provided and the category is C<LC_ALL>, the
502a173a 455result is implementation-dependent. It may be a string of
c052850d 456concatenated locale names (separator also implementation-dependent)
39332f68 457or a single locale name. Please consult your L<setlocale(3)> man page for
502a173a
JH
458details.
459
460If a second argument is given and it corresponds to a valid locale,
461the locale for the category is set to that value, and the function
462returns the now-current locale value. You can then use this in yet
39332f68 463another call to C<setlocale()>. (In some implementations, the return
502a173a
JH
464value may sometimes differ from the value you gave as the second
465argument--think of it as an alias for the value you gave.)
5f05dabc
PP
466
467As the example shows, if the second argument is an empty string, the
468category's locale is returned to the default specified by the
469corresponding environment variables. Generally, this results in a
5a964f20 470return to the default that was in force when Perl started up: changes
54310121 471to the environment made by the application after startup may or may not
5a964f20 472be noticed, depending on your system's C library.
5f05dabc 473
d6ded950
KW
474Note that when a form of C<use locale> that doesn't include all
475categories is specified, Perl ignores the excluded categories.
66cbab2c 476
f170b852 477If C<set_locale()> fails for some reason (for example, an attempt to set
dfcc8045
KW
478to a locale unknown to the system), the locale for the category is not
479changed, and the function returns C<undef>.
480
2619d284 481
39332f68 482For further information about the categories, consult L<setlocale(3)>.
3e6e419a
JH
483
484=head2 Finding locales
485
39332f68 486For locales available in your system, consult also L<setlocale(3)> to
5a964f20
TC
487see whether it leads to the list of available locales (search for the
488I<SEE ALSO> section). If that fails, try the following command lines:
5f05dabc
PP
489
490 locale -a
491
492 nlsinfo
493
494 ls /usr/lib/nls/loc
495
496 ls /usr/lib/locale
497
498 ls /usr/lib/nls
499
b478f28d
JH
500 ls /usr/share/locale
501
5f05dabc
PP
502and see whether they list something resembling these
503
2bdf8add 504 en_US.ISO8859-1 de_DE.ISO8859-1 ru_RU.ISO8859-5
502a173a 505 en_US.iso88591 de_DE.iso88591 ru_RU.iso88595
2bdf8add 506 en_US de_DE ru_RU
14280422 507 en de ru
2bdf8add
JH
508 english german russian
509 english.iso88591 german.iso88591 russian.iso88595
502a173a 510 english.roman8 russian.koi8r
5f05dabc 511
39332f68 512Sadly, even though the calling interface for C<setlocale()> has been
528d65ad 513standardized, names of locales and the directories where the
5a964f20 514configuration resides have not been. The basic form of the name is
528d65ad
JH
515I<language_territory>B<.>I<codeset>, but the latter parts after
516I<language> are not always present. The I<language> and I<country>
517are usually from the standards B<ISO 3166> and B<ISO 639>, the
518two-letter abbreviations for the countries and the languages of the
519world, respectively. The I<codeset> part often mentions some B<ISO
5208859> character set, the Latin codesets. For example, C<ISO 8859-1>
521is the so-called "Western European codeset" that can be used to encode
522most Western European languages adequately. Again, there are several
523ways to write even the name of that one standard. Lamentably.
5f05dabc 524
14280422
DD
525Two special locales are worth particular mention: "C" and "POSIX".
526Currently these are effectively the same locale: the difference is
5a964f20
TC
527mainly that the first one is defined by the C standard, the second by
528the POSIX standard. They define the B<default locale> in which
14280422 529every program starts in the absence of locale information in its
5a964f20 530environment. (The I<default> default locale, if you will.) Its language
39332f68
KW
531is (American) English and its character codeset ASCII or, rarely, a
532superset thereof (such as the "DEC Multinational Character Set
533(DEC-MCS)"). B<Warning>. The C locale delivered by some vendors
534may not actually exactly match what the C standard calls for. So
535beware.
5f05dabc 536
14280422
DD
537B<NOTE>: Not all systems have the "POSIX" locale (not all systems are
538POSIX-conformant), so use "C" when you need explicitly to specify this
539default locale.
5f05dabc 540
3e6e419a
JH
541=head2 LOCALE PROBLEMS
542
5a964f20 543You may encounter the following warning message at Perl startup:
3e6e419a
JH
544
545 perl: warning: Setting locale failed.
546 perl: warning: Please check that your locale settings:
547 LC_ALL = "En_US",
548 LANG = (unset)
549 are supported and installed on your system.
550 perl: warning: Falling back to the standard locale ("C").
551
cb88b78e 552This means that your locale settings had C<LC_ALL> set to "En_US" and
5a964f20
TC
553LANG exists but has no value. Perl tried to believe you but could not.
554Instead, Perl gave up and fell back to the "C" locale, the default locale
65ebb059
KW
555that is supposed to work no matter what. (On Windows, it first tries
556falling back to the system default locale.) This usually means your
557locale settings were wrong, they mention locales your system has never
558heard of, or the locale installation in your system has problems (for
559example, some system files are broken or missing). There are quick and
560temporary fixes to these problems, as well as more thorough and lasting
561fixes.
3e6e419a 562
83fb1bf0
KW
563=head2 Testing for broken locales
564
565If you are building Perl from source, the Perl test suite file
566F<lib/locale.t> can be used to test the locales on your system.
567Setting the environment variable C<PERL_DEBUG_FULL_TEST> to 1
568will cause it to output detailed results. For example, on Linux, you
569could say
570
1d2ab946 571 PERL_DEBUG_FULL_TEST=1 ./perl -T -Ilib lib/locale.t > locale.log 2>&1
83fb1bf0
KW
572
573Besides many other tests, it will test every locale it finds on your
574system to see if they conform to the POSIX standard. If any have
575errors, it will include a summary near the end of the output of which
576locales passed all its tests, and which failed, and why.
577
3e6e419a
JH
578=head2 Temporarily fixing locale problems
579
5a964f20 580The two quickest fixes are either to render Perl silent about any
3e6e419a
JH
581locale inconsistencies or to run Perl under the default locale "C".
582
583Perl's moaning about locale problems can be silenced by setting the
22ff3130 584environment variable C<PERL_BADLANG> to "0" or "".
900bd440
JH
585This method really just sweeps the problem under the carpet: you tell
586Perl to shut up even when Perl sees that something is wrong. Do not
587be surprised if later something locale-dependent misbehaves.
3e6e419a
JH
588
589Perl can be run under the "C" locale by setting the environment
cb88b78e
KW
590variable C<LC_ALL> to "C". This method is perhaps a bit more civilized
591than the C<PERL_BADLANG> approach, but setting C<LC_ALL> (or
5a964f20
TC
592other locale variables) may affect other programs as well, not just
593Perl. In particular, external programs run from within Perl will see
3e6e419a 594these changes. If you make the new settings permanent (read on), all
f979aebc 595programs you run see the changes. See L<"ENVIRONMENT"> for
5a964f20 596the full list of relevant environment variables and L<USING LOCALES>
e05ffc7d 597for their effects in Perl. Effects in other programs are
cb88b78e 598easily deducible. For example, the variable C<LC_COLLATE> may well affect
b432a672 599your B<sort> program (or whatever the program that arranges "records"
3e6e419a
JH
600alphabetically in your system is called).
