This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlxs: Add text about dealing with locales, etc
[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
35UTF-8 locales, except for sorting and string comparisions. (Use
36L<Unicode::Collate> for these.) Perl continues to support the old
37non UTF-8 locales as well.
38
66cbab2c
KW
39(Unicode is also creating C<CLDR>, the "Common Locale Data Repository",
40L<http://cldr.unicode.org/> which includes more types of information than
41are available in the POSIX locale system. At the time of this writing,
42there was no CPAN module that provides access to this XML-encoded data.
43However, many of its locales have the POSIX-only data extracted, and are
31f05a37
KW
44available as UTF-8 locales at
45L<http://unicode.org/Public/cldr/latest/>.)
66cbab2c
KW
46
47=head1 WHAT IS A LOCALE
48
49A locale is a set of data that describes various aspects of how various
50communities in the world categorize their world. These categories are
51broken down into the following types (some of which include a brief
52note here):
53
54=over
55
cb88b78e 56=item Category C<LC_NUMERIC>: Numeric formatting
66cbab2c
KW
57
58This indicates how numbers should be formatted for human readability,
59for example the character used as the decimal point.
60
cb88b78e 61=item Category C<LC_MONETARY>: Formatting of monetary amounts
66cbab2c
KW
62
63=for comment
ebc3223b 64The nbsp below makes this look better (though not great)
66cbab2c
KW
65
66E<160>
67
cb88b78e 68=item Category C<LC_TIME>: Date/Time formatting
66cbab2c
KW
69
70=for comment
ebc3223b 71The nbsp below makes this look better (though not great)
66cbab2c
KW
72
73E<160>
74
cb88b78e 75=item Category C<LC_MESSAGES>: Error and other messages
66cbab2c 76
2619d284 77This is used by Perl itself only for accessing operating system error
03c702c5 78messages via L<$!|perlvar/$ERRNO> and L<$^E|perlvar/$EXTENDED_OS_ERROR>.
66cbab2c 79
cb88b78e 80=item Category C<LC_COLLATE>: Collation
66cbab2c 81
76073c88 82This indicates the ordering of letters for comparison and sorting.
66cbab2c
KW
83In Latin alphabets, for example, "b", generally follows "a".
84
cb88b78e 85=item Category C<LC_CTYPE>: Character Types
66cbab2c
KW
86
87This indicates, for example if a character is an uppercase letter.
88
2619d284
KW
89=item Other categories
90
91Some platforms have other categories, dealing with such things as
92measurement units and paper sizes. None of these are used directly by
93Perl, but outside operations that Perl interacts with may use
d6ded950 94these. See L</Not within the scope of "use locale"> below.
2619d284 95
66cbab2c
KW
96=back
97
2619d284
KW
98More details on the categories used by Perl are given below in L</LOCALE
99CATEGORIES>.
66cbab2c
KW
100
101Together, these categories go a long way towards being able to customize
102a single program to run in many different locations. But there are
103deficiencies, so keep reading.
5f05dabc
PP
104
105=head1 PREPARING TO USE LOCALES
106
2619d284
KW
107Perl itself will not use locales unless specifically requested to (but
108again note that Perl may interact with code that does use them). Even
109if there is such a request, B<all> of the following must be true
b960a36e 110for it to work properly:
5f05dabc
PP
111
112=over 4
113
114=item *
115
116B<Your operating system must support the locale system>. If it does,
39332f68 117you should find that the C<setlocale()> function is a documented part of
5f05dabc
PP
118its C library.
119
120=item *
121
5a964f20 122B<Definitions for locales that you use must be installed>. You, or
14280422
DD
123your system administrator, must make sure that this is the case. The
124available locales, the location in which they are kept, and the manner
5a964f20
TC
125in which they are installed all vary from system to system. Some systems
126provide only a few, hard-wired locales and do not allow more to be
127added. Others allow you to add "canned" locales provided by the system
128supplier. Still others allow you or the system administrator to define
14280422 129and add arbitrary locales. (You may have to ask your supplier to
5a964f20 130provide canned locales that are not delivered with your operating
14280422 131system.) Read your system documentation for further illumination.
5f05dabc
PP
132
133=item *
134
135B<Perl must believe that the locale system is supported>. If it does,
136C<perl -V:d_setlocale> will say that the value for C<d_setlocale> is
137C<define>.
138
139=back
140
141If you want a Perl application to process and present your data
142according to a particular locale, the application code should include
d6ded950 143the S<C<use locale>> pragma (see L<The "use locale" pragma>) where
5f05dabc
PP
144appropriate, and B<at least one> of the following must be true:
145
146=over 4
147
c052850d 148=item 1
5f05dabc 149
66cbab2c 150B<The locale-determining environment variables (see L</"ENVIRONMENT">)
5a964f20 151must be correctly set up> at the time the application is started, either
ef3087ec 152by yourself or by whomever set up your system account; or
5f05dabc 153
c052850d 154=item 2
5f05dabc 155
14280422
DD
156B<The application must set its own locale> using the method described in
157L<The setlocale function>.
5f05dabc
PP
158
159=back
160
161=head1 USING LOCALES
162
d6ded950 163=head2 The C<"use locale"> pragma
5f05dabc 164
2619d284 165By default, Perl itself ignores the current locale. The S<C<use locale>>
66cbab2c 166pragma tells Perl to use the current locale for some operations.
d6ded950
KW
167Starting in v5.16, there are optional parameters to this pragma,
168described below, which restrict which operations are affected by it.
c052850d
KW
169
170The current locale is set at execution time by
171L<setlocale()|/The setlocale function> described below. If that function
172hasn't yet been called in the course of the program's execution, the
66cbab2c 173current locale is that which was determined by the L</"ENVIRONMENT"> in
ebc3223b 174effect at the start of the program.
dfcc8045 175If there is no valid environment, the current locale is whatever the
65ebb059
KW
176system default has been set to. On POSIX systems, it is likely, but
177not necessarily, the "C" locale. On Windows, the default is set via the
178computer's S<C<Control Panel-E<gt>Regional and Language Options>> (or its
179current equivalent).
c052850d
KW
180
181The operations that are affected by locale are:
5f05dabc
PP
182
183=over 4
184
d6ded950 185=item B<Not within the scope of C<"use locale">>
b960a36e 186
d6ded950
KW
187Only certain operations originating outside Perl should be affected, as
188follows:
1d2ab946
KW
189
190=over 4
191
192=item *
2619d284 193
2619d284
KW
194The current locale is also used when going outside of Perl with
195operations like L<system()|perlfunc/system LIST> or
196L<qxE<sol>E<sol>|perlop/qxE<sol>STRINGE<sol>>, if those operations are
197locale-sensitive.
198
1d2ab946
KW
199=item *
200
2619d284
KW
201Also Perl gives access to various C library functions through the
202L<POSIX> module. Some of those functions are always affected by the
203current locale. For example, C<POSIX::strftime()> uses C<LC_TIME>;
204C<POSIX::strtod()> uses C<LC_NUMERIC>; C<POSIX::strcoll()> and
205C<POSIX::strxfrm()> use C<LC_COLLATE>; and character classification
206functions like C<POSIX::isalnum()> use C<LC_CTYPE>. All such functions
207will behave according to the current underlying locale, even if that
1d2ab946 208locale isn't exposed to Perl space.
2619d284 209
1d2ab946
KW
210=item *
211
52686f2a
KW
212XS modules for all categories but C<LC_NUMERIC> get the underlying
213locale, and hence any C library functions they call will use that
8a384d3a 214underlying locale. For more discussion, see L<perlxs/CAVEATS>.
1d2ab946
KW
215
216=back
217
ebc3223b
KW
218=for comment
219The nbsp below makes this look better (though not great)
220
221E<160>
222
223=item B<Lingering effects of C<S<use locale>>>
1d2ab946
KW
224
225Certain Perl operations that are set-up within the scope of a
d6ded950 226C<use locale> retain that effect even outside the scope.
b960a36e
KW
227These include:
228
229=over 4
230
231=item *
232
233The output format of a L<write()|perlfunc/write> is determined by an
234earlier format declaration (L<perlfunc/format>), so whether or not the
235output is affected by locale is determined by if the C<format()> is
d6ded950 236within the scope of a C<use locale>, not whether the C<write()>
b960a36e
KW
237is.
238
239=item *
240
241Regular expression patterns can be compiled using
242L<qrE<sol>E<sol>|perlop/qrE<sol>STRINGE<sol>msixpodual> with actual
243matching deferred to later. Again, it is whether or not the compilation
244was done within the scope of C<use locale> that determines the match
245behavior, not if the matches are done within such a scope or not.
246
247=back
248
ebc3223b
KW
249=for comment
250The nbsp below makes this look better (though not great)
251
d6ded950 252
ebc3223b
KW
253E<160>
254
d6ded950 255=item B<Under C<"use locale";>>
66cbab2c
KW
256
257=over 4
258
259=item *
260
d6ded950 261All the above operations
b960a36e
KW
262
263=item *
264
265B<Format declarations> (L<perlfunc/format>) and hence any subsequent
266C<write()>s use C<LC_NUMERIC>.
66cbab2c
KW
267
268=item *
269
b960a36e
KW
270B<stringification and output> use C<LC_NUMERIC>.
271These include the results of
272C<print()>,
273C<printf()>,
274C<say()>,
275and
276C<sprintf()>.
