This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
locale.t: Don't show 100% pass when there are failures
[perl5.git] / lib / locale.t
CommitLineData
8ebc5c01
PP
1#!./perl -wT
2
66cbab2c
KW
3# This tests plain 'use locale' and adorned 'use locale ":not_characters"'
4# Because these pragmas are compile time, and I (khw) am trying to test
5# without using 'eval' as much as possible, which might cloud the issue, the
6# crucial parts of the code are duplicated in a block for each pragma.
7
6c2e653d
KW
8# To make a TODO test, add the string 'TODO' to its %test_names value
9
e3a2734b
KW
10binmode STDOUT, ':utf8';
11binmode STDERR, ':utf8';
12
8ebc5c01
PP
13BEGIN {
14 chdir 't' if -d 't';
20822f61 15 @INC = '../lib';
f9cbebe1 16 unshift @INC, '.';
b002077a 17 require Config; import Config;
97a0514d 18 if (!$Config{d_setlocale} || $Config{ccflags} =~ /\bD?NO_LOCALE\b/) {
b002077a
CS
19 print "1..0\n";
20 exit;
21 }
2de3dbcc 22 $| = 1;
8ebc5c01
PP
23}
24
25use strict;
26c1569f 26use feature 'fc';
8ebc5c01 27
108a305e 28my $debug = $ENV{PERL_DEBUG_FULL_TEST} // 0;
284102e8 29
6d5d702a
KW
30# Certain tests have been shown to be problematical for a few locales. Don't
31# fail them unless at least this percentage of the tested locales fail.
32my $acceptable_fold_failure_percentage = 5;
33
db4b7445
A
34use Dumpvalue;
35
36my $dumper = Dumpvalue->new(
37 tick => qq{"},
38 quoteHighBit => 0,
39 unctrl => "quote"
40 );
6be75cd7 41sub debug {
db4b7445
A
42 return unless $debug;
43 my($mess) = join "", @_;
44 chop $mess;
45 print $dumper->stringify($mess,1), "\n";
6be75cd7
JH
46}
47
48sub debugf {
49 printf @_ if $debug;
50}
51
8ebc5c01
PP
52my $have_setlocale = 0;
53eval {
54 require POSIX;
55 import POSIX ':locale_h';
56 $have_setlocale++;
57};
58
6dead956 59# Visual C's CRT goes silly on strings of the form "en_US.ISO8859-1"
f6c6487a 60# and mingw32 uses said silly CRT
3a2d1764
SH
61# This doesn't seem to be an issue any more, at least on Windows XP,
62# so re-enable the tests for Windows XP onwards.
63my $winxp = ($^O eq 'MSWin32' && defined &Win32::GetOSVersion &&
64 join('.', (Win32::GetOSVersion())[1..2]) >= 5.1);
65$have_setlocale = 0 if ((($^O eq 'MSWin32' && !$winxp) || $^O eq 'NetWare') &&
66 $Config{cc} =~ /^(cl|gcc)/i);
6dead956 67
36a42ae7 68# UWIN seems to loop after taint tests, just skip for now
cd19b65c
JH
69$have_setlocale = 0 if ($^O =~ /^uwin/);
70
9a66ea41 71sub LC_ALL ();
8ebc5c01 72
0e053d1e 73$a = 'abc %';
8ebc5c01 74
c213d471
KW
75my $test_num = 0;
76
8ebc5c01 77sub ok {
c213d471 78 my ($result, $message) = @_;
e3a2734b 79 $message = "" unless defined $message;
8ebc5c01
PP
80
81 print 'not ' unless ($result);
c213d471 82 print "ok " . ++$test_num;
e3a2734b
KW
83 print " $message";
84 print "\n";
8ebc5c01
PP
85}
86
87# First we'll do a lot of taint checking for locales.
88# This is the easiest to test, actually, as any locale,
89# even the default locale will taint under 'use locale'.
90
91sub is_tainted { # hello, camel two.
9f1b1f2d 92 no warnings 'uninitialized' ;
8ebc5c01 93 my $dummy;
ba74571d 94 local $@;
8ebc5c01
PP
95 not eval { $dummy = join("", @_), kill 0; 1 }
96}
97
a9b7c637
KW
98sub check_taint ($;$) {
99 my $message_tail = $_[1] // "";
100 $message_tail = ": $message_tail" if $message_tail;
101 ok is_tainted($_[0]), "verify that is tainted$message_tail";
8ebc5c01
PP
102}
103
a9b7c637
KW
104sub check_taint_not ($;$) {
105 my $message_tail = $_[1] // "";
106 $message_tail = ": $message_tail" if $message_tail;
107 ok((not is_tainted($_[0])), "verify that isn't tainted$message_tail");
8ebc5c01
PP
108}
109
bf3cd0e6
KW
110"\tb\t" =~ /^m?(\s)(.*)\1$/;
111check_taint_not $&, "not tainted outside 'use locale'";
112;
113
8ebc5c01
PP
114use locale; # engage locale and therefore locale taint.
115
36a42ae7 116check_taint_not $a;
8ebc5c01 117
36a42ae7
KW
118check_taint uc($a);
119check_taint "\U$a";
120check_taint ucfirst($a);
121check_taint "\u$a";
122check_taint lc($a);
26c1569f 123check_taint fc($a);
36a42ae7 124check_taint "\L$a";
26c1569f 125check_taint "\F$a";
36a42ae7
KW
126check_taint lcfirst($a);
127check_taint "\l$a";
8ebc5c01 128
36a42ae7
KW
129check_taint_not sprintf('%e', 123.456);
130check_taint_not sprintf('%f', 123.456);
131check_taint_not sprintf('%g', 123.456);
132check_taint_not sprintf('%d', 123.456);
133check_taint_not sprintf('%x', 123.456);
8ebc5c01
PP
134
135$_ = $a; # untaint $_
136
137$_ = uc($a); # taint $_
138
36a42ae7 139check_taint $_;
8ebc5c01
PP
140
141/(\w)/; # taint $&, $`, $', $+, $1.
36a42ae7
KW
142check_taint $&;
143check_taint $`;
144check_taint $';
145check_taint $+;
146check_taint $1;
147check_taint_not $2;
8ebc5c01
PP
148
149/(.)/; # untaint $&, $`, $', $+, $1.
36a42ae7
KW
150check_taint_not $&;
151check_taint_not $`;
152check_taint_not $';
153check_taint_not $+;
154check_taint_not $1;
155check_taint_not $2;
8ebc5c01
PP
156
157/(\W)/; # taint $&, $`, $', $+, $1.