601
5a964f20
TC
602You can test out changing these variables temporarily, and if the
603new settings seem to help, put those settings into your shell startup
663d437a 604files. Consult your local documentation for the exact details. For
5a964f20 605Bourne-like shells (B<sh>, B<ksh>, B<bash>, B<zsh>):
3e6e419a
JH
606
607 LC_ALL=en_US.ISO8859-1
608 export LC_ALL
609
5a964f20
TC
610This assumes that we saw the locale "en_US.ISO8859-1" using the commands
611discussed above. We decided to try that instead of the above faulty
612locale "En_US"--and in Cshish shells (B<csh>, B<tcsh>)
3e6e419a
JH
613
614 setenv LC_ALL en_US.ISO8859-1
c47ff5f1 615
663d437a 616or if you have the "env" application you can do (in any shell)
c406981e
JH
617
618 env LC_ALL=en_US.ISO8859-1 perl ...
619
5a964f20 620If you do not know what shell you have, consult your local
3e6e419a
JH
621helpdesk or the equivalent.
622
623=head2 Permanently fixing locale problems
624
5a964f20
TC
625The slower but superior fixes are when you may be able to yourself
626fix the misconfiguration of your own environment variables. The
3e6e419a
JH
627mis(sing)configuration of the whole system's locales usually requires
628the help of your friendly system administrator.
629
5a964f20
TC
630First, see earlier in this document about L<Finding locales>. That tells
631how to find which locales are really supported--and more importantly,
632installed--on your system. In our example error message, environment
633variables affecting the locale are listed in the order of decreasing
634importance (and unset variables do not matter). Therefore, having
635LC_ALL set to "En_US" must have been the bad choice, as shown by the
636error message. First try fixing locale settings listed first.
3e6e419a 637
5a964f20
TC
638Second, if using the listed commands you see something B<exactly>
639(prefix matches do not count and case usually counts) like "En_US"
640without the quotes, then you should be okay because you are using a
641locale name that should be installed and available in your system.
4a4eefd0 642In this case, see L<Permanently fixing your system's locale configuration>.
3e6e419a 643
4a4eefd0 644=head2 Permanently fixing your system's locale configuration
3e6e419a 645
5a964f20 646This is when you see something like:
3e6e419a
JH
647
648 perl: warning: Please check that your locale settings:
649 LC_ALL = "En_US",
650 LANG = (unset)
651 are supported and installed on your system.
652
653but then cannot see that "En_US" listed by the above-mentioned
5a964f20
TC
654commands. You may see things like "en_US.ISO8859-1", but that isn't
655the same. In this case, try running under a locale
656that you can list and which somehow matches what you tried. The
3e6e419a 657rules for matching locale names are a bit vague because
e05ffc7d 658standardization is weak in this area. See again the
13a2d996 659L<Finding locales> about general rules.
3e6e419a 660
b687b08b 661=head2 Fixing system locale configuration
3e6e419a 662
5a964f20
TC
663Contact a system administrator (preferably your own) and report the exact
664error message you get, and ask them to read this same documentation you
665are now reading. They should be able to check whether there is something
666wrong with the locale configuration of the system. The L<Finding locales>
667section is unfortunately a bit vague about the exact commands and places
668because these things are not that standardized.
3e6e419a 669
5f05dabc
PP
670=head2 The localeconv function
671
39332f68 672The C<POSIX::localeconv()> function allows you to get particulars of the
14280422 673locale-dependent numeric formatting information specified by the current
a835cd47
KW
674underlying C<LC_NUMERIC> and C<LC_MONETARY> locales (regardless of
675whether called from within the scope of C<S<use locale>> or not). (If
676you just want the name of
39332f68 677the current locale for a particular category, use C<POSIX::setlocale()>
5a964f20 678with a single parameter--see L<The setlocale function>.)
5f05dabc
PP
679
680 use POSIX qw(locale_h);
5f05dabc
PP
681
682 # Get a reference to a hash of locale-dependent info
683 $locale_values = localeconv();
684
685 # Output sorted list of the values
686 for (sort keys %$locale_values) {
14280422 687 printf "%-20s = %s\n", $_, $locale_values->{$_}
5f05dabc
PP
688 }
689
39332f68 690C<localeconv()> takes no arguments, and returns B<a reference to> a hash.
5a964f20 691The keys of this hash are variable names for formatting, such as
502a173a 692C<decimal_point> and C<thousands_sep>. The values are the
cea6626f 693corresponding, er, values. See L<POSIX/localeconv> for a longer
502a173a
JH
694example listing the categories an implementation might be expected to
695provide; some provide more and others fewer. You don't need an
39332f68 696explicit C<use locale>, because C<localeconv()> always observes the
502a173a 697current locale.
5f05dabc 698
5a964f20
TC
699Here's a simple-minded example program that rewrites its command-line
700parameters as integers correctly formatted in the current locale:
5f05dabc 701
ef3087ec
KW
702 use POSIX qw(locale_h);
703
704 # Get some of locale's numeric formatting parameters
705 my ($thousands_sep, $grouping) =
706 @{localeconv()}{'thousands_sep', 'grouping'};
707
708 # Apply defaults if values are missing
709 $thousands_sep = ',' unless $thousands_sep;
710
711 # grouping and mon_grouping are packed lists
712 # of small integers (characters) telling the
713 # grouping (thousand_seps and mon_thousand_seps
714 # being the group dividers) of numbers and
715 # monetary quantities. The integers' meanings:
716 # 255 means no more grouping, 0 means repeat
717 # the previous grouping, 1-254 means use that
718 # as the current grouping. Grouping goes from
719 # right to left (low to high digits). In the
720 # below we cheat slightly by never using anything
721 # else than the first grouping (whatever that is).
722 if ($grouping) {
723 @grouping = unpack("C*", $grouping);
724 } else {
725 @grouping = (3);
726 }
727
728 # Format command line params for current locale
729 for (@ARGV) {
730 $_ = int; # Chop non-integer part
731 1 while
732 s/(\d)(\d{$grouping[0]}($|$thousands_sep))/$1$thousands_sep$2/;
733 print "$_";
734 }
735 print "\n";
5f05dabc 736
03ceeedf
KW
737Note that if the platform doesn't have C<LC_NUMERIC> and/or
738C<LC_MONETARY> available or enabled, the corresponding elements of the
739hash will be missing.
740
74c76037 741=head2 I18N::Langinfo
4bbcc6e8
JH
742
743Another interface for querying locale-dependent information is the
39332f68 744C<I18N::Langinfo::langinfo()> function, available at least in Unix-like
4bbcc6e8
JH
745systems and VMS.
746
39332f68
KW
747The following example will import the C<langinfo()> function itself and
748three constants to be used as arguments to C<langinfo()>: a constant for
74c76037
JH
749the abbreviated first day of the week (the numbering starts from
750Sunday = 1) and two more constants for the affirmative and negative
751answers for a yes/no question in the current locale.
4bbcc6e8 752
74c76037 753 use I18N::Langinfo qw(langinfo ABDAY_1 YESSTR NOSTR);
4bbcc6e8 754
ef3087ec
KW
755 my ($abday_1, $yesstr, $nostr)
756 = map { langinfo } qw(ABDAY_1 YESSTR NOSTR);
4bbcc6e8 757
74c76037 758 print "$abday_1? [$yesstr/$nostr] ";
4bbcc6e8 759
74c76037
JH
760In other words, in the "C" (or English) locale the above will probably
761print something like:
762
e05ffc7d 763 Sun? [yes/no]
4bbcc6e8
JH
764
765See L<I18N::Langinfo> for more information.
766
5f05dabc
PP
767=head1 LOCALE CATEGORIES
768
5a964f20
TC
769The following subsections describe basic locale categories. Beyond these,
770some combination categories allow manipulation of more than one
771basic category at a time. See L<"ENVIRONMENT"> for a discussion of these.