66cbab2c 277
b960a36e
KW
278=item *
279
280B<The comparison operators> (C<lt>, C<le>, C<cmp>, C<ge>, and C<gt>) use
39332f68 281C<LC_COLLATE>. C<sort()> is also affected if used without an
5a964f20 282explicit comparison function, because it uses C<cmp> by default.
14280422 283
5a964f20 284B<Note:> C<eq> and C<ne> are unaffected by locale: they always
de108802 285perform a char-by-char comparison of their scalar operands. What's
14280422
DD
286more, if C<cmp> finds that its operands are equal according to the
287collation sequence specified by the current locale, it goes on to
de108802
RGS
288perform a char-by-char comparison, and only returns I<0> (equal) if the
289operands are char-for-char identical. If you really want to know whether
5a964f20 290two strings--which C<eq> and C<cmp> may consider different--are equal
14280422 291as far as collation in the locale is concerned, see the discussion in
cb88b78e 292L<Category C<LC_COLLATE>: Collation>.
5f05dabc
PP
293
294=item *
295
39332f68
KW
296B<Regular expressions and case-modification functions> (C<uc()>, C<lc()>,
297C<ucfirst()>, and C<lcfirst()>) use C<LC_CTYPE>
5f05dabc 298
2c6ee1a7
KW
299=item *
300
301The variables L<$!|perlvar/$ERRNO> (and its synonyms C<$ERRNO> and
302C<$OS_ERROR>) and L<$^E|perlvar/$EXTENDED_OS_ERROR> (and its synonym
303C<$EXTENDED_OS_ERROR>) when used as strings use C<LC_MESSAGES>.
304
5f05dabc
PP
305=back
306
66cbab2c 307=back
5f05dabc 308
5a964f20 309The default behavior is restored with the S<C<no locale>> pragma, or
ef3087ec 310upon reaching the end of the block enclosing C<use locale>.
66cbab2c
KW
311Note that C<use locale> and C<use locale ':not_characters'> may be
312nested, and that what is in effect within an inner scope will revert to
313the outer scope's rules at the end of the inner scope.
5f05dabc 314
5a964f20 315The string result of any operation that uses locale
14280422
DD
316information is tainted, as it is possible for a locale to be
317untrustworthy. See L<"SECURITY">.
5f05dabc 318
d6ded950
KW
319Starting in Perl v5.16 in a very limited way, and more generally in
320v5.22, you can restrict which category or categories are enabled by this
321particular instance of the pragma by adding parameters to it. For
322example,
323
324 use locale qw(:ctype :numeric);
325
326enables locale awareness within its scope of only those operations
327(listed above) that are affected by C<LC_CTYPE> and C<LC_NUMERIC>.
328
329The possible categories are: C<:collate>, C<:ctype>, C<:messages>,
330C<:monetary>, C<:numeric>, C<:time>, and the pseudo category
331C<:characters> (described below).
332
333Thus you can say
334
335 use locale ':messages';
336
337and only L<$!|perlvar/$ERRNO> and L<$^E|perlvar/$EXTENDED_OS_ERROR>
338will be locale aware. Everything else is unaffected.
339
340Since Perl doesn't currently do anything with the C<LC_MONETARY>
341category, specifying C<:monetary> does effectively nothing. Some
342systems have other categories, such as C<LC_PAPER_SIZE>, but Perl
343also doesn't know anything about them, and there is no way to specify
344them in this pragma's arguments.
345
346You can also easily say to use all categories but one, by either, for
347example,
348
349 use locale ':!ctype';
350 use locale ':not_ctype';
351
352both of which mean to enable locale awarness of all categories but
353C<LC_CTYPE>. Only one category argument may be specified in a
354S<C<use locale>> if it is of the negated form.
355
356Prior to v5.22 only one form of the pragma with arguments is available:
357
358 use locale ':not_characters';
359
360(and you have to say C<not_>; you can't use the bang C<!> form). This
361pseudo category is a shorthand for specifying both C<:collate> and
362C<:ctype>. Hence, in the negated form, it is nearly the same thing as
363saying
364
365 use locale qw(:messages :monetary :numeric :time);
366
367We use the term "nearly", because C<:not_characters> also turns on
368S<C<use feature 'unicode_strings'>> within its scope. This form is
369less useful in v5.20 and later, and is described fully in
370L</Unicode and UTF-8>, but briefly, it tells Perl to not use the
371character portions of the locale definition, that is the C<LC_CTYPE> and
372C<LC_COLLATE> categories. Instead it will use the native character set
373(extended by Unicode). When using this parameter, you are responsible
374for getting the external character set translated into the
375native/Unicode one (which it already will be if it is one of the
376increasingly popular UTF-8 locales). There are convenient ways of doing
377this, as described in L</Unicode and UTF-8>.
378
5f05dabc
PP
379=head2 The setlocale function
380
14280422 381You can switch locales as often as you wish at run time with the
39332f68 382C<POSIX::setlocale()> function:
5f05dabc 383
5f05dabc
PP
384 # Import locale-handling tool set from POSIX module.
385 # This example uses: setlocale -- the function call
386 # LC_CTYPE -- explained below
2619d284
KW
387 # (Showing the testing for success/failure of operations is
388 # omitted in these examples to avoid distracting from the main
ebc3223b 389 # point)
6ea81ccf 390
5f05dabc 391 use POSIX qw(locale_h);
dfcc8045
KW
392 use locale;
393 my $old_locale;
5f05dabc 394
14280422 395 # query and save the old locale
5f05dabc
PP
396 $old_locale = setlocale(LC_CTYPE);
397
398 setlocale(LC_CTYPE, "fr_CA.ISO8859-1");
399 # LC_CTYPE now in locale "French, Canada, codeset ISO 8859-1"
400
401 setlocale(LC_CTYPE, "");
65ebb059
KW
402 # LC_CTYPE now reset to the default defined by the
403 # LC_ALL/LC_CTYPE/LANG environment variables, or to the system
404 # default. See below for documentation.
5f05dabc
PP
405
406 # restore the old locale
407 setlocale(LC_CTYPE, $old_locale);
408
39332f68 409The first argument of C<setlocale()> gives the B<category>, the second the
14280422
DD
410B<locale>. The category tells in what aspect of data processing you
411want to apply locale-specific rules. Category names are discussed in
66cbab2c 412L</LOCALE CATEGORIES> and L</"ENVIRONMENT">. The locale is the name of a
14280422
DD
413collection of customization information corresponding to a particular
414combination of language, country or territory, and codeset. Read on for
415hints on the naming of locales: not all systems name locales as in the
416example.
417
39332f68 418If no second argument is provided and the category is something other
cb88b78e 419than C<LC_ALL>, the function returns a string naming the current locale
502a173a 420for the category. You can use this value as the second argument in a
f170b852
KW
421subsequent call to C<setlocale()>, B<but> on some platforms the string
422is opaque, not something that most people would be able to decipher as
423to what locale it means.
502a173a 424
cb88b78e 425If no second argument is provided and the category is C<LC_ALL>, the
502a173a 426result is implementation-dependent. It may be a string of
c052850d 427concatenated locale names (separator also implementation-dependent)
39332f68 428or a single locale name. Please consult your L<setlocale(3)> man page for
502a173a
JH
429details.
430
431If a second argument is given and it corresponds to a valid locale,
432the locale for the category is set to that value, and the function
433returns the now-current locale value. You can then use this in yet
39332f68 434another call to C<setlocale()>. (In some implementations, the return
502a173a
JH
435value may sometimes differ from the value you gave as the second
436argument--think of it as an alias for the value you gave.)
5f05dabc
PP
437
438As the example shows, if the second argument is an empty string, the
439category's locale is returned to the default specified by the
440corresponding environment variables. Generally, this results in a
5a964f20 441return to the default that was in force when Perl started up: changes
54310121 442to the environment made by the application after startup may or may not
5a964f20 443be noticed, depending on your system's C library.
5f05dabc 444
d6ded950
KW
445Note that when a form of C<use locale> that doesn't include all
446categories is specified, Perl ignores the excluded categories.
66cbab2c 447
f170b852 448If C<set_locale()> fails for some reason (for example, an attempt to set
dfcc8045
KW
449to a locale unknown to the system), the locale for the category is not
450changed, and the function returns C<undef>.
451
2619d284 452
39332f68 453For further information about the categories, consult L<setlocale(3)>.
3e6e419a
JH
454
455=head2 Finding locales
456
39332f68 457For locales available in your system, consult also L<setlocale(3)> to
5a964f20
TC
458see whether it leads to the list of available locales (search for the
459I<SEE ALSO> section). If that fails, try the following command lines:
5f05dabc
PP
460
461 locale -a
462
463 nlsinfo
464
465 ls /usr/lib/nls/loc
466
467 ls /usr/lib/locale
468
469 ls /usr/lib/nls
470
b478f28d
JH
471 ls /usr/share/locale
472
5f05dabc
PP
473and see whether they list something resembling these
474
2bdf8add 475 en_US.ISO8859-1 de_DE.ISO8859-1 ru_RU.ISO8859-5
502a173a 476 en_US.iso88591 de_DE.iso88591 ru_RU.iso88595
2bdf8add 477 en_US de_DE ru_RU
14280422 478 en de ru
2bdf8add
JH
479 english german russian
480 english.iso88591 german.iso88591 russian.iso88595
502a173a 481 english.roman8 russian.koi8r
5f05dabc 482
39332f68 483Sadly, even though the calling interface for C<setlocale()> has been
528d65ad 484standardized, names of locales and the directories where the
5a964f20 485configuration resides have not been. The basic form of the name is
528d65ad
JH
486I<language_territory>B<.>I<codeset>, but the latter parts after
487I<language> are not always present. The I<language> and I<country>
488are usually from the standards B<ISO 3166> and B<ISO 639>, the
489two-letter abbreviations for the countries and the languages of the
490world, respectively. The I<codeset> part often mentions some B<ISO
4918859> character set, the Latin codesets. For example, C<ISO 8859-1>
492is the so-called "Western European codeset" that can be used to encode
493most Western European languages adequately. Again, there are several
494ways to write even the name of that one standard. Lamentably.