36a42ae7
KW
158check_taint $&;
159check_taint $`;
160check_taint $';
161check_taint $+;
162check_taint $1;
163check_taint_not $2;
8ebc5c01
PP
164
165/(\s)/; # taint $&, $`, $', $+, $1.
36a42ae7
KW
166check_taint $&;
167check_taint $`;
168check_taint $';
169check_taint $+;
170check_taint $1;
171check_taint_not $2;
8ebc5c01
PP
172
173/(\S)/; # taint $&, $`, $', $+, $1.
36a42ae7
KW
174check_taint $&;
175check_taint $`;
176check_taint $';
177check_taint $+;
178check_taint $1;
179check_taint_not $2;
8ebc5c01
PP
180
181$_ = $a; # untaint $_
182
36a42ae7 183check_taint_not $_;
8ebc5c01
PP
184
185/(b)/; # this must not taint
36a42ae7
KW
186check_taint_not $&;
187check_taint_not $`;
188check_taint_not $';
189check_taint_not $+;
190check_taint_not $1;
191check_taint_not $2;
8ebc5c01
PP
192
193$_ = $a; # untaint $_
194
36a42ae7 195check_taint_not $_;
8ebc5c01
PP
196
197$b = uc($a); # taint $b
198s/(.+)/$b/; # this must taint only the $_
199
36a42ae7
KW
200check_taint $_;
201check_taint_not $&;
202check_taint_not $`;
203check_taint_not $';
204check_taint_not $+;
205check_taint_not $1;
206check_taint_not $2;
8ebc5c01
PP
207
208$_ = $a; # untaint $_
209
210s/(.+)/b/; # this must not taint
36a42ae7
KW
211check_taint_not $_;
212check_taint_not $&;
213check_taint_not $`;
214check_taint_not $';
215check_taint_not $+;
216check_taint_not $1;
217check_taint_not $2;
8ebc5c01
PP
218
219$b = $a; # untaint $b
220
221($b = $a) =~ s/\w/$&/;
36a42ae7
KW
222check_taint $b; # $b should be tainted.
223check_taint_not $a; # $a should be not.
8ebc5c01
PP
224
225$_ = $a; # untaint $_
226
227s/(\w)/\l$1/; # this must taint
36a42ae7
KW
228check_taint $_;
229check_taint $&;
230check_taint $`;
231check_taint $';
232check_taint $+;
233check_taint $1;
234check_taint_not $2;
8ebc5c01
PP
235
236$_ = $a; # untaint $_
237
238s/(\w)/\L$1/; # this must taint
36a42ae7
KW
239check_taint $_;
240check_taint $&;
241check_taint $`;
242check_taint $';
243check_taint $+;
244check_taint $1;
245check_taint_not $2;
8ebc5c01
PP
246
247$_ = $a; # untaint $_
248
249s/(\w)/\u$1/; # this must taint
36a42ae7
KW
250check_taint $_;
251check_taint $&;
252check_taint $`;
253check_taint $';
254check_taint $+;
255check_taint $1;
256check_taint_not $2;
8ebc5c01
PP
257
258$_ = $a; # untaint $_
259
260s/(\w)/\U$1/; # this must taint
36a42ae7
KW
261check_taint $_;
262check_taint $&;
263check_taint $`;
264check_taint $';
265check_taint $+;
266check_taint $1;
267check_taint_not $2;
8ebc5c01
PP
268
269# After all this tainting $a should be cool.
270
36a42ae7 271check_taint_not $a;
8ebc5c01 272
66cbab2c
KW
273{ # This is just the previous tests copied here with a different
274 # compile-time pragma.
275
276 use locale ':not_characters'; # engage restricted locale with different
277 # tainting rules
278
279 check_taint_not $a;
280
281 check_taint_not uc($a);
282 check_taint_not "\U$a";
283 check_taint_not ucfirst($a);
284 check_taint_not "\u$a";
285 check_taint_not lc($a);
26c1569f 286 check_taint_not fc($a);
66cbab2c 287 check_taint_not "\L$a";
26c1569f 288 check_taint_not "\F$a";
66cbab2c
KW
289 check_taint_not lcfirst($a);
290 check_taint_not "\l$a";
291
292 check_taint_not sprintf('%e', 123.456);
293 check_taint_not sprintf('%f', 123.456);
294 check_taint_not sprintf('%g', 123.456);
295 check_taint_not sprintf('%d', 123.456);
296 check_taint_not sprintf('%x', 123.456);
297
298 $_ = $a; # untaint $_
299
300 $_ = uc($a); # taint $_
301
302 check_taint_not $_;
303
304 /(\w)/; # taint $&, $`, $', $+, $1.
305 check_taint_not $&;
306 check_taint_not $`;
307 check_taint_not $';
308 check_taint_not $+;
309 check_taint_not $1;
310 check_taint_not $2;
311
312 /(.)/; # untaint $&, $`, $', $+, $1.
313 check_taint_not $&;
314 check_taint_not $`;
315 check_taint_not $';
316 check_taint_not $+;
317 check_taint_not $1;
318 check_taint_not $2;
319
320 /(\W)/; # taint $&, $`, $', $+, $1.
321 check_taint_not $&;
322 check_taint_not $`;
323 check_taint_not $';
324 check_taint_not $+;
325 check_taint_not $1;
326 check_taint_not $2;
327
328 /(\s)/; # taint $&, $`, $', $+, $1.
329 check_taint_not $&;
330 check_taint_not $`;
331 check_taint_not $';
332 check_taint_not $+;
333 check_taint_not $1;
334 check_taint_not $2;
335
336 /(\S)/; # taint $&, $`, $', $+, $1.
337 check_taint_not $&;
338 check_taint_not $`;
339 check_taint_not $';
340 check_taint_not $+;
341 check_taint_not $1;
342 check_taint_not $2;
343
344 $_ = $a; # untaint $_
345
346 check_taint_not $_;
347
348 /(b)/; # this must not taint
349 check_taint_not $&;
350 check_taint_not $`;
351 check_taint_not $';
352 check_taint_not $+;
353 check_taint_not $1;
354 check_taint_not $2;
355
356 $_ = $a; # untaint $_
357
358 check_taint_not $_;
359
360 $b = uc($a); # taint $b
361 s/(.+)/$b/; # this must taint only the $_
362
363 check_taint_not $_;
364 check_taint_not $&;
365 check_taint_not $`;
366 check_taint_not $';
367 check_taint_not $+;
368 check_taint_not $1;
369 check_taint_not $2;
370
371 $_ = $a; # untaint $_
372
373 s/(.+)/b/; # this must not taint
374 check_taint_not $_;
375 check_taint_not $&;
376 check_taint_not $`;
377 check_taint_not $';
378 check_taint_not $+;
379 check_taint_not $1;
380 check_taint_not $2;
381
382 $b = $a; # untaint $b
383
384 ($b = $a) =~ s/\w/$&/;
385 check_taint_not $b; # $b should be tainted.