5f05dabc 772
ff52fcf1 773=head2 Category C<LC_COLLATE>: Collation: Text Comparisons and Sorting
5f05dabc 774
d6ded950
KW
775In the scope of a S<C<use locale>> form that includes collation, Perl
776looks to the C<LC_COLLATE>
5a964f20 777environment variable to determine the application's notions on collation
b4ffc3db
TC
778(ordering) of characters. For example, "b" follows "a" in Latin
779alphabets, but where do "E<aacute>" and "E<aring>" belong? And while
f87fa335 780"color" follows "chocolate" in English, what about in traditional Spanish?
5f05dabc 781
60f0fa02 782The following collations all make sense and you may meet any of them
dbf3c4d7 783if you C<"use locale">.
60f0fa02
JH
784
785 A B C D E a b c d e
35316ca3 786 A a B b C c D d E e
60f0fa02
JH
787 a A b B c C d D e E
788 a b c d e A B C D E
789
f1cbbd6e 790Here is a code snippet to tell what "word"
5a964f20 791characters are in the current locale, in that locale's order:
5f05dabc
PP
792
793 use locale;
35316ca3 794 print +(sort grep /\w/, map { chr } 0..255), "\n";
5f05dabc 795
14280422
DD
796Compare this with the characters that you see and their order if you
797state explicitly that the locale should be ignored:
5f05dabc
PP
798
799 no locale;
35316ca3 800 print +(sort grep /\w/, map { chr } 0..255), "\n";
5f05dabc
PP
801
802This machine-native collation (which is what you get unless S<C<use
803locale>> has appeared earlier in the same block) must be used for
804sorting raw binary data, whereas the locale-dependent collation of the
b0c42ed9 805first example is useful for natural text.
5f05dabc 806
14280422
DD
807As noted in L<USING LOCALES>, C<cmp> compares according to the current
808collation locale when C<use locale> is in effect, but falls back to a
de108802 809char-by-char comparison for strings that the locale says are equal. You
39332f68 810can use C<POSIX::strcoll()> if you don't want this fall-back:
14280422
DD
811
812 use POSIX qw(strcoll);
813 $equal_in_locale =
814 !strcoll("space and case ignored", "SpaceAndCaseIgnored");
815
39332f68 816C<$equal_in_locale> will be true if the collation locale specifies a
5a964f20 817dictionary-like ordering that ignores space characters completely and
9e3a2af8 818which folds case.
14280422 819
a4a439fb
KW
820Perl uses the platform's C library collation functions C<strcoll()> and
821C<strxfrm()>. That means you get whatever they give. On some
822platforms, these functions work well on UTF-8 locales, giving
823a reasonable default collation for the code points that are important in
824that locale. (And if they aren't working well, the problem may only be
825that the locale definition is deficient, so can be fixed by using a
826better definition file. Unicode's definitions (see L</Freely available
827locale definitions>) provide reasonable UTF-8 locale collation
828definitions.) Starting in Perl v5.26, Perl's use of these functions has
829been made more seamless. This may be sufficient for your needs. For
830more control, and to make sure strings containing any code point (not
831just the ones important in the locale) collate properly, the
832L<Unicode::Collate> module is suggested.
833
834In non-UTF-8 locales (hence single byte), code points above 0xFF are
835technically invalid. But if present, again starting in v5.26, they will
836collate to the same position as the highest valid code point does. This
837generally gives good results, but the collation order may be skewed if
838the valid code point gets special treatment when it forms particular
839sequences with other characters as defined by the locale.
840When two strings collate identically, the code point order is used as a
841tie breaker.
842
843If Perl detects that there are problems with the locale collation order,
844it reverts to using non-locale collation rules for that locale.
31f05a37 845
79f120c8
KW
846If Perl detects that there are problems with the locale collation order,
847it reverts to using non-locale collation rules for that locale.
848
5a964f20 849If you have a single string that you want to check for "equality in
14280422 850locale" against several others, you might think you could gain a little
39332f68 851efficiency by using C<POSIX::strxfrm()> in conjunction with C<eq>:
14280422
DD
852
853 use POSIX qw(strxfrm);
854 $xfrm_string = strxfrm("Mixed-case string");
855 print "locale collation ignores spaces\n"
856 if $xfrm_string eq strxfrm("Mixed-casestring");
857 print "locale collation ignores hyphens\n"
858 if $xfrm_string eq strxfrm("Mixedcase string");
859 print "locale collation ignores case\n"
860 if $xfrm_string eq strxfrm("mixed-case string");
861
39332f68 862C<strxfrm()> takes a string and maps it into a transformed string for use
de108802 863in char-by-char comparisons against other transformed strings during
14280422 864collation. "Under the hood", locale-affected Perl comparison operators
39332f68
KW
865call C<strxfrm()> for both operands, then do a char-by-char
866comparison of the transformed strings. By calling C<strxfrm()> explicitly
14280422 867and using a non locale-affected comparison, the example attempts to save
5a964f20 868a couple of transformations. But in fact, it doesn't save anything: Perl
2ae324a7 869magic (see L<perlguts/Magic Variables>) creates the transformed version of a
5a964f20 870string the first time it's needed in a comparison, then keeps this version around
14280422 871in case it's needed again. An example rewritten the easy way with
e38874e2 872C<cmp> runs just about as fast. It also copes with null characters
39332f68 873embedded in strings; if you call C<strxfrm()> directly, it treats the first
5a964f20
TC
874null it finds as a terminator. don't expect the transformed strings
875it produces to be portable across systems--or even from one revision
39332f68 876of your operating system to the next. In short, don't call C<strxfrm()>
e38874e2 877directly: let Perl do it for you.
14280422 878
5a964f20 879Note: C<use locale> isn't shown in some of these examples because it isn't
dfcc8045
KW
880needed: C<strcoll()> and C<strxfrm()> are POSIX functions
881which use the standard system-supplied C<libc> functions that
882always obey the current C<LC_COLLATE> locale.
5f05dabc 883
cb88b78e 884=head2 Category C<LC_CTYPE>: Character Types
5f05dabc 885
d6ded950
KW
886In the scope of a S<C<use locale>> form that includes C<LC_CTYPE>, Perl
887obeys the C<LC_CTYPE> locale
14280422 888setting. This controls the application's notion of which characters are
ebc3223b
KW
889alphabetic, numeric, punctuation, I<etc>. This affects Perl's C<\w>
890regular expression metanotation,
f1cbbd6e 891which stands for alphanumeric characters--that is, alphabetic,
ebc3223b
KW
892numeric, and the platform's native underscore.
893(Consult L<perlre> for more information about
14280422 894regular expressions.) Thanks to C<LC_CTYPE>, depending on your locale
b4ffc3db
TC
895setting, characters like "E<aelig>", "E<eth>", "E<szlig>", and
896"E<oslash>" may be understood as C<\w> characters.
ebc3223b
KW
897It also affects things like C<\s>, C<\D>, and the POSIX character
898classes, like C<[[:graph:]]>. (See L<perlrecharclass> for more
899information on all these.)
5f05dabc 900
2c268ad5 901The C<LC_CTYPE> locale also provides the map used in transliterating
68dc0745 902characters between lower and uppercase. This affects the case-mapping
663d437a
KW
903functions--C<fc()>, C<lc()>, C<lcfirst()>, C<uc()>, and C<ucfirst()>;
904case-mapping
b9cc4f69
KW
905interpolation with C<\F>, C<\l>, C<\L>, C<\u>, or C<\U> in double-quoted
906strings and C<s///> substitutions; and case-independent regular expression
e38874e2
DD
907pattern matching using the C<i> modifier.