5f05dabc 495
14280422
DD
496Two special locales are worth particular mention: "C" and "POSIX".
497Currently these are effectively the same locale: the difference is
5a964f20
TC
498mainly that the first one is defined by the C standard, the second by
499the POSIX standard. They define the B<default locale> in which
14280422 500every program starts in the absence of locale information in its
5a964f20 501environment. (The I<default> default locale, if you will.) Its language
39332f68
KW
502is (American) English and its character codeset ASCII or, rarely, a
503superset thereof (such as the "DEC Multinational Character Set
504(DEC-MCS)"). B<Warning>. The C locale delivered by some vendors
505may not actually exactly match what the C standard calls for. So
506beware.
5f05dabc 507
14280422
DD
508B<NOTE>: Not all systems have the "POSIX" locale (not all systems are
509POSIX-conformant), so use "C" when you need explicitly to specify this
510default locale.
5f05dabc 511
3e6e419a
JH
512=head2 LOCALE PROBLEMS
513
5a964f20 514You may encounter the following warning message at Perl startup:
3e6e419a
JH
515
516 perl: warning: Setting locale failed.
517 perl: warning: Please check that your locale settings:
518 LC_ALL = "En_US",
519 LANG = (unset)
520 are supported and installed on your system.
521 perl: warning: Falling back to the standard locale ("C").
522
cb88b78e 523This means that your locale settings had C<LC_ALL> set to "En_US" and
5a964f20
TC
524LANG exists but has no value. Perl tried to believe you but could not.
525Instead, Perl gave up and fell back to the "C" locale, the default locale
65ebb059
KW
526that is supposed to work no matter what. (On Windows, it first tries
527falling back to the system default locale.) This usually means your
528locale settings were wrong, they mention locales your system has never
529heard of, or the locale installation in your system has problems (for
530example, some system files are broken or missing). There are quick and
531temporary fixes to these problems, as well as more thorough and lasting
532fixes.
3e6e419a 533
83fb1bf0
KW
534=head2 Testing for broken locales
535
536If you are building Perl from source, the Perl test suite file
537F<lib/locale.t> can be used to test the locales on your system.
538Setting the environment variable C<PERL_DEBUG_FULL_TEST> to 1
539will cause it to output detailed results. For example, on Linux, you
540could say
541
1d2ab946 542 PERL_DEBUG_FULL_TEST=1 ./perl -T -Ilib lib/locale.t > locale.log 2>&1
83fb1bf0
KW
543
544Besides many other tests, it will test every locale it finds on your
545system to see if they conform to the POSIX standard. If any have
546errors, it will include a summary near the end of the output of which
547locales passed all its tests, and which failed, and why.
548
3e6e419a
JH
549=head2 Temporarily fixing locale problems
550
5a964f20 551The two quickest fixes are either to render Perl silent about any
3e6e419a
JH
552locale inconsistencies or to run Perl under the default locale "C".
553
554Perl's moaning about locale problems can be silenced by setting the
cb88b78e 555environment variable C<PERL_BADLANG> to a zero value, for example "0".
900bd440
JH
556This method really just sweeps the problem under the carpet: you tell
557Perl to shut up even when Perl sees that something is wrong. Do not
558be surprised if later something locale-dependent misbehaves.
3e6e419a
JH
559
560Perl can be run under the "C" locale by setting the environment
cb88b78e
KW
561variable C<LC_ALL> to "C". This method is perhaps a bit more civilized
562than the C<PERL_BADLANG> approach, but setting C<LC_ALL> (or
5a964f20
TC
563other locale variables) may affect other programs as well, not just
564Perl. In particular, external programs run from within Perl will see
3e6e419a 565these changes. If you make the new settings permanent (read on), all
f979aebc 566programs you run see the changes. See L<"ENVIRONMENT"> for
5a964f20 567the full list of relevant environment variables and L<USING LOCALES>
e05ffc7d 568for their effects in Perl. Effects in other programs are
cb88b78e 569easily deducible. For example, the variable C<LC_COLLATE> may well affect
b432a672 570your B<sort> program (or whatever the program that arranges "records"
3e6e419a
JH
571alphabetically in your system is called).
572
5a964f20
TC
573You can test out changing these variables temporarily, and if the
574new settings seem to help, put those settings into your shell startup
575files. Consult your local documentation for the exact details. For in
576Bourne-like shells (B<sh>, B<ksh>, B<bash>, B<zsh>):
3e6e419a
JH
577
578 LC_ALL=en_US.ISO8859-1
579 export LC_ALL
580
5a964f20
TC
581This assumes that we saw the locale "en_US.ISO8859-1" using the commands
582discussed above. We decided to try that instead of the above faulty
583locale "En_US"--and in Cshish shells (B<csh>, B<tcsh>)
3e6e419a
JH
584
585 setenv LC_ALL en_US.ISO8859-1
c47ff5f1 586
c406981e
JH
587or if you have the "env" application you can do in any shell
588
589 env LC_ALL=en_US.ISO8859-1 perl ...
590
5a964f20 591If you do not know what shell you have, consult your local
3e6e419a
JH
592helpdesk or the equivalent.
593
594=head2 Permanently fixing locale problems
595
5a964f20
TC
596The slower but superior fixes are when you may be able to yourself
597fix the misconfiguration of your own environment variables. The
3e6e419a
JH
598mis(sing)configuration of the whole system's locales usually requires
599the help of your friendly system administrator.
600
5a964f20
TC
601First, see earlier in this document about L<Finding locales>. That tells
602how to find which locales are really supported--and more importantly,
603installed--on your system. In our example error message, environment
604variables affecting the locale are listed in the order of decreasing
605importance (and unset variables do not matter). Therefore, having
606LC_ALL set to "En_US" must have been the bad choice, as shown by the
607error message. First try fixing locale settings listed first.
3e6e419a 608
5a964f20
TC
609Second, if using the listed commands you see something B<exactly>
610(prefix matches do not count and case usually counts) like "En_US"
611without the quotes, then you should be okay because you are using a
612locale name that should be installed and available in your system.
4a4eefd0 613In this case, see L<Permanently fixing your system's locale configuration>.
3e6e419a 614
4a4eefd0 615=head2 Permanently fixing your system's locale configuration
3e6e419a 616
5a964f20 617This is when you see something like:
3e6e419a
JH
618
619 perl: warning: Please check that your locale settings:
620 LC_ALL = "En_US",
621 LANG = (unset)
622 are supported and installed on your system.
623
624but then cannot see that "En_US" listed by the above-mentioned
5a964f20
TC
625commands. You may see things like "en_US.ISO8859-1", but that isn't
626the same. In this case, try running under a locale
627that you can list and which somehow matches what you tried. The
3e6e419a 628rules for matching locale names are a bit vague because
e05ffc7d 629standardization is weak in this area. See again the
13a2d996 630L<Finding locales> about general rules.
3e6e419a 631
b687b08b 632=head2 Fixing system locale configuration
3e6e419a 633
5a964f20
TC
634Contact a system administrator (preferably your own) and report the exact
635error message you get, and ask them to read this same documentation you
636are now reading. They should be able to check whether there is something
637wrong with the locale configuration of the system. The L<Finding locales>
638section is unfortunately a bit vague about the exact commands and places
639because these things are not that standardized.
3e6e419a 640
5f05dabc
PP
641=head2 The localeconv function
642
39332f68 643The C<POSIX::localeconv()> function allows you to get particulars of the
14280422 644locale-dependent numeric formatting information specified by the current
a835cd47
KW
645underlying C<LC_NUMERIC> and C<LC_MONETARY> locales (regardless of
646whether called from within the scope of C<S<use locale>> or not). (If
647you just want the name of
39332f68 648the current locale for a particular category, use C<POSIX::setlocale()>
5a964f20 649with a single parameter--see L<The setlocale function>.)
5f05dabc
PP
650
651 use POSIX qw(locale_h);
5f05dabc
PP
652
653 # Get a reference to a hash of locale-dependent info
654 $locale_values = localeconv();
655
656 # Output sorted list of the values
657 for (sort keys %$locale_values) {
14280422 658 printf "%-20s = %s\n", $_, $locale_values->{$_}
5f05dabc
PP
659 }
660
39332f68 661C<localeconv()> takes no arguments, and returns B<a reference to> a hash.