386 check_taint_not $a; # $a should be not.
387
388 $_ = $a; # untaint $_
389
390 s/(\w)/\l$1/; # this must taint
391 check_taint_not $_;
392 check_taint_not $&;
393 check_taint_not $`;
394 check_taint_not $';
395 check_taint_not $+;
396 check_taint_not $1;
397 check_taint_not $2;
398
399 $_ = $a; # untaint $_
400
401 s/(\w)/\L$1/; # this must taint
402 check_taint_not $_;
403 check_taint_not $&;
404 check_taint_not $`;
405 check_taint_not $';
406 check_taint_not $+;
407 check_taint_not $1;
408 check_taint_not $2;
409
410 $_ = $a; # untaint $_
411
412 s/(\w)/\u$1/; # this must taint
413 check_taint_not $_;
414 check_taint_not $&;
415 check_taint_not $`;
416 check_taint_not $';
417 check_taint_not $+;
418 check_taint_not $1;
419 check_taint_not $2;
420
421 $_ = $a; # untaint $_
422
423 s/(\w)/\U$1/; # this must taint
424 check_taint_not $_;
425 check_taint_not $&;
426 check_taint_not $`;
427 check_taint_not $';
428 check_taint_not $+;
429 check_taint_not $1;
430 check_taint_not $2;
431
432 # After all this tainting $a should be cool.
433
434 check_taint_not $a;
435}
436
437# Here are in scope of 'use locale'
438
8ebc5c01
PP
439# I think we've seen quite enough of taint.
440# Let us do some *real* locale work now,
284102e8 441# unless setlocale() is missing (i.e. minitest).
8ebc5c01 442
fdf053ee
KW
443unless ($have_setlocale) {
444 print "1..$test_num\n";
445 exit;
446}
8ebc5c01 447
6cf0b567 448# The test number before our first setlocale()
66330f13 449my $final_without_setlocale = $test_num;
6cf0b567 450
284102e8
JH
451# Find locales.
452
6be75cd7
JH
453debug "# Scanning for locales...\n";
454
455# Note that it's okay that some languages have their native names
456# capitalized here even though that's not "right". They are lowercased
457# anyway later during the scanning process (and besides, some clueless
98dc9551 458# vendor might have them capitalized erroneously anyway).
6be75cd7 459
284102e8 460my $locales = <<EOF;
6be75cd7 461Afrikaans:af:za:1 15
284102e8 462Arabic:ar:dz eg sa:6 arabic8
6be75cd7
JH
463Brezhoneg Breton:br:fr:1 15
464Bulgarski Bulgarian:bg:bg:5
dd8482fc 465Chinese:zh:cn tw:cn.EUC eucCN eucTW euc.CN euc.TW Big5 GB2312 tw.EUC
6be75cd7
JH
466Hrvatski Croatian:hr:hr:2
467Cymraeg Welsh:cy:cy:1 14 15
284102e8 468Czech:cs:cz:2
df8a53a3 469Dansk Danish:da:dk:1 15
6be75cd7 470Nederlands Dutch:nl:be nl:1 15
dd8482fc 471English American British:en:au ca gb ie nz us uk zw:1 15 cp850
6be75cd7
JH
472Esperanto:eo:eo:3
473Eesti Estonian:et:ee:4 6 13
474Suomi Finnish:fi:fi:1 15
475Flamish::fl:1 15
6be75cd7
JH
476Deutsch German:de:at be ch de lu:1 15
477Euskaraz Basque:eu:es fr:1 15
6be75cd7
JH
478Galego Galician:gl:es:1 15
479Ellada Greek:el:gr:7 g8
6be75cd7
JH
480Frysk:fy:nl:1 15
481Greenlandic:kl:gl:4 6
284102e8
JH
482Hebrew:iw:il:8 hebrew8
483Hungarian:hu:hu:2
df8a53a3 484Indonesian:id:id:1 15
6be75cd7
JH
485Gaeilge Irish:ga:IE:1 14 15
486Italiano Italian:it:ch it:1 15
487Nihongo Japanese:ja:jp:euc eucJP jp.EUC sjis
284102e8 488Korean:ko:kr:
6be75cd7
JH
489Latine Latin:la:va:1 15
490Latvian:lv:lv:4 6 13
491Lithuanian:lt:lt:4 6 13
492Macedonian:mk:mk:1 15
493Maltese:mt:mt:3
dd8482fc 494Moldovan:mo:mo:2
df8a53a3 495Norsk Norwegian:no no\@nynorsk nb nn:no:1 15
6be75cd7
JH
496Occitan:oc:es:1 15
497Polski Polish:pl:pl:2
284102e8 498Rumanian:ro:ro:2
a528dad0 499Russki Russian:ru:ru su ua:5 koi8 koi8r KOI8-R koi8u cp1251 cp866
6be75cd7 500Serbski Serbian:sr:yu:5
284102e8 501Slovak:sk:sk:2
6be75cd7 502Slovene Slovenian:sl:si:2
d43ce814
JH
503Sqhip Albanian:sq:sq:1 15
504Svenska Swedish:sv:fi se:1 15
6be75cd7 505Thai:th:th:11 tis620
284102e8 506Turkish:tr:tr:9 turkish8
dd8482fc 507Yiddish:yi::1 15
284102e8
JH
508EOF
509
ee50adbe 510if ($^O eq 'os390') {
dd8482fc 511 # These cause heartburn. Broken locales?