908
31f05a37
KW
909Starting in v5.20, Perl supports UTF-8 locales for C<LC_CTYPE>, but
910otherwise Perl only supports single-byte locales, such as the ISO 8859
911series. This means that wide character locales, for example for Asian
8c6180a9
KW
912languages, are not well-supported. (If the platform has the capability
913for Perl to detect such a locale, starting in Perl v5.22,
914L<Perl will warn, default enabled|warnings/Category Hierarchy>,
915using the C<locale> warning category, whenever such a locale is switched
916into.) The UTF-8 locale support is actually a
31f05a37 917superset of POSIX locales, because it is really full Unicode behavior
13af4fd9
KW
918as if no C<LC_CTYPE> locale were in effect at all (except for tainting;
919see L</SECURITY>). POSIX locales, even UTF-8 ones,
31f05a37
KW
920are lacking certain concepts in Unicode, such as the idea that changing
921the case of a character could expand to be more than one character.
922Perl in a UTF-8 locale, will give you that expansion. Prior to v5.20,
923Perl treated a UTF-8 locale on some platforms like an ISO 8859-1 one,
924with some restrictions, and on other platforms more like the "C" locale.
925For releases v5.16 and v5.18, C<S<use locale 'not_characters>> could be
926used as a workaround for this (see L</Unicode and UTF-8>).
927
5d63e270 928Note that there are quite a few things that are unaffected by the
663d437a
KW
929current locale. Any literal character is the native character for the
930given platform. Hence 'A' means the character at code point 65 on ASCII
931platforms, and 193 on EBCDIC. That may or may not be an 'A' in the
932current locale, if that locale even has an 'A'.
933Similarly, all the escape sequences for particular characters,
5d63e270
KW
934C<\n> for example, always mean the platform's native one. This means,
935for example, that C<\N> in regular expressions (every character
1d2ab946 936but new-line) works on the platform character set.
5d63e270 937
8c6180a9
KW
938Starting in v5.22, Perl will by default warn when switching into a
939locale that redefines any ASCII printable character (plus C<\t> and
84035de0
KW
940C<\n>) into a different class than expected. This is likely to
941happen on modern locales only on EBCDIC platforms, where, for example,
942a CCSID 0037 locale on a CCSID 1047 machine moves C<"[">, but it can
943happen on ASCII platforms with the ISO 646 and other
8c6180a9
KW
9447-bit locales that are essentially obsolete. Things may still work,
945depending on what features of Perl are used by the program. For
946example, in the example from above where C<"|"> becomes a C<\w>, and
947there are no regular expressions where this matters, the program may
948still work properly. The warning lists all the characters that
949it can determine could be adversely affected.
950
14280422
DD
951B<Note:> A broken or malicious C<LC_CTYPE> locale definition may result
952in clearly ineligible characters being considered to be alphanumeric by
e199995e 953your application. For strict matching of (mundane) ASCII letters and
5a964f20 954digits--for example, in command strings--locale-aware applications
e199995e 955should use C<\w> with the C</a> regular expression modifier. See L<"SECURITY">.
5f05dabc 956
fee33030 957=head2 Category C<LC_NUMERIC>: Numeric Formatting
5f05dabc 958
d6ded950
KW
959After a proper C<POSIX::setlocale()> call, and within the scope of
960of a C<use locale> form that includes numerics, Perl obeys the
961C<LC_NUMERIC> locale information, which controls an application's idea
962of how numbers should be formatted for human readability.
b960a36e 963In most implementations the only effect is to
b4ffc3db 964change the character used for the decimal point--perhaps from "." to ",".
b960a36e 965The functions aren't aware of such niceties as thousands separation and
2095dafa 966so on. (See L<The localeconv function> if you care about these things.)
5a964f20 967
b960a36e
KW
968 use POSIX qw(strtod setlocale LC_NUMERIC);
969 use locale;
5f05dabc 970
b960a36e 971 setlocale LC_NUMERIC, "";
14280422 972
b960a36e 973 $n = 5/2; # Assign numeric 2.5 to $n
5f05dabc 974
b960a36e 975 $a = " $n"; # Locale-dependent conversion to string
5f05dabc 976
b960a36e 977 print "half five is $n\n"; # Locale-dependent output
5f05dabc 978
b960a36e 979 printf "half five is %g\n", $n; # Locale-dependent output
5f05dabc 980
b960a36e
KW
981 print "DECIMAL POINT IS COMMA\n"
982 if $n == (strtod("2,5"))[0]; # Locale-dependent conversion
5f05dabc 983
4bbcc6e8
JH
984See also L<I18N::Langinfo> and C<RADIXCHAR>.
985
cb88b78e 986=head2 Category C<LC_MONETARY>: Formatting of monetary amounts
5f05dabc 987
e199995e 988The C standard defines the C<LC_MONETARY> category, but not a function
5a964f20 989that is affected by its contents. (Those with experience of standards
b0c42ed9 990committees will recognize that the working group decided to punt on the
fa9b773e
KW
991issue.) Consequently, Perl essentially takes no notice of it. If you
992really want to use C<LC_MONETARY>, you can query its contents--see
e05ffc7d
KW
993L<The localeconv function>--and use the information that it returns in your
994application's own formatting of currency amounts. However, you may well
995find that the information, voluminous and complex though it may be, still
996does not quite meet your requirements: currency formatting is a hard nut
13a2d996 997to crack.
5f05dabc 998
4bbcc6e8
JH
999See also L<I18N::Langinfo> and C<CRNCYSTR>.
1000
ff52fcf1 1001=head2 Category C<LC_TIME>: Respresentation of time
5f05dabc 1002
39332f68 1003Output produced by C<POSIX::strftime()>, which builds a formatted
5f05dabc
PP
1004human-readable date/time string, is affected by the current C<LC_TIME>
1005locale. Thus, in a French locale, the output produced by the C<%B>
1006format element (full month name) for the first month of the year would
5a964f20 1007be "janvier". Here's how to get a list of long month names in the
5f05dabc
PP
1008current locale:
1009
1010 use POSIX qw(strftime);
14280422
DD
1011 for (0..11) {
1012 $long_month_name[$_] =
1013 strftime("%B", 0, 0, 0, 1, $_, 96);
5f05dabc
PP
1014 }
1015
2619d284
KW
1016Note: C<use locale> isn't needed in this example: C<strftime()> is a POSIX
1017function which uses the standard system-supplied C<libc> function that
1018always obeys the current C<LC_TIME> locale.
5f05dabc 1019
4bbcc6e8 1020See also L<I18N::Langinfo> and C<ABDAY_1>..C<ABDAY_7>, C<DAY_1>..C<DAY_7>,
2a2bf5f4 1021C<ABMON_1>..C<ABMON_12>, and C<ABMON_1>..C<ABMON_12>.
4bbcc6e8 1022
5f05dabc
PP
1023=head2 Other categories
1024
2619d284
KW
1025The remaining locale categories are not currently used by Perl itself.
1026But again note that things Perl interacts with may use these, including
1027extensions outside the standard Perl distribution, and by the
98a6f11e 1028operating system and its utilities. Note especially that the string
1029value of C<$!> and the error messages given by external utilities may
1030be changed by C<LC_MESSAGES>. If you want to have portable error
265f5c4a 1031codes, use C<%!>. See L<Errno>.
14280422
DD
1032
1033=head1 SECURITY
1034
5a964f20 1035Although the main discussion of Perl security issues can be found in
14280422
DD
1036L<perlsec>, a discussion of Perl's locale handling would be incomplete
1037if it did not draw your attention to locale-dependent security issues.
5a964f20
TC
1038Locales--particularly on systems that allow unprivileged users to
1039build their own locales--are untrustworthy. A malicious (or just plain
14280422
DD
1040broken) locale can make a locale-aware application give unexpected
1041results. Here are a few possibilities:
1042
1043=over 4
1044
1045=item *
1046
1047Regular expression checks for safe file names or mail addresses using
5a964f20 1048C<\w> may be spoofed by an C<LC_CTYPE> locale that claims that
dbf3c4d7 1049characters such as C<"E<gt>"> and C<"|"> are alphanumeric.