5a964f20 662The keys of this hash are variable names for formatting, such as
502a173a 663C<decimal_point> and C<thousands_sep>. The values are the
cea6626f 664corresponding, er, values. See L<POSIX/localeconv> for a longer
502a173a
JH
665example listing the categories an implementation might be expected to
666provide; some provide more and others fewer. You don't need an
39332f68 667explicit C<use locale>, because C<localeconv()> always observes the
502a173a 668current locale.
5f05dabc 669
5a964f20
TC
670Here's a simple-minded example program that rewrites its command-line
671parameters as integers correctly formatted in the current locale:
5f05dabc 672
ef3087ec
KW
673 use POSIX qw(locale_h);
674
675 # Get some of locale's numeric formatting parameters
676 my ($thousands_sep, $grouping) =
677 @{localeconv()}{'thousands_sep', 'grouping'};
678
679 # Apply defaults if values are missing
680 $thousands_sep = ',' unless $thousands_sep;
681
682 # grouping and mon_grouping are packed lists
683 # of small integers (characters) telling the
684 # grouping (thousand_seps and mon_thousand_seps
685 # being the group dividers) of numbers and
686 # monetary quantities. The integers' meanings:
687 # 255 means no more grouping, 0 means repeat
688 # the previous grouping, 1-254 means use that
689 # as the current grouping. Grouping goes from
690 # right to left (low to high digits). In the
691 # below we cheat slightly by never using anything
692 # else than the first grouping (whatever that is).
693 if ($grouping) {
694 @grouping = unpack("C*", $grouping);
695 } else {
696 @grouping = (3);
697 }
698
699 # Format command line params for current locale
700 for (@ARGV) {
701 $_ = int; # Chop non-integer part
702 1 while
703 s/(\d)(\d{$grouping[0]}($|$thousands_sep))/$1$thousands_sep$2/;
704 print "$_";
705 }
706 print "\n";
5f05dabc 707
03ceeedf
KW
708Note that if the platform doesn't have C<LC_NUMERIC> and/or
709C<LC_MONETARY> available or enabled, the corresponding elements of the
710hash will be missing.
711
74c76037 712=head2 I18N::Langinfo
4bbcc6e8
JH
713
714Another interface for querying locale-dependent information is the
39332f68 715C<I18N::Langinfo::langinfo()> function, available at least in Unix-like
4bbcc6e8
JH
716systems and VMS.
717
39332f68
KW
718The following example will import the C<langinfo()> function itself and
719three constants to be used as arguments to C<langinfo()>: a constant for
74c76037
JH
720the abbreviated first day of the week (the numbering starts from
721Sunday = 1) and two more constants for the affirmative and negative
722answers for a yes/no question in the current locale.
4bbcc6e8 723
74c76037 724 use I18N::Langinfo qw(langinfo ABDAY_1 YESSTR NOSTR);
4bbcc6e8 725
ef3087ec
KW
726 my ($abday_1, $yesstr, $nostr)
727 = map { langinfo } qw(ABDAY_1 YESSTR NOSTR);
4bbcc6e8 728
74c76037 729 print "$abday_1? [$yesstr/$nostr] ";
4bbcc6e8 730
74c76037
JH
731In other words, in the "C" (or English) locale the above will probably
732print something like:
733
e05ffc7d 734 Sun? [yes/no]
4bbcc6e8
JH
735
736See L<I18N::Langinfo> for more information.
737
5f05dabc
PP
738=head1 LOCALE CATEGORIES
739
5a964f20
TC
740The following subsections describe basic locale categories. Beyond these,
741some combination categories allow manipulation of more than one
742basic category at a time. See L<"ENVIRONMENT"> for a discussion of these.
5f05dabc 743
cb88b78e 744=head2 Category C<LC_COLLATE>: Collation
5f05dabc 745
d6ded950
KW
746In the scope of a S<C<use locale>> form that includes collation, Perl
747looks to the C<LC_COLLATE>
5a964f20 748environment variable to determine the application's notions on collation
b4ffc3db
TC
749(ordering) of characters. For example, "b" follows "a" in Latin
750alphabets, but where do "E<aacute>" and "E<aring>" belong? And while
f87fa335 751"color" follows "chocolate" in English, what about in traditional Spanish?
5f05dabc 752
60f0fa02
JH
753The following collations all make sense and you may meet any of them
754if you "use locale".
755
756 A B C D E a b c d e
35316ca3 757 A a B b C c D d E e
60f0fa02
JH
758 a A b B c C d D e E
759 a b c d e A B C D E
760
f1cbbd6e 761Here is a code snippet to tell what "word"
5a964f20 762characters are in the current locale, in that locale's order:
5f05dabc
PP
763
764 use locale;
35316ca3 765 print +(sort grep /\w/, map { chr } 0..255), "\n";
5f05dabc 766
14280422
DD
767Compare this with the characters that you see and their order if you
768state explicitly that the locale should be ignored:
5f05dabc
PP
769
770 no locale;
35316ca3 771 print +(sort grep /\w/, map { chr } 0..255), "\n";
5f05dabc
PP
772
773This machine-native collation (which is what you get unless S<C<use
774locale>> has appeared earlier in the same block) must be used for
775sorting raw binary data, whereas the locale-dependent collation of the
b0c42ed9 776first example is useful for natural text.
5f05dabc 777
14280422
DD
778As noted in L<USING LOCALES>, C<cmp> compares according to the current
779collation locale when C<use locale> is in effect, but falls back to a
de108802 780char-by-char comparison for strings that the locale says are equal. You
39332f68 781can use C<POSIX::strcoll()> if you don't want this fall-back:
14280422
DD
782
783 use POSIX qw(strcoll);
784 $equal_in_locale =
785 !strcoll("space and case ignored", "SpaceAndCaseIgnored");
786
39332f68 787C<$equal_in_locale> will be true if the collation locale specifies a
5a964f20 788dictionary-like ordering that ignores space characters completely and
9e3a2af8 789which folds case.
14280422 790
31f05a37
KW
791Perl only supports single-byte locales for C<LC_COLLATE>. This means
792that a UTF-8 locale likely will just give you machine-native ordering.
793Use L<Unicode::Collate> for the full implementation of the Unicode
794Collation Algorithm.
795
5a964f20 796If you have a single string that you want to check for "equality in
14280422 797locale" against several others, you might think you could gain a little
39332f68 798efficiency by using C<POSIX::strxfrm()> in conjunction with C<eq>:
14280422
DD
799
800 use POSIX qw(strxfrm);
801 $xfrm_string = strxfrm("Mixed-case string");
802 print "locale collation ignores spaces\n"
803 if $xfrm_string eq strxfrm("Mixed-casestring");
804 print "locale collation ignores hyphens\n"
805 if $xfrm_string eq strxfrm("Mixedcase string");
806 print "locale collation ignores case\n"
807 if $xfrm_string eq strxfrm("mixed-case string");
808
39332f68 809C<strxfrm()> takes a string and maps it into a transformed string for use
de108802 810in char-by-char comparisons against other transformed strings during
14280422 811collation. "Under the hood", locale-affected Perl comparison operators
39332f68
KW
812call C<strxfrm()> for both operands, then do a char-by-char
813comparison of the transformed strings. By calling C<strxfrm()> explicitly
14280422 814and using a non locale-affected comparison, the example attempts to save
5a964f20 815a couple of transformations. But in fact, it doesn't save anything: Perl
2ae324a7 816magic (see L<perlguts/Magic Variables>) creates the transformed version of a
5a964f20 817string the first time it's needed in a comparison, then keeps this version around
14280422 818in case it's needed again. An example rewritten the easy way with
e38874e2 819C<cmp> runs just about as fast. It also copes with null characters
39332f68 820embedded in strings; if you call C<strxfrm()> directly, it treats the first
5a964f20
TC
821null it finds as a terminator. don't expect the transformed strings
822it produces to be portable across systems--or even from one revision
39332f68 823of your operating system to the next. In short, don't call C<strxfrm()>
e38874e2 824directly: let Perl do it for you.
14280422 825
5a964f20 826Note: C<use locale> isn't shown in some of these examples because it isn't
dfcc8045
KW
827needed: C<strcoll()> and C<strxfrm()> are POSIX functions
828which use the standard system-supplied C<libc> functions that
829always obey the current C<LC_COLLATE> locale.
5f05dabc 830
cb88b78e 831=head2 Category C<LC_CTYPE>: Character Types
5f05dabc 832
d6ded950
KW
833In the scope of a S<C<use locale>> form that includes C<LC_CTYPE>, Perl
834obeys the C<LC_CTYPE> locale
14280422 835setting. This controls the application's notion of which characters are
ebc3223b
KW
836alphabetic, numeric, punctuation, I<etc>. This affects Perl's C<\w>
837regular expression metanotation,
f1cbbd6e 838which stands for alphanumeric characters--that is, alphabetic,
ebc3223b
KW
839numeric, and the platform's native underscore.
840(Consult L<perlre> for more information about
14280422 841regular expressions.) Thanks to C<LC_CTYPE>, depending on your locale
b4ffc3db
TC
842setting, characters like "E<aelig>", "E<eth>", "E<szlig>", and
843"E<oslash>" may be understood as C<\w> characters.
ebc3223b
KW
844It also affects things like C<\s>, C<\D>, and the POSIX character
845classes, like C<[[:graph:]]>. (See L<perlrecharclass> for more
846information on all these.)