ee50adbe
PP
512 $locales =~ s/Svenska Swedish:sv:fi se:1 15\n//;
513 $locales =~ s/Thai:th:th:11 tis620\n//;
514}
515
ef4a39e5 516sub in_utf8 () { $^H & 0x08 || (${^OPEN} || "") =~ /:utf8/ }
f9cbebe1
JH
517
518if (in_utf8) {
8a6cb2cb 519 require "lib/locale/utf8";
f9cbebe1 520} else {
8a6cb2cb 521 require "lib/locale/latin1";
f9cbebe1
JH
522}
523
284102e8
JH
524my @Locale;
525my $Locale;
526my @Alnum_;
527
284102e8
JH
528sub trylocale {
529 my $locale = shift;
0b9f254b 530 return if grep { $locale eq $_ } @Locale;
e439cacb
KW
531 return unless setlocale(LC_ALL, $locale);
532 my $badutf8;
533 {
534 local $SIG{__WARN__} = sub {
535 $badutf8 = $_[0] =~ /Malformed UTF-8/;
536 };
537 $Locale =~ /UTF-?8/i;
284102e8 538 }
e439cacb
KW
539
540 if ($badutf8) {
541 ok(0, "Locale name contains malformed utf8");
542 return;
543 }
544 push @Locale, $locale;
284102e8 545}
8ebc5c01 546
284102e8
JH
547sub decode_encodings {
548 my @enc;
8ebc5c01 549
284102e8
JH
550 foreach (split(/ /, shift)) {
551 if (/^(\d+)$/) {
552 push @enc, "ISO8859-$1";
553 push @enc, "iso8859$1"; # HP
554 if ($1 eq '1') {
555 push @enc, "roman8"; # HP
556 }
557 } else {
558 push @enc, $_;
dd8482fc 559 push @enc, "$_.UTF-8";
8ebc5c01
PP
560 }
561 }
ee50adbe
PP
562 if ($^O eq 'os390') {
563 push @enc, qw(IBM-037 IBM-819 IBM-1047);
564 }
8ebc5c01 565
284102e8 566 return @enc;
8ebc5c01
PP
567}
568
284102e8
JH
569trylocale("C");
570trylocale("POSIX");
571foreach (0..15) {
572 trylocale("ISO8859-$_");
284102e8 573 trylocale("iso8859$_");
097ee67d
JH
574 trylocale("iso8859-$_");
575 trylocale("iso_8859_$_");
576 trylocale("isolatin$_");
577 trylocale("isolatin-$_");
578 trylocale("iso_latin_$_");
8ebc5c01
PP
579}
580
645e49ed
JH
581# Sanitize the environment so that we can run the external 'locale'
582# program without the taint mode getting grumpy.
cce5967e
JH
583
584# $ENV{PATH} is special in VMS.
585delete $ENV{PATH} if $^O ne 'VMS' or $Config{d_setenv};
586
587# Other subversive stuff.
588delete @ENV{qw(IFS CDPATH ENV BASH_ENV)};
dd8482fc 589
21477fb4 590if (-x "/usr/bin/locale" && open(LOCALES, "/usr/bin/locale -a 2>/dev/null|")) {
dd8482fc 591 while (<LOCALES>) {
d281a6ac
NC
592 # It seems that /usr/bin/locale steadfastly outputs 8 bit data, which
593 # ain't great when we're running this testPERL_UNICODE= so that utf8
594 # locales will cause all IO hadles to default to (assume) utf8
595 next unless utf8::valid($_);
dd8482fc
JH
596 chomp;
597 trylocale($_);
284102e8 598 }
dd8482fc 599 close(LOCALES);
a6259068 600} elsif ($^O eq 'VMS' && defined($ENV{'SYS$I18N_LOCALE'}) && -d 'SYS$I18N_LOCALE') {
71e5cbb3 601# The SYS$I18N_LOCALE logical name search list was not present on
a6259068
PP
602# VAX VMS V5.5-12, but was on AXP && VAX VMS V6.2 as well as later versions.
603 opendir(LOCALES, "SYS\$I18N_LOCALE:");
604 while ($_ = readdir(LOCALES)) {
605 chomp;
606 trylocale($_);
607 }
608 close(LOCALES);
87e33296
SP
609} elsif ($^O eq 'openbsd' && -e '/usr/share/locale') {
610
611 # OpenBSD doesn't have a locale executable, so reading /usr/share/locale
612 # is much easier and faster than the last resort method.
613
614 opendir(LOCALES, '/usr/share/locale');
615 while ($_ = readdir(LOCALES)) {
616 chomp;
617 trylocale($_);
618 }
619 close(LOCALES);
dd8482fc
JH
620} else {
621
622 # This is going to be slow.
623
624 foreach my $locale (split(/\n/, $locales)) {
625 my ($locale_name, $language_codes, $country_codes, $encodings) =
626 split(/:/, $locale);
627 my @enc = decode_encodings($encodings);
628 foreach my $loc (split(/ /, $locale_name)) {
629 trylocale($loc);
284102e8 630 foreach my $enc (@enc) {
dd8482fc 631 trylocale("$loc.$enc");
284102e8 632 }
dd8482fc 633 $loc = lc $loc;
284102e8 634 foreach my $enc (@enc) {
dd8482fc
JH
635 trylocale("$loc.$enc");
636 }
637 }
638 foreach my $lang (split(/ /, $language_codes)) {
639 trylocale($lang);
640 foreach my $country (split(/ /, $country_codes)) {
641 my $lc = "${lang}_${country}";
642 trylocale($lc);
643 foreach my $enc (@enc) {
644 trylocale("$lc.$enc");
645 }
646 my $lC = "${lang}_\U${country}";
647 trylocale($lC);
648 foreach my $enc (@enc) {
649 trylocale("$lC.$enc");
650 }
284102e8
JH
651 }
652 }
653 }
654}
4599a1de 655
d43ce814
JH
656setlocale(LC_ALL, "C");
657
86f50d7d 658if ($^O eq 'darwin') {
4373e181 659 # Darwin 8/Mac OS X 10.4 and 10.5 have bad Basque locales: perl bug #35895,
86f50d7d 660 # Apple bug ID# 4139653. It also has a problem in Byelorussian.