14280422
DD
1050
1051=item *
1052
e38874e2 1053String interpolation with case-mapping, as in, say, C<$dest =
cb88b78e 1054"C:\U$name.$ext">, may produce dangerous results if a bogus C<LC_CTYPE>
e38874e2
DD
1055case-mapping table is in effect.
1056
1057=item *
1058
14280422
DD
1059A sneaky C<LC_COLLATE> locale could result in the names of students with
1060"D" grades appearing ahead of those with "A"s.
1061
1062=item *
1063
5a964f20 1064An application that takes the trouble to use information in
14280422 1065C<LC_MONETARY> may format debits as if they were credits and vice versa
5a964f20 1066if that locale has been subverted. Or it might make payments in US
14280422
DD
1067dollars instead of Hong Kong dollars.
1068
1069=item *
1070
39332f68 1071The date and day names in dates formatted by C<strftime()> could be
14280422 1072manipulated to advantage by a malicious user able to subvert the
5a964f20 1073C<LC_DATE> locale. ("Look--it says I wasn't in the building on
14280422
DD
1074Sunday.")
1075
1076=back
1077
1078Such dangers are not peculiar to the locale system: any aspect of an
5a964f20 1079application's environment which may be modified maliciously presents
14280422 1080similar challenges. Similarly, they are not specific to Perl: any
5a964f20 1081programming language that allows you to write programs that take
14280422
DD
1082account of their environment exposes you to these issues.
1083
5a964f20
TC
1084Perl cannot protect you from all possibilities shown in the
1085examples--there is no substitute for your own vigilance--but, when
14280422 1086C<use locale> is in effect, Perl uses the tainting mechanism (see
5a964f20 1087L<perlsec>) to mark string results that become locale-dependent, and
14280422 1088which may be untrustworthy in consequence. Here is a summary of the
5a964f20 1089tainting behavior of operators and functions that may be affected by
14280422
DD
1090the locale:
1091
1092=over 4
1093
551e1d92
RB
1094=item *
1095
1096B<Comparison operators> (C<lt>, C<le>, C<ge>, C<gt> and C<cmp>):
14280422
DD
1097
1098Scalar true/false (or less/equal/greater) result is never tainted.
1099
551e1d92
RB
1100=item *
1101
1d2ab946 1102B<Case-mapping interpolation> (with C<\l>, C<\L>, C<\u>, C<\U>, or C<\F>)
e38874e2 1103
d6ded950
KW
1104The result string containing interpolated material is tainted if
1105a C<use locale> form that includes C<LC_CTYPE> is in effect.
e38874e2 1106
551e1d92
RB
1107=item *
1108
1109B<Matching operator> (C<m//>):
14280422
DD
1110
1111Scalar true/false result never tainted.
1112
1d2ab946 1113All subpatterns, either delivered as a list-context result or as C<$1>
d6ded950
KW
1114I<etc>., are tainted if a C<use locale> form that includes
1115C<LC_CTYPE> is in effect, and the subpattern
63baef57
KW
1116regular expression contains a locale-dependent construct. These
1117constructs include C<\w> (to match an alphanumeric character), C<\W>
1118(non-alphanumeric character), C<\b> and C<\B> (word-boundary and
1119non-boundardy, which depend on what C<\w> and C<\W> match), C<\s>
1120(whitespace character), C<\S> (non whitespace character), C<\d> and
1121C<\D> (digits and non-digits), and the POSIX character classes, such as
1122C<[:alpha:]> (see L<perlrecharclass/POSIX Character Classes>).
1123
1124Tainting is also likely if the pattern is to be matched
1125case-insensitively (via C</i>). The exception is if all the code points
1126to be matched this way are above 255 and do not have folds under Unicode
1127rules to below 256. Tainting is not done for these because Perl
1128only uses Unicode rules for such code points, and those rules are the
1129same no matter what the current locale.
1130
1d2ab946
KW
1131The matched-pattern variables, C<$&>, C<$`> (pre-match), C<$'>
1132(post-match), and C<$+> (last match) also are tainted.
14280422 1133
551e1d92
RB
1134=item *
1135
1136B<Substitution operator> (C<s///>):
14280422 1137
e38874e2 1138Has the same behavior as the match operator. Also, the left
d6ded950
KW
1139operand of C<=~> becomes tainted when a C<use locale>
1140form that includes C<LC_CTYPE> is in effect, if modified as
66cbab2c 1141a result of a substitution based on a regular
1d2ab946
KW
1142expression match involving any of the things mentioned in the previous
1143item, or of case-mapping, such as C<\l>, C<\L>,C<\u>, C<\U>, or C<\F>.
14280422 1144
551e1d92
RB
1145=item *
1146
39332f68 1147B<Output formatting functions> (C<printf()> and C<write()>):
14280422 1148
3cf03d68
JH
1149Results are never tainted because otherwise even output from print,
1150for example C<print(1/7)>, should be tainted if C<use locale> is in
1151effect.
14280422 1152
551e1d92
RB
1153=item *
1154
39332f68 1155B<Case-mapping functions> (C<lc()>, C<lcfirst()>, C<uc()>, C<ucfirst()>):
14280422 1156
d6ded950
KW
1157Results are tainted if a C<use locale> form that includes C<LC_CTYPE> is
1158in effect.
14280422 1159
551e1d92
RB
1160=item *
1161
39332f68
KW
1162B<POSIX locale-dependent functions> (C<localeconv()>, C<strcoll()>,
1163C<strftime()>, C<strxfrm()>):
14280422
DD
1164
1165Results are never tainted.
1166
14280422
DD
1167=back
1168
1169Three examples illustrate locale-dependent tainting.
1170The first program, which ignores its locale, won't run: a value taken
54310121 1171directly from the command line may not be used to name an output file
14280422
DD
1172when taint checks are enabled.
1173
1174 #/usr/local/bin/perl -T
1175 # Run with taint checking
1176
54310121 1177 # Command line sanity check omitted...
14280422
DD
1178 $tainted_output_file = shift;
1179
1180 open(F, ">$tainted_output_file")
3183d96c 1181 or warn "Open of $tainted_output_file failed: $!\n";
14280422
DD
1182
1183The program can be made to run by "laundering" the tainted value through
5a964f20
TC
1184a regular expression: the second example--which still ignores locale
1185information--runs, creating the file named on its command line
14280422
DD
1186if it can.
1187
1188 #/usr/local/bin/perl -T
1189
1190 $tainted_output_file = shift;
1191 $tainted_output_file =~ m%[\w/]+%;
1192 $untainted_output_file = $&;
1193
1194 open(F, ">$untainted_output_file")
1195 or warn "Open of $untainted_output_file failed: $!\n";
1196
5a964f20 1197Compare this with a similar but locale-aware program:
14280422
DD
1198
1199 #/usr/local/bin/perl -T
1200
1201 $tainted_output_file = shift;
1202 use locale;
1203 $tainted_output_file =~ m%[\w/]+%;
1204 $localized_output_file = $&;
1205
1206 open(F, ">$localized_output_file")
1207 or warn "Open of $localized_output_file failed: $!\n";
1208
1d2ab946 1209This third program fails to run because C<$&> is tainted: it is the result
5a964f20 1210of a match involving C<\w> while C<use locale> is in effect.
5f05dabc
PP
1211
1212=head1 ENVIRONMENT
1213
1214=over 12
1215
ee1ec05f
KW
1216=item PERL_SKIP_LOCALE_INIT
1217
c5c88224
KW
1218This environment variable, available starting in Perl v5.20, if set
1219(to any value), tells Perl to not use the rest of the
ee1ec05f
KW
1220environment variables to initialize with. Instead, Perl uses whatever
1221the current locale settings are. This is particularly useful in
1222embedded environments, see
1223L<perlembed/Using embedded Perl with POSIX locales>.