5f05dabc 847
2c268ad5 848The C<LC_CTYPE> locale also provides the map used in transliterating
68dc0745 849characters between lower and uppercase. This affects the case-mapping
39332f68 850functions--C<fc()>, C<lc()>, C<lcfirst()>, C<uc()>, and C<ucfirst()>; case-mapping
b9cc4f69
KW
851interpolation with C<\F>, C<\l>, C<\L>, C<\u>, or C<\U> in double-quoted
852strings and C<s///> substitutions; and case-independent regular expression
e38874e2
DD
853pattern matching using the C<i> modifier.
854
2da736a2 855Finally, C<LC_CTYPE> affects the (deprecated) POSIX character-class test
1d2ab946
KW
856functions--C<POSIX::isalpha()>, C<POSIX::islower()>, and so on. For
857example, if you move from the "C" locale to a 7-bit Scandinavian one,
858you may find--possibly to your surprise--that "|" moves from the
859C<POSIX::ispunct()> class to C<POSIX::isalpha()>.
ef3087ec
KW
860Unfortunately, this creates big problems for regular expressions. "|" still
861means alternation even though it matches C<\w>.
5f05dabc 862
31f05a37
KW
863Starting in v5.20, Perl supports UTF-8 locales for C<LC_CTYPE>, but
864otherwise Perl only supports single-byte locales, such as the ISO 8859
865series. This means that wide character locales, for example for Asian
866languages, are not supported. The UTF-8 locale support is actually a
867superset of POSIX locales, because it is really full Unicode behavior
868as if no locale were in effect at all (except for tainting; see
869L</SECURITY>). POSIX locales, even UTF-8 ones,
870are lacking certain concepts in Unicode, such as the idea that changing
871the case of a character could expand to be more than one character.
872Perl in a UTF-8 locale, will give you that expansion. Prior to v5.20,
873Perl treated a UTF-8 locale on some platforms like an ISO 8859-1 one,
874with some restrictions, and on other platforms more like the "C" locale.
875For releases v5.16 and v5.18, C<S<use locale 'not_characters>> could be
876used as a workaround for this (see L</Unicode and UTF-8>).
877
5d63e270
KW
878Note that there are quite a few things that are unaffected by the
879current locale. All the escape sequences for particular characters,
880C<\n> for example, always mean the platform's native one. This means,
881for example, that C<\N> in regular expressions (every character
1d2ab946 882but new-line) works on the platform character set.
5d63e270 883
14280422
DD
884B<Note:> A broken or malicious C<LC_CTYPE> locale definition may result
885in clearly ineligible characters being considered to be alphanumeric by
e199995e 886your application. For strict matching of (mundane) ASCII letters and
5a964f20 887digits--for example, in command strings--locale-aware applications
e199995e 888should use C<\w> with the C</a> regular expression modifier. See L<"SECURITY">.
5f05dabc 889
fee33030 890=head2 Category C<LC_NUMERIC>: Numeric Formatting
5f05dabc 891
d6ded950
KW
892After a proper C<POSIX::setlocale()> call, and within the scope of
893of a C<use locale> form that includes numerics, Perl obeys the
894C<LC_NUMERIC> locale information, which controls an application's idea
895of how numbers should be formatted for human readability.
b960a36e 896In most implementations the only effect is to
b4ffc3db 897change the character used for the decimal point--perhaps from "." to ",".
b960a36e 898The functions aren't aware of such niceties as thousands separation and
2095dafa 899so on. (See L<The localeconv function> if you care about these things.)
5a964f20 900
b960a36e
KW
901 use POSIX qw(strtod setlocale LC_NUMERIC);
902 use locale;
5f05dabc 903
b960a36e 904 setlocale LC_NUMERIC, "";
14280422 905
b960a36e 906 $n = 5/2; # Assign numeric 2.5 to $n
5f05dabc 907
b960a36e 908 $a = " $n"; # Locale-dependent conversion to string
5f05dabc 909
b960a36e 910 print "half five is $n\n"; # Locale-dependent output
5f05dabc 911
b960a36e 912 printf "half five is %g\n", $n; # Locale-dependent output
5f05dabc 913
b960a36e
KW
914 print "DECIMAL POINT IS COMMA\n"
915 if $n == (strtod("2,5"))[0]; # Locale-dependent conversion
5f05dabc 916
4bbcc6e8
JH
917See also L<I18N::Langinfo> and C<RADIXCHAR>.
918
cb88b78e 919=head2 Category C<LC_MONETARY>: Formatting of monetary amounts
5f05dabc 920
e199995e 921The C standard defines the C<LC_MONETARY> category, but not a function
5a964f20 922that is affected by its contents. (Those with experience of standards
b0c42ed9 923committees will recognize that the working group decided to punt on the
fa9b773e
KW
924issue.) Consequently, Perl essentially takes no notice of it. If you
925really want to use C<LC_MONETARY>, you can query its contents--see
e05ffc7d
KW
926L<The localeconv function>--and use the information that it returns in your
927application's own formatting of currency amounts. However, you may well
928find that the information, voluminous and complex though it may be, still
929does not quite meet your requirements: currency formatting is a hard nut
13a2d996 930to crack.
5f05dabc 931
4bbcc6e8
JH
932See also L<I18N::Langinfo> and C<CRNCYSTR>.
933
cb88b78e 934=head2 C<LC_TIME>
5f05dabc 935
39332f68 936Output produced by C<POSIX::strftime()>, which builds a formatted
5f05dabc
PP
937human-readable date/time string, is affected by the current C<LC_TIME>
938locale. Thus, in a French locale, the output produced by the C<%B>
939format element (full month name) for the first month of the year would
5a964f20 940be "janvier". Here's how to get a list of long month names in the
5f05dabc
PP
941current locale:
942
943 use POSIX qw(strftime);
14280422
DD
944 for (0..11) {
945 $long_month_name[$_] =
946 strftime("%B", 0, 0, 0, 1, $_, 96);
5f05dabc
PP
947 }
948
2619d284
KW
949Note: C<use locale> isn't needed in this example: C<strftime()> is a POSIX
950function which uses the standard system-supplied C<libc> function that
951always obeys the current C<LC_TIME> locale.
5f05dabc 952
4bbcc6e8 953See also L<I18N::Langinfo> and C<ABDAY_1>..C<ABDAY_7>, C<DAY_1>..C<DAY_7>,
2a2bf5f4 954C<ABMON_1>..C<ABMON_12>, and C<ABMON_1>..C<ABMON_12>.
4bbcc6e8 955
5f05dabc
PP
956=head2 Other categories
957
2619d284
KW
958The remaining locale categories are not currently used by Perl itself.
959But again note that things Perl interacts with may use these, including
960extensions outside the standard Perl distribution, and by the
98a6f11e 961operating system and its utilities. Note especially that the string
962value of C<$!> and the error messages given by external utilities may
963be changed by C<LC_MESSAGES>. If you want to have portable error
265f5c4a 964codes, use C<%!>. See L<Errno>.
14280422
DD
965
966=head1 SECURITY
967
5a964f20 968Although the main discussion of Perl security issues can be found in
14280422
DD
969L<perlsec>, a discussion of Perl's locale handling would be incomplete
970if it did not draw your attention to locale-dependent security issues.
5a964f20
TC
971Locales--particularly on systems that allow unprivileged users to
972build their own locales--are untrustworthy. A malicious (or just plain
14280422
DD
973broken) locale can make a locale-aware application give unexpected
974results. Here are a few possibilities:
975
976=over 4
977
978=item *
979
980Regular expression checks for safe file names or mail addresses using
5a964f20 981C<\w> may be spoofed by an C<LC_CTYPE> locale that claims that
14280422
DD
982characters such as "E<gt>" and "|" are alphanumeric.
983
984=item *
985
e38874e2 986String interpolation with case-mapping, as in, say, C<$dest =
cb88b78e 987"C:\U$name.$ext">, may produce dangerous results if a bogus C<LC_CTYPE>
e38874e2
DD
988case-mapping table is in effect.
989
990=item *
991
14280422
DD
992A sneaky C<LC_COLLATE> locale could result in the names of students with
993"D" grades appearing ahead of those with "A"s.
994
995=item *
996
5a964f20 997An application that takes the trouble to use information in
14280422 998C<LC_MONETARY> may format debits as if they were credits and vice versa
5a964f20 999if that locale has been subverted. Or it might make payments in US
14280422
DD
1000dollars instead of Hong Kong dollars.
1001
1002=item *
1003
39332f68 1004The date and day names in dates formatted by C<strftime()> could be
14280422 1005manipulated to advantage by a malicious user able to subvert the
5a964f20 1006C<LC_DATE> locale. ("Look--it says I wasn't in the building on
14280422
DD
1007Sunday.")
1008
1009=back
1010
1011Such dangers are not peculiar to the locale system: any aspect of an
5a964f20 1012application's environment which may be modified maliciously presents
14280422 1013similar challenges. Similarly, they are not specific to Perl: any
5a964f20 1014programming language that allows you to write programs that take
14280422
DD
1015account of their environment exposes you to these issues.