4373e181
RGS
661 (my $v) = $Config{osvers} =~ /^(\d+)/;
662 if ($v >= 8 and $v < 10) {
86f50d7d 663 debug "# Skipping eu_ES, be_BY locales -- buggy in Darwin\n";
a5ec937f 664 @Locale = grep ! m/^(eu_ES(?:\..*)?|be_BY\.CP1131)$/, @Locale;
dfa5c78f 665 } elsif ($v < 12) {
a44d0896
NC
666 debug "# Skipping be_BY locales -- buggy in Darwin\n";
667 @Locale = grep ! m/^be_BY\.CP1131$/, @Locale;
a5ec937f 668 }
86f50d7d
DD
669}
670
4599a1de
JH
671@Locale = sort @Locale;
672
887ef7ed
PP
673debug "# Locales =\n";
674for ( @Locale ) {
675 debug "# $_\n";
676}
8ebc5c01 677
284102e8 678my %Problem;
2a680da6
JH
679my %Okay;
680my %Testing;
30032ef4 681my @Added_alpha; # Alphas that aren't in the C locale.
c08acc4c 682my %test_names;
284102e8 683
30032ef4
KW
684sub report_result {
685 my ($Locale, $i, $pass_fail, $message) = @_;
15bbd6a2
KW
686 $message //= "";
687 $message = " ($message)" if $message;
30032ef4 688 unless ($pass_fail) {
2a680da6 689 $Problem{$i}{$Locale} = 1;
baae13cb 690 debug "# failed $i ($test_names{$i}) with locale '$Locale'$message\n";
2a680da6
JH
691 } else {
692 push @{$Okay{$i}}, $Locale;
693 }
694}
695
7c844d17
KW
696sub report_multi_result {
697 my ($Locale, $i, $results_ref) = @_;
698
699 # $results_ref points to an array, each element of which is a character that was
700 # in error for this test numbered '$i'. If empty, the test passed
701
702 my $message = "";
703 if (@$results_ref) {
704 $message = join " ", "for", map { sprintf '\\x%02X', ord $_ } @$results_ref;
705 }
706 report_result($Locale, $i, @$results_ref == 0, $message);
707}
708
c4093d7d
KW
709my $first_locales_test_number = $final_without_setlocale + 1;
710my $locales_test_number;
711my $not_necessarily_a_problem_test_number;
6d5d702a
KW
712my $first_casing_test_number;
713my $final_casing_test_number;
c4093d7d
KW
714my %setlocale_failed; # List of locales that setlocale() didn't work on
715
284102e8 716foreach $Locale (@Locale) {
c4093d7d 717 $locales_test_number = $first_locales_test_number - 1;
284102e8 718 debug "# Locale = $Locale\n";
284102e8
JH
719
720 unless (setlocale(LC_ALL, $Locale)) {
c4093d7d 721 $setlocale_failed{$Locale} = $Locale;
284102e8 722 next;
8ebc5c01 723 }
8ebc5c01 724
66cbab2c
KW
725 # We test UTF-8 locales only under ':not_characters'; otherwise they have
726 # documented deficiencies. Non- UTF-8 locales are tested only under plain
727 # 'use locale', as otherwise we would have to convert everything in them
728 # to Unicode.
729 my $is_utf8_locale = $Locale =~ /UTF-?8/i;
730
95eaa1bf
KW
731 my %UPPER = (); # All alpha X for which uc(X) == X and lc(X) != X
732 my %lower = (); # All alpha X for which lc(X) == X and uc(X) != X
733 my %BoThCaSe = (); # All alpha X for which uc(X) == lc(X) == X
66cbab2c
KW
734
735 if (! $is_utf8_locale) {
736 use locale;
71e5cbb3 737 @Alnum_ = sort grep /\w/, map { chr } 0..255;
f07538a5 738
71e5cbb3 739 debug "# w = ", join("",@Alnum_), "\n";
e5272a46 740
71e5cbb3
KW
741 # Sieve the uppercase and the lowercase.
742
743 for (@Alnum_) {
744 if (/[^\d_]/) { # skip digits and the _
745 if (uc($_) eq $_) {
746 $UPPER{$_} = $_;
747 }
748 if (lc($_) eq $_) {
749 $lower{$_} = $_;
750 }
751 }
752 }
66cbab2c
KW
753 }
754 else {
755 use locale ':not_characters';
756 @Alnum_ = sort grep /\w/, map { chr } 0..255;
757 debug "# w = ", join("",@Alnum_), "\n";
758 for (@Alnum_) {
759 if (/[^\d_]/) { # skip digits and the _
760 if (uc($_) eq $_) {
761 $UPPER{$_} = $_;
762 }
763 if (lc($_) eq $_) {
764 $lower{$_} = $_;
765 }
766 }
767 }
768 }
284102e8 769 foreach (keys %UPPER) {
097ee67d 770 $BoThCaSe{$_}++ if exists $lower{$_};
284102e8
JH
771 }
772 foreach (keys %lower) {
097ee67d 773 $BoThCaSe{$_}++ if exists $UPPER{$_};
284102e8 774 }
097ee67d 775 foreach (keys %BoThCaSe) {
284102e8
JH
776 delete $UPPER{$_};
777 delete $lower{$_};
778 }
779
db4b7445
A
780 debug "# UPPER = ", join("", sort keys %UPPER ), "\n";
781 debug "# lower = ", join("", sort keys %lower ), "\n";
782 debug "# BoThCaSe = ", join("", sort keys %BoThCaSe), "\n";
284102e8 783
baa71cfd 784 my @failures;
3da38613 785 my @fold_failures;
baa71cfd
KW
786 foreach my $x (sort keys %UPPER) {
787 my $ok;
3da38613 788 my $fold_ok;
baa71cfd
KW
789 if ($is_utf8_locale) {
790 use locale ':not_characters';
791 $ok = $x =~ /[[:upper:]]/;
3da38613 792 $fold_ok = $x =~ /[[:lower:]]/i;
baa71cfd
KW
793 }
794 else {
795 use locale;
796 $ok = $x =~ /[[:upper:]]/;
3da38613 797 $fold_ok = $x =~ /[[:lower:]]/i;
baa71cfd
KW
798 }
799 push @failures, $x unless $ok;
3da38613 800 push @fold_failures, $x unless $fold_ok;
baa71cfd 801 }
baa71cfd 802 $locales_test_number++;
6d5d702a 803 $first_casing_test_number = $locales_test_number;
95eaa1bf 804 $test_names{$locales_test_number} = 'Verify that /[[:upper:]]/ matches all alpha X for which uc(X) == X and lc(X) != X';
7c844d17 805 report_multi_result($Locale, $locales_test_number, \@failures);
6d5d702a 806
3da38613 807 $locales_test_number++;
6d5d702a 808
95eaa1bf 809 $test_names{$locales_test_number} = 'Verify that /[[:lower:]]/i matches all alpha X for which uc(X) == X and lc(X) != X';
7c844d17 810 report_multi_result($Locale, $locales_test_number, \@fold_failures);
baa71cfd 811
baa71cfd 812 undef @failures;
3da38613 813 undef @fold_failures;
baa71cfd
KW
814
815 foreach my $x (sort keys %lower) {
816 my $ok;
3da38613 817 my $fold_ok;
baa71cfd
KW
818 if ($is_utf8_locale) {
819 use locale ':not_characters';
820 $ok = $x =~ /[[:lower:]]/;
3da38613 821 $fold_ok = $x =~ /[[:upper:]]/i;
baa71cfd
KW
822 }
823 else {
824 use locale;
825 $ok = $x =~ /[[:lower:]]/;
3da38613 826 $fold_ok = $x =~ /[[:upper:]]/i;
baa71cfd
KW
827 }
828 push @failures, $x unless $ok;
3da38613 829 push @fold_failures, $x unless $fold_ok;
baa71cfd
KW
830 }
831
832 $locales_test_number++;
95eaa1bf 833 $test_names{$locales_test_number} = 'Verify that /[[:lower:]]/ matches all alpha X for which lc(X) == X and uc(X) != X';
7c844d17
KW
834 report_multi_result($Locale, $locales_test_number, \@failures);
835
3da38613 836 $locales_test_number++;
95eaa1bf 837 $test_names{$locales_test_number} = 'Verify that /[[:upper:]]/i matches all alpha X for which lc(X) == X and uc(X) != X';
7c844d17 838 report_multi_result($Locale, $locales_test_number, \@fold_failures);
baa71cfd 839
9445c837
KW
840 { # Find the alphabetic characters that are not considered alphabetics
841 # in the default (C) locale.