1224
5f05dabc
PP
1225=item PERL_BADLANG
1226
14280422 1227A string that can suppress Perl's warning about failed locale settings
54310121 1228at startup. Failure can occur if the locale support in the operating
5a964f20 1229system is lacking (broken) in some way--or if you mistyped the name of
900bd440 1230a locale when you set up your environment. If this environment
22ff3130
HS
1231variable is absent, or has a value other than "0" or "", Perl will
1232complain about locale setting failures.
5f05dabc 1233
cb88b78e 1234B<NOTE>: C<PERL_BADLANG> only gives you a way to hide the warning message.
14280422
DD
1235The message tells about some problem in your system's locale support,
1236and you should investigate what the problem is.
5f05dabc
PP
1237
1238=back
1239
1240The following environment variables are not specific to Perl: They are
39332f68 1241part of the standardized (ISO C, XPG4, POSIX 1.c) C<setlocale()> method
b385bb4d
KW
1242for controlling an application's opinion on data. Windows is non-POSIX,
1243but Perl arranges for the following to work as described anyway.
65ebb059
KW
1244If the locale given by an environment variable is not valid, Perl tries
1245the next lower one in priority. If none are valid, on Windows, the
1246system default locale is then tried. If all else fails, the C<"C">
1247locale is used. If even that doesn't work, something is badly broken,
c5e9a8e7 1248but Perl tries to forge ahead with whatever the locale settings might
65ebb059 1249be.
5f05dabc
PP
1250
1251=over 12
1252
cb88b78e 1253=item C<LC_ALL>
5f05dabc 1254
5a964f20 1255C<LC_ALL> is the "override-all" locale environment variable. If
5f05dabc
PP
1256set, it overrides all the rest of the locale environment variables.
1257
cb88b78e 1258=item C<LANGUAGE>
528d65ad
JH
1259
1260B<NOTE>: C<LANGUAGE> is a GNU extension, it affects you only if you
1261are using the GNU libc. This is the case if you are using e.g. Linux.
e1020413 1262If you are using "commercial" Unixes you are most probably I<not>
22b6f60d
JH
1263using GNU libc and you can ignore C<LANGUAGE>.
1264
1265However, in the case you are using C<LANGUAGE>: it affects the
1266language of informational, warning, and error messages output by
1267commands (in other words, it's like C<LC_MESSAGES>) but it has higher
96090e4f 1268priority than C<LC_ALL>. Moreover, it's not a single value but
22b6f60d
JH
1269instead a "path" (":"-separated list) of I<languages> (not locales).
1270See the GNU C<gettext> library documentation for more information.
528d65ad 1271
3ee1a09c 1272=item C<LC_CTYPE>
5f05dabc
PP
1273
1274In the absence of C<LC_ALL>, C<LC_CTYPE> chooses the character type
1275locale. In the absence of both C<LC_ALL> and C<LC_CTYPE>, C<LANG>
1276chooses the character type locale.
1277
cb88b78e 1278=item C<LC_COLLATE>
5f05dabc 1279
14280422
DD
1280In the absence of C<LC_ALL>, C<LC_COLLATE> chooses the collation
1281(sorting) locale. In the absence of both C<LC_ALL> and C<LC_COLLATE>,
1282C<LANG> chooses the collation locale.
5f05dabc 1283
cb88b78e 1284=item C<LC_MONETARY>
5f05dabc 1285
14280422
DD
1286In the absence of C<LC_ALL>, C<LC_MONETARY> chooses the monetary
1287formatting locale. In the absence of both C<LC_ALL> and C<LC_MONETARY>,
1288C<LANG> chooses the monetary formatting locale.
5f05dabc 1289
cb88b78e 1290=item C<LC_NUMERIC>
5f05dabc
PP
1291
1292In the absence of C<LC_ALL>, C<LC_NUMERIC> chooses the numeric format
1293locale. In the absence of both C<LC_ALL> and C<LC_NUMERIC>, C<LANG>
1294chooses the numeric format.
1295
cb88b78e 1296=item C<LC_TIME>
5f05dabc 1297
14280422
DD
1298In the absence of C<LC_ALL>, C<LC_TIME> chooses the date and time
1299formatting locale. In the absence of both C<LC_ALL> and C<LC_TIME>,
1300C<LANG> chooses the date and time formatting locale.
5f05dabc 1301
cb88b78e 1302=item C<LANG>
5f05dabc 1303
14280422
DD
1304C<LANG> is the "catch-all" locale environment variable. If it is set, it
1305is used as the last resort after the overall C<LC_ALL> and the
3ee1a09c 1306category-specific C<LC_I<foo>>.
5f05dabc
PP
1307
1308=back
1309
7e4353e9
RGS
1310=head2 Examples
1311
cb88b78e 1312The C<LC_NUMERIC> controls the numeric output:
7e4353e9 1313
ef3087ec
KW
1314 use locale;
1315 use POSIX qw(locale_h); # Imports setlocale() and the LC_ constants.
1316 setlocale(LC_NUMERIC, "fr_FR") or die "Pardon";
1317 printf "%g\n", 1.23; # If the "fr_FR" succeeded, probably shows 1,23.
7e4353e9 1318
39332f68 1319and also how strings are parsed by C<POSIX::strtod()> as numbers:
7e4353e9 1320
ef3087ec
KW
1321 use locale;
1322 use POSIX qw(locale_h strtod);
1323 setlocale(LC_NUMERIC, "de_DE") or die "Entschuldigung";
1324 my $x = strtod("2,34") + 5;
1325 print $x, "\n"; # Probably shows 7,34.
7e4353e9 1326
5f05dabc
PP
1327=head1 NOTES
1328
b960a36e
KW
1329=head2 String C<eval> and C<LC_NUMERIC>
1330
1331A string L<eval|perlfunc/eval EXPR> parses its expression as standard
1332Perl. It is therefore expecting the decimal point to be a dot. If
1333C<LC_NUMERIC> is set to have this be a comma instead, the parsing will
1334be confused, perhaps silently.
1335
1336 use locale;
1337 use POSIX qw(locale_h);
1338 setlocale(LC_NUMERIC, "fr_FR") or die "Pardon";
1339 my $a = 1.2;
1340 print eval "$a + 1.5";
1341 print "\n";
1342
1343prints C<13,5>. This is because in that locale, the comma is the
1344decimal point character. The C<eval> thus expands to:
1345
1346 eval "1,2 + 1.5"
1347
1348and the result is not what you likely expected. No warnings are
1349generated. If you do string C<eval>'s within the scope of
1350S<C<use locale>>, you should instead change the C<eval> line to do
1351something like:
1352
1353 print eval "no locale; $a + 1.5";
1354
1355This prints C<2.7>.
1356
d6ded950
KW
1357You could also exclude C<LC_NUMERIC>, if you don't need it, by
1358
1359 use locale ':!numeric';
1360
5f05dabc
PP
1361=head2 Backward compatibility
1362
b0c42ed9 1363Versions of Perl prior to 5.004 B<mostly> ignored locale information,
5a964f20
TC
1364generally behaving as if something similar to the C<"C"> locale were
1365always in force, even if the program environment suggested otherwise
1366(see L<The setlocale function>). By default, Perl still behaves this
1367way for backward compatibility. If you want a Perl application to pay
1368attention to locale information, you B<must> use the S<C<use locale>>
d6ded950 1369pragma (see L<The "use locale" pragma>) or, in the unlikely event
062ca197 1370that you want to do so for just pattern matching, the
70709c68
KW
1371C</l> regular expression modifier (see L<perlre/Character set
1372modifiers>) to instruct it to do so.
b0c42ed9
JH
1373
1374Versions of Perl from 5.002 to 5.003 did use the C<LC_CTYPE>
5a964f20
TC
1375information if available; that is, C<\w> did understand what
1376were the letters according to the locale environment variables.
b0c42ed9
JH
1377The problem was that the user had no control over the feature:
1378if the C library supported locales, Perl used them.