1016
5a964f20
TC
1017Perl cannot protect you from all possibilities shown in the
1018examples--there is no substitute for your own vigilance--but, when
14280422 1019C<use locale> is in effect, Perl uses the tainting mechanism (see
5a964f20 1020L<perlsec>) to mark string results that become locale-dependent, and
14280422 1021which may be untrustworthy in consequence. Here is a summary of the
5a964f20 1022tainting behavior of operators and functions that may be affected by
14280422
DD
1023the locale:
1024
1025=over 4
1026
551e1d92
RB
1027=item *
1028
1029B<Comparison operators> (C<lt>, C<le>, C<ge>, C<gt> and C<cmp>):
14280422
DD
1030
1031Scalar true/false (or less/equal/greater) result is never tainted.
1032
551e1d92
RB
1033=item *
1034
1d2ab946 1035B<Case-mapping interpolation> (with C<\l>, C<\L>, C<\u>, C<\U>, or C<\F>)
e38874e2 1036
d6ded950
KW
1037The result string containing interpolated material is tainted if
1038a C<use locale> form that includes C<LC_CTYPE> is in effect.
e38874e2 1039
551e1d92
RB
1040=item *
1041
1042B<Matching operator> (C<m//>):
14280422
DD
1043
1044Scalar true/false result never tainted.
1045
1d2ab946 1046All subpatterns, either delivered as a list-context result or as C<$1>
d6ded950
KW
1047I<etc>., are tainted if a C<use locale> form that includes
1048C<LC_CTYPE> is in effect, and the subpattern
63baef57
KW
1049regular expression contains a locale-dependent construct. These
1050constructs include C<\w> (to match an alphanumeric character), C<\W>
1051(non-alphanumeric character), C<\b> and C<\B> (word-boundary and
1052non-boundardy, which depend on what C<\w> and C<\W> match), C<\s>
1053(whitespace character), C<\S> (non whitespace character), C<\d> and
1054C<\D> (digits and non-digits), and the POSIX character classes, such as
1055C<[:alpha:]> (see L<perlrecharclass/POSIX Character Classes>).
1056
1057Tainting is also likely if the pattern is to be matched
1058case-insensitively (via C</i>). The exception is if all the code points
1059to be matched this way are above 255 and do not have folds under Unicode
1060rules to below 256. Tainting is not done for these because Perl
1061only uses Unicode rules for such code points, and those rules are the
1062same no matter what the current locale.
1063
1d2ab946
KW
1064The matched-pattern variables, C<$&>, C<$`> (pre-match), C<$'>
1065(post-match), and C<$+> (last match) also are tainted.
14280422 1066
551e1d92
RB
1067=item *
1068
1069B<Substitution operator> (C<s///>):
14280422 1070
e38874e2 1071Has the same behavior as the match operator. Also, the left
d6ded950
KW
1072operand of C<=~> becomes tainted when a C<use locale>
1073form that includes C<LC_CTYPE> is in effect, if modified as
66cbab2c 1074a result of a substitution based on a regular
1d2ab946
KW
1075expression match involving any of the things mentioned in the previous
1076item, or of case-mapping, such as C<\l>, C<\L>,C<\u>, C<\U>, or C<\F>.
14280422 1077
551e1d92
RB
1078=item *
1079
39332f68 1080B<Output formatting functions> (C<printf()> and C<write()>):
14280422 1081
3cf03d68
JH
1082Results are never tainted because otherwise even output from print,
1083for example C<print(1/7)>, should be tainted if C<use locale> is in
1084effect.
14280422 1085
551e1d92
RB
1086=item *
1087
39332f68 1088B<Case-mapping functions> (C<lc()>, C<lcfirst()>, C<uc()>, C<ucfirst()>):
14280422 1089
d6ded950
KW
1090Results are tainted if a C<use locale> form that includes C<LC_CTYPE> is
1091in effect.
14280422 1092
551e1d92
RB
1093=item *
1094
39332f68
KW
1095B<POSIX locale-dependent functions> (C<localeconv()>, C<strcoll()>,
1096C<strftime()>, C<strxfrm()>):
14280422
DD
1097
1098Results are never tainted.
1099
551e1d92
RB
1100=item *
1101
1d2ab946
KW
1102B<POSIX character class tests> (C<POSIX::isalnum()>,
1103C<POSIX::isalpha()>, C<POSIX::isdigit()>, C<POSIX::isgraph()>,
1104C<POSIX::islower()>, C<POSIX::isprint()>, C<POSIX::ispunct()>,
1105C<POSIX::isspace()>, C<POSIX::isupper()>, C<POSIX::isxdigit()>):
14280422
DD
1106
1107True/false results are never tainted.
1108
1109=back
1110
1111Three examples illustrate locale-dependent tainting.
1112The first program, which ignores its locale, won't run: a value taken
54310121 1113directly from the command line may not be used to name an output file
14280422
DD
1114when taint checks are enabled.
1115
1116 #/usr/local/bin/perl -T
1117 # Run with taint checking
1118
54310121 1119 # Command line sanity check omitted...
14280422
DD
1120 $tainted_output_file = shift;
1121
1122 open(F, ">$tainted_output_file")
3183d96c 1123 or warn "Open of $tainted_output_file failed: $!\n";
14280422
DD
1124
1125The program can be made to run by "laundering" the tainted value through
5a964f20
TC
1126a regular expression: the second example--which still ignores locale
1127information--runs, creating the file named on its command line
14280422
DD
1128if it can.
1129
1130 #/usr/local/bin/perl -T
1131
1132 $tainted_output_file = shift;
1133 $tainted_output_file =~ m%[\w/]+%;
1134 $untainted_output_file = $&;
1135
1136 open(F, ">$untainted_output_file")
1137 or warn "Open of $untainted_output_file failed: $!\n";
1138
5a964f20 1139Compare this with a similar but locale-aware program:
14280422
DD
1140
1141 #/usr/local/bin/perl -T
1142
1143 $tainted_output_file = shift;
1144 use locale;
1145 $tainted_output_file =~ m%[\w/]+%;
1146 $localized_output_file = $&;
1147
1148 open(F, ">$localized_output_file")
1149 or warn "Open of $localized_output_file failed: $!\n";
1150
1d2ab946 1151This third program fails to run because C<$&> is tainted: it is the result
5a964f20 1152of a match involving C<\w> while C<use locale> is in effect.
5f05dabc
PP
1153
1154=head1 ENVIRONMENT
1155
1156=over 12
1157
ee1ec05f
KW
1158=item PERL_SKIP_LOCALE_INIT
1159
1160This environment variable, available starting in Perl v5.20, and if it
1161evaluates to a TRUE value, tells Perl to not use the rest of the
1162environment variables to initialize with. Instead, Perl uses whatever
1163the current locale settings are. This is particularly useful in
1164embedded environments, see
1165L<perlembed/Using embedded Perl with POSIX locales>.
1166
5f05dabc
PP
1167=item PERL_BADLANG
1168
14280422 1169A string that can suppress Perl's warning about failed locale settings
54310121 1170at startup. Failure can occur if the locale support in the operating
5a964f20 1171system is lacking (broken) in some way--or if you mistyped the name of
900bd440
JH
1172a locale when you set up your environment. If this environment
1173variable is absent, or has a value that does not evaluate to integer
1174zero--that is, "0" or ""-- Perl will complain about locale setting
1175failures.
5f05dabc 1176
cb88b78e 1177B<NOTE>: C<PERL_BADLANG> only gives you a way to hide the warning message.
14280422
DD
1178The message tells about some problem in your system's locale support,
1179and you should investigate what the problem is.
5f05dabc
PP
1180
1181=back
1182
1183The following environment variables are not specific to Perl: They are
39332f68 1184part of the standardized (ISO C, XPG4, POSIX 1.c) C<setlocale()> method
b385bb4d
KW
1185for controlling an application's opinion on data. Windows is non-POSIX,
1186but Perl arranges for the following to work as described anyway.
65ebb059
KW
1187If the locale given by an environment variable is not valid, Perl tries
1188the next lower one in priority. If none are valid, on Windows, the
1189system default locale is then tried. If all else fails, the C<"C">
1190locale is used. If even that doesn't work, something is badly broken,
c5e9a8e7 1191but Perl tries to forge ahead with whatever the locale settings might
65ebb059 1192be.
5f05dabc
PP
1193
1194=over 12
1195
cb88b78e 1196=item C<LC_ALL>
5f05dabc 1197
5a964f20 1198C<LC_ALL> is the "override-all" locale environment variable. If
5f05dabc
PP
1199set, it overrides all the rest of the locale environment variables.
1200
cb88b78e 1201=item C<LANGUAGE>
528d65ad
JH
1202
1203B<NOTE>: C<LANGUAGE> is a GNU extension, it affects you only if you
1204are using the GNU libc. This is the case if you are using e.g. Linux.
e1020413 1205If you are using "commercial" Unixes you are most probably I<not>
22b6f60d
JH
1206using GNU libc and you can ignore C<LANGUAGE>.
1207
1208However, in the case you are using C<LANGUAGE>: it affects the
1209language of informational, warning, and error messages output by
1210commands (in other words, it's like C<LC_MESSAGES>) but it has higher
96090e4f 1211priority than C<LC_ALL>. Moreover, it's not a single value but
22b6f60d
JH
1212instead a "path" (":"-separated list) of I<languages> (not locales).
1213See the GNU C<gettext> library documentation for more information.
528d65ad 1214
cb88b78e 1215=item C<LC_CTYPE>.
5f05dabc
PP
1216
1217In the absence of C<LC_ALL>, C<LC_CTYPE> chooses the character type
1218locale. In the absence of both C<LC_ALL> and C<LC_CTYPE>, C<LANG>
1219chooses the character type locale.