8ebc5c01 842
284102e8 843 no locale;
71e5cbb3 844
30032ef4 845 @Added_alpha = ();
5e7a1028 846 for (keys %UPPER, keys %lower, keys %BoThCaSe) {
30032ef4 847 push(@Added_alpha, $_) if (/\W/);
284102e8 848 }
8ebc5c01 849 }
8ebc5c01 850
30032ef4 851 @Added_alpha = sort @Added_alpha;
8ebc5c01 852
30032ef4 853 debug "# Added_alpha = ", join("",@Added_alpha), "\n";
8ebc5c01 854
db31898d 855 # Cross-check the whole 8-bit character set.
8ebc5c01 856
db31898d
KW
857 ++$locales_test_number;
858 my @f;
859 $test_names{$locales_test_number} = 'Verify that \w and [:word:] are identical';
860 for (map { chr } 0..255) {
861 if ($is_utf8_locale) {
862 use locale ':not_characters';
863 push @f, $_ unless /[[:word:]]/ == /\w/;
864 }
865 else {
866 push @f, $_ unless /[[:word:]]/ == /\w/;
867 }
868 }
869 report_multi_result($Locale, $locales_test_number, \@f);
8ebc5c01 870
db31898d
KW
871 ++$locales_test_number;
872 undef @f;
873 $test_names{$locales_test_number} = 'Verify that \d and [:digit:] are identical';
874 for (map { chr } 0..255) {
875 if ($is_utf8_locale) {
876 use locale ':not_characters';
877 push @f, $_ unless /[[:digit:]]/ == /\d/;
878 }
879 else {
880 push @f, $_ unless /[[:digit:]]/ == /\d/;
881 }
882 }
883 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 884
db31898d
KW
885 ++$locales_test_number;
886 undef @f;
887 $test_names{$locales_test_number} = 'Verify that \s and [:space:] are identical';
888 for (map { chr } 0..255) {
889 if ($is_utf8_locale) {
890 use locale ':not_characters';
891 push @f, $_ unless /[[:space:]]/ == /\s/;
892 }
893 else {
894 push @f, $_ unless /[[:space:]]/ == /\s/;
895 }
896 }
897 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 898
db31898d
KW
899 ++$locales_test_number;
900 undef @f;
901 $test_names{$locales_test_number} = 'Verify that [:posix:] and [:^posix:] are mutually exclusive';
902 for (map { chr } 0..255) {
903 if ($is_utf8_locale) {
904 use locale ':not_characters';
905 push @f, $_ unless (/[[:alpha:]]/ xor /[[:^alpha:]]/) ||
906 (/[[:alnum:]]/ xor /[[:^alnum:]]/) ||
907 (/[[:ascii:]]/ xor /[[:^ascii:]]/) ||
908 (/[[:blank:]]/ xor /[[:^blank:]]/) ||
909 (/[[:cntrl:]]/ xor /[[:^cntrl:]]/) ||
910 (/[[:digit:]]/ xor /[[:^digit:]]/) ||
911 (/[[:graph:]]/ xor /[[:^graph:]]/) ||
912 (/[[:lower:]]/ xor /[[:^lower:]]/) ||
913 (/[[:print:]]/ xor /[[:^print:]]/) ||
914 (/[[:space:]]/ xor /[[:^space:]]/) ||
915 (/[[:upper:]]/ xor /[[:^upper:]]/) ||
916 (/[[:word:]]/ xor /[[:^word:]]/) ||
917 (/[[:xdigit:]]/ xor /[[:^xdigit:]]/) ||
918
919 # effectively is what [:cased:] would be if it existed.