1379
1380=head2 I18N:Collate obsolete
1381
5a964f20 1382In versions of Perl prior to 5.004, per-locale collation was possible
b0c42ed9
JH
1383using the C<I18N::Collate> library module. This module is now mildly
1384obsolete and should be avoided in new applications. The C<LC_COLLATE>
1385functionality is now integrated into the Perl core language: One can
1386use locale-specific scalar data completely normally with C<use locale>,
1387so there is no longer any need to juggle with the scalar references of
1388C<I18N::Collate>.
5f05dabc 1389
14280422 1390=head2 Sort speed and memory use impacts
5f05dabc
PP
1391
1392Comparing and sorting by locale is usually slower than the default
14280422
DD
1393sorting; slow-downs of two to four times have been observed. It will
1394also consume more memory: once a Perl scalar variable has participated
1395in any string comparison or sorting operation obeying the locale
1396collation rules, it will take 3-15 times more memory than before. (The
1397exact multiplier depends on the string's contents, the operating system
1398and the locale.) These downsides are dictated more by the operating
1399system's implementation of the locale system than by Perl.
5f05dabc 1400
5f05dabc
PP
1401=head2 Freely available locale definitions
1402
66cbab2c
KW
1403The Unicode CLDR project extracts the POSIX portion of many of its
1404locales, available at
1405
e784ce68
KW
1406 http://unicode.org/Public/cldr/2.0.1/
1407
1408(Newer versions of CLDR require you to compute the POSIX data yourself.
1409See L<http://unicode.org/Public/cldr/latest/>.)
66cbab2c 1410
08d7a6b2
LB
1411There is a large collection of locale definitions at:
1412
1413 http://std.dkuug.dk/i18n/WG15-collection/locales/
1414
1415You should be aware that it is
14280422 1416unsupported, and is not claimed to be fit for any purpose. If your
5a964f20 1417system allows installation of arbitrary locales, you may find the
14280422
DD
1418definitions useful as they are, or as a basis for the development of
1419your own locales.
5f05dabc 1420
14280422 1421=head2 I18n and l10n
5f05dabc 1422
b0c42ed9
JH
1423"Internationalization" is often abbreviated as B<i18n> because its first
1424and last letters are separated by eighteen others. (You may guess why
1425the internalin ... internaliti ... i18n tends to get abbreviated.) In
1426the same way, "localization" is often abbreviated to B<l10n>.
14280422
DD
1427
1428=head2 An imperfect standard
1429
1430Internationalization, as defined in the C and POSIX standards, can be
1431criticized as incomplete, ungainly, and having too large a granularity.
1432(Locales apply to a whole process, when it would arguably be more useful
1433to have them apply to a single thread, window group, or whatever.) They
1434also have a tendency, like standards groups, to divide the world into
1435nations, when we all know that the world can equally well be divided
e199995e 1436into bankers, bikers, gamers, and so on.
5f05dabc 1437
b310b053
JH
1438=head1 Unicode and UTF-8
1439
7ee2ae1e 1440The support of Unicode is new starting from Perl version v5.6, and more fully
31f05a37
KW
1441implemented in versions v5.8 and later. See L<perluniintro>.
1442
a4a439fb
KW
1443Starting in Perl v5.20, UTF-8 locales are supported in Perl, except
1444C<LC_COLLATE> is only partially supported; collation support is improved
1445in Perl v5.26 to a level that may be sufficient for your needs
1446(see L</Category C<LC_COLLATE>: Collation: Text Comparisons and Sorting>).
1447
1448If you have Perl v5.16 or v5.18 and can't upgrade, you can use
66cbab2c
KW
1449
1450 use locale ':not_characters';
1451
1452When this form of the pragma is used, only the non-character portions of
1453locales are used by Perl, for example C<LC_NUMERIC>. Perl assumes that
1454you have translated all the characters it is to operate on into Unicode
1455(actually the platform's native character set (ASCII or EBCDIC) plus
1456Unicode). For data in files, this can conveniently be done by also
1457specifying
1458
1459 use open ':locale';
1460
1461This pragma arranges for all inputs from files to be translated into
1462Unicode from the current locale as specified in the environment (see
1463L</ENVIRONMENT>), and all outputs to files to be translated back
1464into the locale. (See L<open>). On a per-filehandle basis, you can
1465instead use the L<PerlIO::locale> module, or the L<Encode::Locale>
1466module, both available from CPAN. The latter module also has methods to
1467ease the handling of C<ARGV> and environment variables, and can be used
31f05a37 1468on individual strings. If you know that all your locales will be
66cbab2c
KW
1469UTF-8, as many are these days, you can use the L<B<-C>|perlrun/-C>
1470command line switch.
1471
1472This form of the pragma allows essentially seamless handling of locales
31f05a37 1473with Unicode. The collation order will be by Unicode code point order.
a4a439fb 1474L<Unicode::Collate> can be used to get Unicode rules collation.
66cbab2c 1475
31f05a37
KW
1476All the modules and switches just described can be used in v5.20 with
1477just plain C<use locale>, and, should the input locales not be UTF-8,
1478you'll get the less than ideal behavior, described below, that you get
1479with pre-v5.16 Perls, or when you use the locale pragma without the
1480C<:not_characters> parameter in v5.16 and v5.18. If you are using
1481exclusively UTF-8 locales in v5.20 and higher, the rest of this section
1482does not apply to you.
1483
1484There are two cases, multi-byte and single-byte locales. First
1485multi-byte:
1486
1487The only multi-byte (or wide character) locale that Perl is ever likely
1488to support is UTF-8. This is due to the difficulty of implementation,
1489the fact that high quality UTF-8 locales are now published for every
e784ce68
KW
1490area of the world (L<http://unicode.org/Public/cldr/2.0.1/> for
1491ones that are already set-up, but from an earlier version;
1492L<http://unicode.org/Public/cldr/latest/> for the most up-to-date, but
1493you have to extract the POSIX information yourself), and that
31f05a37
KW
1494failing all that you can use the L<Encode> module to translate to/from
1495your locale. So, you'll have to do one of those things if you're using
1496one of these locales, such as Big5 or Shift JIS. For UTF-8 locales, in
1497Perls (pre v5.20) that don't have full UTF-8 locale support, they may
1498work reasonably well (depending on your C library implementation)
1499simply because both
dc4bfc4b
KW
1500they and Perl store characters that take up multiple bytes the same way.
1501However, some, if not most, C library implementations may not process
1502the characters in the upper half of the Latin-1 range (128 - 255)
cb88b78e 1503properly under C<LC_CTYPE>. To see if a character is a particular type
dc4bfc4b
KW
1504under a locale, Perl uses the functions like C<isalnum()>. Your C
1505library may not work for UTF-8 locales with those functions, instead
dbf3c4d7
KW
1506only working under the newer wide library functions like C<iswalnum()>,
1507which Perl does not use.