1220
cb88b78e 1221=item C<LC_COLLATE>
5f05dabc 1222
14280422
DD
1223In the absence of C<LC_ALL>, C<LC_COLLATE> chooses the collation
1224(sorting) locale. In the absence of both C<LC_ALL> and C<LC_COLLATE>,
1225C<LANG> chooses the collation locale.
5f05dabc 1226
cb88b78e 1227=item C<LC_MONETARY>
5f05dabc 1228
14280422
DD
1229In the absence of C<LC_ALL>, C<LC_MONETARY> chooses the monetary
1230formatting locale. In the absence of both C<LC_ALL> and C<LC_MONETARY>,
1231C<LANG> chooses the monetary formatting locale.
5f05dabc 1232
cb88b78e 1233=item C<LC_NUMERIC>
5f05dabc
PP
1234
1235In the absence of C<LC_ALL>, C<LC_NUMERIC> chooses the numeric format
1236locale. In the absence of both C<LC_ALL> and C<LC_NUMERIC>, C<LANG>
1237chooses the numeric format.
1238
cb88b78e 1239=item C<LC_TIME>
5f05dabc 1240
14280422
DD
1241In the absence of C<LC_ALL>, C<LC_TIME> chooses the date and time
1242formatting locale. In the absence of both C<LC_ALL> and C<LC_TIME>,
1243C<LANG> chooses the date and time formatting locale.
5f05dabc 1244
cb88b78e 1245=item C<LANG>
5f05dabc 1246
14280422
DD
1247C<LANG> is the "catch-all" locale environment variable. If it is set, it
1248is used as the last resort after the overall C<LC_ALL> and the
cb88b78e 1249category-specific C<LC_I<foo>>
5f05dabc
PP
1250
1251=back
1252
7e4353e9
RGS
1253=head2 Examples
1254
cb88b78e 1255The C<LC_NUMERIC> controls the numeric output:
7e4353e9 1256
ef3087ec
KW
1257 use locale;
1258 use POSIX qw(locale_h); # Imports setlocale() and the LC_ constants.
1259 setlocale(LC_NUMERIC, "fr_FR") or die "Pardon";
1260 printf "%g\n", 1.23; # If the "fr_FR" succeeded, probably shows 1,23.
7e4353e9 1261
39332f68 1262and also how strings are parsed by C<POSIX::strtod()> as numbers:
7e4353e9 1263
ef3087ec
KW
1264 use locale;
1265 use POSIX qw(locale_h strtod);
1266 setlocale(LC_NUMERIC, "de_DE") or die "Entschuldigung";
1267 my $x = strtod("2,34") + 5;
1268 print $x, "\n"; # Probably shows 7,34.
7e4353e9 1269
5f05dabc
PP
1270=head1 NOTES
1271
b960a36e
KW
1272=head2 String C<eval> and C<LC_NUMERIC>
1273
1274A string L<eval|perlfunc/eval EXPR> parses its expression as standard
1275Perl. It is therefore expecting the decimal point to be a dot. If
1276C<LC_NUMERIC> is set to have this be a comma instead, the parsing will
1277be confused, perhaps silently.
1278
1279 use locale;
1280 use POSIX qw(locale_h);
1281 setlocale(LC_NUMERIC, "fr_FR") or die "Pardon";
1282 my $a = 1.2;
1283 print eval "$a + 1.5";
1284 print "\n";
1285
1286prints C<13,5>. This is because in that locale, the comma is the
1287decimal point character. The C<eval> thus expands to:
1288
1289 eval "1,2 + 1.5"
1290
1291and the result is not what you likely expected. No warnings are
1292generated. If you do string C<eval>'s within the scope of
1293S<C<use locale>>, you should instead change the C<eval> line to do
1294something like:
1295
1296 print eval "no locale; $a + 1.5";
1297
1298This prints C<2.7>.
1299
d6ded950
KW
1300You could also exclude C<LC_NUMERIC>, if you don't need it, by
1301
1302 use locale ':!numeric';
1303
5f05dabc
PP
1304=head2 Backward compatibility
1305
b0c42ed9 1306Versions of Perl prior to 5.004 B<mostly> ignored locale information,
5a964f20
TC
1307generally behaving as if something similar to the C<"C"> locale were
1308always in force, even if the program environment suggested otherwise
1309(see L<The setlocale function>). By default, Perl still behaves this
1310way for backward compatibility. If you want a Perl application to pay
1311attention to locale information, you B<must> use the S<C<use locale>>
d6ded950 1312pragma (see L<The "use locale" pragma>) or, in the unlikely event
062ca197 1313that you want to do so for just pattern matching, the
70709c68
KW
1314C</l> regular expression modifier (see L<perlre/Character set
1315modifiers>) to instruct it to do so.
b0c42ed9
JH
1316
1317Versions of Perl from 5.002 to 5.003 did use the C<LC_CTYPE>
5a964f20
TC
1318information if available; that is, C<\w> did understand what
1319were the letters according to the locale environment variables.
b0c42ed9
JH
1320The problem was that the user had no control over the feature:
1321if the C library supported locales, Perl used them.
1322
1323=head2 I18N:Collate obsolete
1324
5a964f20 1325In versions of Perl prior to 5.004, per-locale collation was possible
b0c42ed9
JH
1326using the C<I18N::Collate> library module. This module is now mildly
1327obsolete and should be avoided in new applications. The C<LC_COLLATE>
1328functionality is now integrated into the Perl core language: One can
1329use locale-specific scalar data completely normally with C<use locale>,
1330so there is no longer any need to juggle with the scalar references of
1331C<I18N::Collate>.
5f05dabc 1332
14280422 1333=head2 Sort speed and memory use impacts
5f05dabc
PP
1334
1335Comparing and sorting by locale is usually slower than the default
14280422
DD
1336sorting; slow-downs of two to four times have been observed. It will
1337also consume more memory: once a Perl scalar variable has participated
1338in any string comparison or sorting operation obeying the locale
1339collation rules, it will take 3-15 times more memory than before. (The
1340exact multiplier depends on the string's contents, the operating system
1341and the locale.) These downsides are dictated more by the operating
1342system's implementation of the locale system than by Perl.
5f05dabc 1343
5f05dabc
PP
1344=head2 Freely available locale definitions
1345
66cbab2c
KW
1346The Unicode CLDR project extracts the POSIX portion of many of its
1347locales, available at
1348
1349 http://unicode.org/Public/cldr/latest/
1350
08d7a6b2
LB
1351There is a large collection of locale definitions at:
1352
1353 http://std.dkuug.dk/i18n/WG15-collection/locales/
1354
1355You should be aware that it is
14280422 1356unsupported, and is not claimed to be fit for any purpose. If your
5a964f20 1357system allows installation of arbitrary locales, you may find the
14280422
DD
1358definitions useful as they are, or as a basis for the development of
1359your own locales.
5f05dabc 1360
14280422 1361=head2 I18n and l10n
5f05dabc 1362
b0c42ed9
JH
1363"Internationalization" is often abbreviated as B<i18n> because its first
1364and last letters are separated by eighteen others. (You may guess why
1365the internalin ... internaliti ... i18n tends to get abbreviated.) In
1366the same way, "localization" is often abbreviated to B<l10n>.
14280422
DD
1367
1368=head2 An imperfect standard
1369
1370Internationalization, as defined in the C and POSIX standards, can be
1371criticized as incomplete, ungainly, and having too large a granularity.
1372(Locales apply to a whole process, when it would arguably be more useful
1373to have them apply to a single thread, window group, or whatever.) They
1374also have a tendency, like standards groups, to divide the world into
1375nations, when we all know that the world can equally well be divided
e199995e 1376into bankers, bikers, gamers, and so on.
5f05dabc 1377
b310b053
JH
1378=head1 Unicode and UTF-8
1379
7ee2ae1e 1380The support of Unicode is new starting from Perl version v5.6, and more fully
31f05a37
KW
1381implemented in versions v5.8 and later. See L<perluniintro>.
1382
1383Starting in Perl v5.20, UTF-8 locales are supported in Perl, except for
1384C<LC_COLLATE> (use L<Unicode::Collate> instead). If you have Perl v5.16
1385or v5.18 and can't upgrade, you can use
66cbab2c
KW
1386
1387 use locale ':not_characters';
1388
1389When this form of the pragma is used, only the non-character portions of
1390locales are used by Perl, for example C<LC_NUMERIC>. Perl assumes that
1391you have translated all the characters it is to operate on into Unicode
1392(actually the platform's native character set (ASCII or EBCDIC) plus
1393Unicode). For data in files, this can conveniently be done by also
1394specifying
1395
1396 use open ':locale';
1397
1398This pragma arranges for all inputs from files to be translated into
1399Unicode from the current locale as specified in the environment (see
1400L</ENVIRONMENT>), and all outputs to files to be translated back
1401into the locale. (See L<open>). On a per-filehandle basis, you can
1402instead use the L<PerlIO::locale> module, or the L<Encode::Locale>
1403module, both available from CPAN. The latter module also has methods to
1404ease the handling of C<ARGV> and environment variables, and can be used
31f05a37 1405on individual strings. If you know that all your locales will be
66cbab2c
KW
1406UTF-8, as many are these days, you can use the L<B<-C>|perlrun/-C>
1407command line switch.