920 (/[[:upper:]]/i xor /[[:^upper:]]/i);
921 }
922 else {
923 push @f, $_ unless (/[[:alpha:]]/ xor /[[:^alpha:]]/) ||
924 (/[[:alnum:]]/ xor /[[:^alnum:]]/) ||
925 (/[[:ascii:]]/ xor /[[:^ascii:]]/) ||
926 (/[[:blank:]]/ xor /[[:^blank:]]/) ||
927 (/[[:cntrl:]]/ xor /[[:^cntrl:]]/) ||
928 (/[[:digit:]]/ xor /[[:^digit:]]/) ||
929 (/[[:graph:]]/ xor /[[:^graph:]]/) ||
930 (/[[:lower:]]/ xor /[[:^lower:]]/) ||
931 (/[[:print:]]/ xor /[[:^print:]]/) ||
932 (/[[:space:]]/ xor /[[:^space:]]/) ||
933 (/[[:upper:]]/ xor /[[:^upper:]]/) ||
934 (/[[:word:]]/ xor /[[:^word:]]/) ||
935 (/[[:xdigit:]]/ xor /[[:^xdigit:]]/) ||
936 (/[[:upper:]]/i xor /[[:^upper:]]/i);
937 }
938 }
939 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 940
db31898d
KW
941 # The rules for the relationships are given in:
942 # http://www.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap07.html
32e8aa3f 943
db31898d
KW
944 ++$locales_test_number;
945 undef @f;
946 $test_names{$locales_test_number} = 'Verify that [:lower:] is a subset of [:alpha:]';
947 for (map { chr } 0..255) {
948 if ($is_utf8_locale) {
949 use locale ':not_characters';
950 push @f, $_ if /[[:lower:]]/ and ! /[[:alpha:]]/;
951 }
952 else {
953 push @f, $_ if /[[:lower:]]/ and ! /[[:alpha:]]/;
954 }
955 }
956 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 957
db31898d
KW
958 ++$locales_test_number;
959 undef @f;
960 $test_names{$locales_test_number} = 'Verify that [:upper:] is a subset of [:alpha:]';
961 for (map { chr } 0..255) {
962 if ($is_utf8_locale) {
963 use locale ':not_characters';
964 push @f, $_ if /[[:upper:]]/ and ! /[[:alpha:]]/;
965 }
966 else {
967 push @f, $_ if /[[:upper:]]/ and ! /[[:alpha:]]/;
968 }
969 }
970 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 971
db31898d
KW
972 ++$locales_test_number;
973 undef @f;
974 $test_names{$locales_test_number} = 'Verify that /[[:lower:]]/i is a subset of [:alpha:]';
975 for (map { chr } 0..255) {
976 if ($is_utf8_locale) {
977 use locale ':not_characters';
978 push @f, $_ if /[[:lower:]]/i and ! /[[:alpha:]]/;
979 }
980 else {
981 push @f, $_ if /[[:lower:]]/i and ! /[[:alpha:]]/;
982 }
983 }
984 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 985
db31898d
KW
986 ++$locales_test_number;
987 undef @f;
988 $test_names{$locales_test_number} = 'Verify that [:alpha:] is a subset of [:alnum:]';
989 for (map { chr } 0..255) {
990 if ($is_utf8_locale) {
991 use locale ':not_characters';
992 push @f, $_ if /[[:alpha:]]/ and ! /[[:alnum:]]/;
993 }
994 else {
995 push @f, $_ if /[[:alpha:]]/ and ! /[[:alnum:]]/;
996 }
997 }
998 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 999
db31898d
KW
1000 ++$locales_test_number;
1001 undef @f;
1002 $test_names{$locales_test_number} = 'Verify that [:digit:] is a subset of [:alnum:]';
1003 for (map { chr } 0..255) {
1004 if ($is_utf8_locale) {
1005 use locale ':not_characters';
1006 push @f, $_ if /[[:digit:]]/ and ! /[[:alnum:]]/;
1007 }
1008 else {
1009 push @f, $_ if /[[:digit:]]/ and ! /[[:alnum:]]/;
1010 }
1011 }
1012 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1013
db31898d
KW
1014 ++$locales_test_number;
1015 undef @f;
1016 $test_names{$locales_test_number} = 'Verify that [:digit:] is a subset of [:xdigit:]';
1017 for (map { chr } 0..255) {
1018 if ($is_utf8_locale) {
1019 use locale ':not_characters';
1020 push @f, $_ if /[[:digit:]]/ and ! /[[:xdigit:]]/;
1021 }
1022 else {
1023 push @f, $_ if /[[:digit:]]/ and ! /[[:xdigit:]]/;
1024 }
1025 }
1026 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1027
db31898d
KW
1028 ++$locales_test_number;
1029 undef @f;
1030 $test_names{$locales_test_number} = 'Verify that [:alnum:] is a subset of [:graph:]';
1031 for (map { chr } 0..255) {
1032 if ($is_utf8_locale) {
1033 use locale ':not_characters';
1034 push @f, $_ if /[[:alnum:]]/ and ! /[[:graph:]]/;
1035 }
1036 else {
1037 push @f, $_ if /[[:alnum:]]/ and ! /[[:graph:]]/;
1038 }
1039 }
1040 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1041
db31898d 1042 # Note that xdigit doesn't have to be a subset of alnum
32e8aa3f 1043
db31898d
KW
1044 ++$locales_test_number;
1045 undef @f;
1046 $test_names{$locales_test_number} = 'Verify that [:xdigit:] is a subset of [:graph:]';
1047 for (map { chr } 0..255) {
1048 if ($is_utf8_locale) {
1049 use locale ':not_characters';
1050 push @f, $_ if /[[:xdigit:]]/ and ! /[[:graph:]]/;
1051 }
1052 else {
1053 push @f, $_ if /[[:xdigit:]]/ and ! /[[:graph:]]/;
1054 }
1055 }
1056 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1057
db31898d
KW
1058 ++$locales_test_number;
1059 undef @f;
1060 $test_names{$locales_test_number} = 'Verify that [:punct:] is a subset of [:graph:]';
1061 for (map { chr } 0..255) {
1062 if ($is_utf8_locale) {
1063 use locale ':not_characters';
1064 push @f, $_ if /[[:punct:]]/ and ! /[[:graph:]]/;
1065 }
1066 else {
1067 push @f, $_ if /[[:punct:]]/ and ! /[[:graph:]]/;
1068 }
1069 }
1070 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1071
db31898d
KW
1072 ++$locales_test_number;
1073 undef @f;
1074 $test_names{$locales_test_number} = 'Verify that [:blank:] is a subset of [:space:]';
1075 for (map { chr } 0..255) {
1076 if ($is_utf8_locale) {
1077 use locale ':not_characters';
1078 push @f, $_ if /[[:blank:]]/ and ! /[[:space:]]/;
1079 }
1080 else {
1081 push @f, $_ if /[[:blank:]]/ and ! /[[:space:]]/;
1082 }
1083 }
1084 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1085
db31898d
KW
1086 ++$locales_test_number;