1508These multi-byte locales are treated like single-byte locales, and will
1509have the restrictions described below. Starting in Perl v5.22 a warning
1510message is raised when Perl detects a multi-byte locale that it doesn't
1511fully support.
e199995e 1512
31f05a37 1513For single-byte locales,
e199995e 1514Perl generally takes the tack to use locale rules on code points that can fit
66cbab2c
KW
1515in a single byte, and Unicode rules for those that can't (though this
1516isn't uniformly applied, see the note at the end of this section). This
1517prevents many problems in locales that aren't UTF-8. Suppose the locale
1518is ISO8859-7, Greek. The character at 0xD7 there is a capital Chi. But
1519in the ISO8859-1 locale, Latin1, it is a multiplication sign. The POSIX
1520regular expression character class C<[[:alpha:]]> will magically match
15210xD7 in the Greek locale but not in the Latin one.
e199995e 1522
1d2ab946 1523However, there are places where this breaks down. Certain Perl constructs are
b4ffc3db
TC
1524for Unicode only, such as C<\p{Alpha}>. They assume that 0xD7 always has its
1525Unicode meaning (or the equivalent on EBCDIC platforms). Since Latin1 is a
1526subset of Unicode and 0xD7 is the multiplication sign in both Latin1 and
1527Unicode, C<\p{Alpha}> will never match it, regardless of locale. A similar
31f05a37
KW
1528issue occurs with C<\N{...}>. Prior to v5.20, It is therefore a bad
1529idea to use C<\p{}> or
66cbab2c 1530C<\N{}> under plain C<use locale>--I<unless> you can guarantee that the
dbf3c4d7 1531locale will be ISO8859-1. Use POSIX character classes instead.
66cbab2c
KW
1532
1533Another problem with this approach is that operations that cross the
1534single byte/multiple byte boundary are not well-defined, and so are
4a70680a 1535disallowed. (This boundary is between the codepoints at 255/256.)
66cbab2c
KW
1536For example, lower casing LATIN CAPITAL LETTER Y WITH DIAERESIS (U+0178)
1537should return LATIN SMALL LETTER Y WITH DIAERESIS (U+00FF). But in the
1538Greek locale, for example, there is no character at 0xFF, and Perl
1539has no way of knowing what the character at 0xFF is really supposed to
1540represent. Thus it disallows the operation. In this mode, the
1541lowercase of U+0178 is itself.
1542
1543The same problems ensue if you enable automatic UTF-8-ification of your
e199995e 1544standard file handles, default C<open()> layer, and C<@ARGV> on non-ISO8859-1,
b4ffc3db
TC
1545non-UTF-8 locales (by using either the B<-C> command line switch or the
1546C<PERL_UNICODE> environment variable; see L<perlrun>).
1547Things are read in as UTF-8, which would normally imply a Unicode
1548interpretation, but the presence of a locale causes them to be interpreted
1549in that locale instead. For example, a 0xD7 code point in the Unicode
1550input, which should mean the multiplication sign, won't be interpreted by
66cbab2c 1551Perl that way under the Greek locale. This is not a problem
b4ffc3db 1552I<provided> you make certain that all locales will always and only be either
66cbab2c 1553an ISO8859-1, or, if you don't have a deficient C library, a UTF-8 locale.
b4ffc3db 1554
1d2ab946
KW
1555Still another problem is that this approach can lead to two code
1556points meaning the same character. Thus in a Greek locale, both U+03A7
1557and U+00D7 are GREEK CAPITAL LETTER CHI.
1558
613abc6d
KW
1559Because of all these problems, starting in v5.22, Perl will raise a
1560warning if a multi-byte (hence Unicode) code point is used when a
1561single-byte locale is in effect. (Although it doesn't check for this if
1562doing so would unreasonably slow execution down.)
1563
b4ffc3db
TC
1564Vendor locales are notoriously buggy, and it is difficult for Perl to test
1565its locale-handling code because this interacts with code that Perl has no
1566control over; therefore the locale-handling code in Perl may be buggy as
66cbab2c
KW
1567well. (However, the Unicode-supplied locales should be better, and
1568there is a feed back mechanism to correct any problems. See
1569L</Freely available locale definitions>.)
1570
7ee2ae1e 1571If you have Perl v5.16, the problems mentioned above go away if you use
66cbab2c 1572the C<:not_characters> parameter to the locale pragma (except for vendor
7ee2ae1e 1573bugs in the non-character portions). If you don't have v5.16, and you
66cbab2c
KW
1574I<do> have locales that work, using them may be worthwhile for certain
1575specific purposes, as long as you keep in mind the gotchas already
1576mentioned. For example, if the collation for your locales works, it
1577runs faster under locales than under L<Unicode::Collate>; and you gain
1578access to such things as the local currency symbol and the names of the
7ee2ae1e 1579months and days of the week. (But to hammer home the point, in v5.16,
66cbab2c
KW
1580you get this access without the downsides of locales by using the
1581C<:not_characters> form of the pragma.)
1582
1583Note: The policy of using locale rules for code points that can fit in a
1584byte, and Unicode rules for those that can't is not uniformly applied.
7ee2ae1e 1585Pre-v5.12, it was somewhat haphazard; in v5.12 it was applied fairly
66cbab2c 1586consistently to regular expression matching except for bracketed
7ee2ae1e 1587character classes; in v5.14 it was extended to all regex matches; and in
663d437a 1588v5.16 to the casing operations such as C<\L> and C<uc()>. For
dbf3c4d7 1589collation, in all releases so far, the system's C<strxfrm()> function is
a4a439fb
KW
1590called, and whatever it does is what you get. Starting in v5.26, various
1591bugs are fixed with the way perl uses this function.
b310b053 1592
5f05dabc
PP
1593=head1 BUGS
1594
4e615abd
KW
1595=head2 Collation of strings containing embedded C<NUL> characters
1596
6696cfa7
KW
1597C<NUL> characters will sort the same as the lowest collating control
1598character does, or to C<"\001"> in the unlikely event that there are no
1599control characters at all in the locale. In cases where the strings
1600don't contain this non-C<NUL> control, the results will be correct, and
1601in many locales, this control, whatever it might be, will rarely be
1602encountered. But there are cases where a C<NUL> should sort before this
1603control, but doesn't. If two strings do collate identically, the one
1604containing the C<NUL> will sort to earlier.
4e615abd 1605
5f05dabc
PP
1606=head2 Broken systems
1607
5a964f20 1608In certain systems, the operating system's locale support
2bdf8add 1609is broken and cannot be fixed or used by Perl. Such deficiencies can
b4ffc3db 1610and will result in mysterious hangs and/or Perl core dumps when
2bdf8add 1611C<use locale> is in effect. When confronted with such a system,
7f2de2d2 1612please report in excruciating detail to <F<perlbug@perl.org>>, and
b4ffc3db 1613also contact your vendor: bug fixes may exist for these problems
2bdf8add 1614in your operating system. Sometimes such bug fixes are called an
83fb1bf0
KW
1615operating system upgrade. If you have the source for Perl, include in
1616the perlbug email the output of the test described above in L</Testing
1617for broken locales>.
5f05dabc
PP
1618
1619=head1 SEE ALSO
1620
b310b053
JH
1621L<I18N::Langinfo>, L<perluniintro>, L<perlunicode>, L<open>,
1622L<POSIX/isalnum>, L<POSIX/isalpha>,
4bbcc6e8
JH
1623L<POSIX/isdigit>, L<POSIX/isgraph>, L<POSIX/islower>,
1624L<POSIX/isprint>, L<POSIX/ispunct>, L<POSIX/isspace>,
1625L<POSIX/isupper>, L<POSIX/isxdigit>, L<POSIX/localeconv>,
1626L<POSIX/setlocale>, L<POSIX/strcoll>, L<POSIX/strftime>,
1627L<POSIX/strtod>, L<POSIX/strxfrm>.
5f05dabc 1628
ccd65d51
KW
1629For special considerations when Perl is embedded in a C program,
1630see L<perlembed/Using embedded Perl with POSIX locales>.
1631
5f05dabc
PP
1632=head1 HISTORY
1633
b0c42ed9 1634Jarkko Hietaniemi's original F<perli18n.pod> heavily hacked by Dominic
5a964f20 1635Dunlop, assisted by the perl5-porters. Prose worked over a bit by
c052850d 1636Tom Christiansen, and updated by Perl 5 porters.