1408
1409This form of the pragma allows essentially seamless handling of locales
31f05a37
KW
1410with Unicode. The collation order will be by Unicode code point order.
1411It is strongly
66cbab2c
KW
1412recommended that when you need to order and sort strings that you use
1413the standard module L<Unicode::Collate> which gives much better results
1414in many instances than you can get with the old-style locale handling.
1415
31f05a37
KW
1416All the modules and switches just described can be used in v5.20 with
1417just plain C<use locale>, and, should the input locales not be UTF-8,
1418you'll get the less than ideal behavior, described below, that you get
1419with pre-v5.16 Perls, or when you use the locale pragma without the
1420C<:not_characters> parameter in v5.16 and v5.18. If you are using
1421exclusively UTF-8 locales in v5.20 and higher, the rest of this section
1422does not apply to you.
1423
1424There are two cases, multi-byte and single-byte locales. First
1425multi-byte:
1426
1427The only multi-byte (or wide character) locale that Perl is ever likely
1428to support is UTF-8. This is due to the difficulty of implementation,
1429the fact that high quality UTF-8 locales are now published for every
1430area of the world (L<http://unicode.org/Public/cldr/latest/>), and that
1431failing all that you can use the L<Encode> module to translate to/from
1432your locale. So, you'll have to do one of those things if you're using
1433one of these locales, such as Big5 or Shift JIS. For UTF-8 locales, in
1434Perls (pre v5.20) that don't have full UTF-8 locale support, they may
1435work reasonably well (depending on your C library implementation)
1436simply because both
dc4bfc4b
KW
1437they and Perl store characters that take up multiple bytes the same way.
1438However, some, if not most, C library implementations may not process
1439the characters in the upper half of the Latin-1 range (128 - 255)
cb88b78e 1440properly under C<LC_CTYPE>. To see if a character is a particular type
dc4bfc4b
KW
1441under a locale, Perl uses the functions like C<isalnum()>. Your C
1442library may not work for UTF-8 locales with those functions, instead
1443only working under the newer wide library functions like C<iswalnum()>.
31f05a37
KW
1444However, they are treated like single-byte locales, and will have the
1445restrictions described below.
e199995e 1446
31f05a37 1447For single-byte locales,
e199995e 1448Perl generally takes the tack to use locale rules on code points that can fit
66cbab2c
KW
1449in a single byte, and Unicode rules for those that can't (though this
1450isn't uniformly applied, see the note at the end of this section). This
1451prevents many problems in locales that aren't UTF-8. Suppose the locale
1452is ISO8859-7, Greek. The character at 0xD7 there is a capital Chi. But
1453in the ISO8859-1 locale, Latin1, it is a multiplication sign. The POSIX
1454regular expression character class C<[[:alpha:]]> will magically match
14550xD7 in the Greek locale but not in the Latin one.
e199995e 1456
1d2ab946 1457However, there are places where this breaks down. Certain Perl constructs are
b4ffc3db
TC
1458for Unicode only, such as C<\p{Alpha}>. They assume that 0xD7 always has its
1459Unicode meaning (or the equivalent on EBCDIC platforms). Since Latin1 is a
1460subset of Unicode and 0xD7 is the multiplication sign in both Latin1 and
1461Unicode, C<\p{Alpha}> will never match it, regardless of locale. A similar
31f05a37
KW
1462issue occurs with C<\N{...}>. Prior to v5.20, It is therefore a bad
1463idea to use C<\p{}> or
66cbab2c
KW
1464C<\N{}> under plain C<use locale>--I<unless> you can guarantee that the
1465locale will be a ISO8859-1. Use POSIX character classes instead.
1466
1467Another problem with this approach is that operations that cross the
1468single byte/multiple byte boundary are not well-defined, and so are
4a70680a 1469disallowed. (This boundary is between the codepoints at 255/256.)
66cbab2c
KW
1470For example, lower casing LATIN CAPITAL LETTER Y WITH DIAERESIS (U+0178)
1471should return LATIN SMALL LETTER Y WITH DIAERESIS (U+00FF). But in the
1472Greek locale, for example, there is no character at 0xFF, and Perl
1473has no way of knowing what the character at 0xFF is really supposed to
1474represent. Thus it disallows the operation. In this mode, the
1475lowercase of U+0178 is itself.
1476
1477The same problems ensue if you enable automatic UTF-8-ification of your
e199995e 1478standard file handles, default C<open()> layer, and C<@ARGV> on non-ISO8859-1,
b4ffc3db
TC
1479non-UTF-8 locales (by using either the B<-C> command line switch or the
1480C<PERL_UNICODE> environment variable; see L<perlrun>).
1481Things are read in as UTF-8, which would normally imply a Unicode
1482interpretation, but the presence of a locale causes them to be interpreted
1483in that locale instead. For example, a 0xD7 code point in the Unicode
1484input, which should mean the multiplication sign, won't be interpreted by
66cbab2c 1485Perl that way under the Greek locale. This is not a problem
b4ffc3db 1486I<provided> you make certain that all locales will always and only be either
66cbab2c 1487an ISO8859-1, or, if you don't have a deficient C library, a UTF-8 locale.
b4ffc3db 1488
1d2ab946
KW
1489Still another problem is that this approach can lead to two code
1490points meaning the same character. Thus in a Greek locale, both U+03A7
1491and U+00D7 are GREEK CAPITAL LETTER CHI.
1492
b4ffc3db
TC
1493Vendor locales are notoriously buggy, and it is difficult for Perl to test
1494its locale-handling code because this interacts with code that Perl has no
1495control over; therefore the locale-handling code in Perl may be buggy as
66cbab2c
KW
1496well. (However, the Unicode-supplied locales should be better, and
1497there is a feed back mechanism to correct any problems. See
1498L</Freely available locale definitions>.)
1499
7ee2ae1e 1500If you have Perl v5.16, the problems mentioned above go away if you use
66cbab2c 1501the C<:not_characters> parameter to the locale pragma (except for vendor
7ee2ae1e 1502bugs in the non-character portions). If you don't have v5.16, and you
66cbab2c
KW
1503I<do> have locales that work, using them may be worthwhile for certain
1504specific purposes, as long as you keep in mind the gotchas already
1505mentioned. For example, if the collation for your locales works, it
1506runs faster under locales than under L<Unicode::Collate>; and you gain
1507access to such things as the local currency symbol and the names of the
7ee2ae1e 1508months and days of the week. (But to hammer home the point, in v5.16,
66cbab2c
KW
1509you get this access without the downsides of locales by using the
1510C<:not_characters> form of the pragma.)
1511
1512Note: The policy of using locale rules for code points that can fit in a
1513byte, and Unicode rules for those that can't is not uniformly applied.
7ee2ae1e 1514Pre-v5.12, it was somewhat haphazard; in v5.12 it was applied fairly
66cbab2c 1515consistently to regular expression matching except for bracketed
7ee2ae1e
KW
1516character classes; in v5.14 it was extended to all regex matches; and in
1517v5.16 to the casing operations such as C<"\L"> and C<uc()>. For
66cbab2c
KW
1518collation, in all releases, the system's C<strxfrm()> function is called,
1519and whatever it does is what you get.
b310b053 1520
5f05dabc
PP
1521=head1 BUGS
1522
1523=head2 Broken systems
1524
5a964f20 1525In certain systems, the operating system's locale support
2bdf8add 1526is broken and cannot be fixed or used by Perl. Such deficiencies can
b4ffc3db 1527and will result in mysterious hangs and/or Perl core dumps when
2bdf8add 1528C<use locale> is in effect. When confronted with such a system,
7f2de2d2 1529please report in excruciating detail to <F<perlbug@perl.org>>, and
b4ffc3db 1530also contact your vendor: bug fixes may exist for these problems
2bdf8add 1531in your operating system. Sometimes such bug fixes are called an
83fb1bf0
KW
1532operating system upgrade. If you have the source for Perl, include in
1533the perlbug email the output of the test described above in L</Testing
1534for broken locales>.
5f05dabc
PP
1535
1536=head1 SEE ALSO
1537
b310b053
JH
1538L<I18N::Langinfo>, L<perluniintro>, L<perlunicode>, L<open>,
1539L<POSIX/isalnum>, L<POSIX/isalpha>,
4bbcc6e8
JH
1540L<POSIX/isdigit>, L<POSIX/isgraph>, L<POSIX/islower>,
1541L<POSIX/isprint>, L<POSIX/ispunct>, L<POSIX/isspace>,
1542L<POSIX/isupper>, L<POSIX/isxdigit>, L<POSIX/localeconv>,
1543L<POSIX/setlocale>, L<POSIX/strcoll>, L<POSIX/strftime>,
1544L<POSIX/strtod>, L<POSIX/strxfrm>.
5f05dabc 1545
ccd65d51
KW
1546For special considerations when Perl is embedded in a C program,
1547see L<perlembed/Using embedded Perl with POSIX locales>.
1548
5f05dabc
PP
1549=head1 HISTORY
1550
b0c42ed9 1551Jarkko Hietaniemi's original F<perli18n.pod> heavily hacked by Dominic
5a964f20 1552Dunlop, assisted by the perl5-porters. Prose worked over a bit by
c052850d 1553Tom Christiansen, and updated by Perl 5 porters.