1087 undef @f;
1088 $test_names{$locales_test_number} = 'Verify that [:graph:] is a subset of [:print:]';
1089 for (map { chr } 0..255) {
1090 if ($is_utf8_locale) {
1091 use locale ':not_characters';
1092 push @f, $_ if /[[:graph:]]/ and ! /[[:print:]]/;
1093 }
1094 else {
1095 push @f, $_ if /[[:graph:]]/ and ! /[[:print:]]/;
1096 }
1097 }
1098 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1099
db31898d
KW
1100 ++$locales_test_number;
1101 undef @f;
1102 $test_names{$locales_test_number} = 'Verify that isn\'t both [:cntrl:] and [:print:]';
1103 for (map { chr } 0..255) {
1104 if ($is_utf8_locale) {
1105 use locale ':not_characters';
1106 push @f, $_ if (/[[:print:]]/ and /[[:cntrl:]]/);
1107 }
1108 else {
1109 push @f, $_ if (/[[:print:]]/ and /[[:cntrl:]]/);
1110 }
1111 }
1112 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1113
db31898d
KW
1114 ++$locales_test_number;
1115 undef @f;
1116 $test_names{$locales_test_number} = 'Verify that isn\'t both [:alnum:] and [:punct:]';
1117 for (map { chr } 0..255) {
1118 if ($is_utf8_locale) {
1119 use locale ':not_characters';
1120 push @f, $_ if /[[:alnum:]]/ and /[[:punct:]]/;
1121 }
1122 else {
1123 push @f, $_ if /[[:alnum:]]/ and /[[:punct:]]/;
1124 }
1125 }
1126 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1127
db31898d
KW
1128 ++$locales_test_number;
1129 undef @f;
1130 $test_names{$locales_test_number} = 'Verify that isn\'t both [:xdigit:] and [:punct:]';
1131 for (map { chr } 0..255) {
1132 if ($is_utf8_locale) {
1133 use locale ':not_characters';
1134 push @f, $_ if (/[[:punct:]]/ and /[[:xdigit:]]/);
1135 }
1136 else {
1137 push @f, $_ if (/[[:punct:]]/ and /[[:xdigit:]]/);
1138 }
1139 }
1140 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1141
db31898d
KW
1142 ++$locales_test_number;
1143 undef @f;
1144 $test_names{$locales_test_number} = 'Verify that isn\'t both [:graph:] and [:space:]';
1145 for (map { chr } 0..255) {
1146 if ($is_utf8_locale) {
1147 use locale ':not_characters';
1148 push @f, $_ if (/[[:graph:]]/ and /[[:space:]]/);
1149 }
1150 else {
1151 push @f, $_ if (/[[:graph:]]/ and /[[:space:]]/);
1152 }
1153 }
1154 report_multi_result($Locale, $locales_test_number, \@f);
32e8aa3f 1155
db31898d 1156 $final_casing_test_number = $locales_test_number;
32e8aa3f 1157
db31898d
KW
1158 # Test for read-only scalars' locale vs non-locale comparisons.
1159
1160 {
1161 no locale;
1162 my $ok;
1163 $a = "qwerty";
1164 if ($is_utf8_locale) {
1165 use locale ':not_characters';
1166 $ok = ($a cmp "qwerty") == 0;
1167 }
1168 else {
1169 use locale;
1170 $ok = ($a cmp "qwerty") == 0;
1171 }
1172 report_result($Locale, ++$locales_test_number, $ok);
1173 $test_names{$locales_test_number} = 'Verify that cmp works with a read-only scalar; no- vs locale';
1174 }
8ebc5c01 1175
db31898d
KW
1176 {
1177 my ($from, $to, $lesser, $greater,
1178 @test, %test, $test, $yes, $no, $sign);
284102e8 1179
db31898d
KW
1180 ++$locales_test_number;
1181 $test_names{$locales_test_number} = 'Verify that "le", "ne", etc work';
1182 $not_necessarily_a_problem_test_number = $locales_test_number;
1183 for (0..9) {
1184 # Select a slice.
1185 $from = int(($_*@Alnum_)/10);
1186 $to = $from + int(@Alnum_/10);
1187 $to = $#Alnum_ if ($to > $#Alnum_);
1188 $lesser = join('', @Alnum_[$from..$to]);
1189 # Select a slice one character on.
1190 $from++; $to++;
1191 $to = $#Alnum_ if ($to > $#Alnum_);
1192 $greater = join('', @Alnum_[$from..$to]);
66cbab2c
KW
1193 if ($is_utf8_locale) {
1194 use locale ':not_characters';
db31898d
KW
1195 ($yes, $no, $sign) = ($lesser lt $greater
1196 ? (" ", "not ", 1)
1197 : ("not ", " ", -1));
66cbab2c
KW
1198 }
1199 else {
1200 use locale;
db31898d
KW
1201 ($yes, $no, $sign) = ($lesser lt $greater
1202 ? (" ", "not ", 1)
1203 : ("not ", " ", -1));
66cbab2c 1204 }
db31898d
KW
1205 # all these tests should FAIL (return 0). Exact lt or gt cannot
1206 # be tested because in some locales, say, eacute and E may test
1207 # equal.
1208 @test =
1209 (
1210 $no.' ($lesser le $greater)', # 1
1211 'not ($lesser ne $greater)', # 2
1212 ' ($lesser eq $greater)', # 3
1213 $yes.' ($lesser ge $greater)', # 4
1214 $yes.' ($lesser ge $greater)', # 5
1215 $yes.' ($greater le $lesser )', # 7
1216 'not ($greater ne $lesser )', # 8
1217 ' ($greater eq $lesser )', # 9
1218 $no.' ($greater ge $lesser )', # 10
1219 'not (($lesser cmp $greater) == -($sign))' # 11
1220 );
1221 @test{@test} = 0 x @test;
1222 $test = 0;
1223 for my $ti (@test) {
66cbab2c
KW
1224 if ($is_utf8_locale) {
1225 use locale ':not_characters';
db31898d 1226 $test{$ti} = eval $ti;
66cbab2c
KW
1227 }
1228 else {
db31898d
KW
1229 # Already in 'use locale';
1230 $test{$ti} = eval $ti;
66cbab2c 1231 }
db31898d
KW
1232 $test ||= $test{$ti}
1233 }
1234 report_result($Locale, $locales_test_number, $test == 0);
1235 if ($test) {
1236 debug "# lesser = '$lesser'\n";
1237 debug "# greater = '$greater'\n";
1238 debug "# lesser cmp greater = ",
1239 $lesser cmp $greater, "\n";
1240 debug "# greater cmp lesser = ",
1241 $greater cmp $lesser, "\n";
1242 debug "# (greater) from = $from, to = $to\n";
1243 for my $ti (@test) {
1244 debugf("# %-40s %-4s", $ti,
1245 $test{$ti} ? 'FAIL' : 'ok');
1246 if ($ti =~ /\(\.*(\$.+ +cmp +\$[^\)]+)\.*\)/) {
1247 debugf("(%s == %4d)", $1, eval $1);
66cbab2c 1248 }
db31898d
KW
1249 debug "\n#";
1250 }
1251
1252 last;
1253 }
1254 }
1255 }
c4093d7d 1256
66cbab2c
KW
1257 my $ok1;
1258 my $ok2;
1259 my $ok3;
1260 my $ok4;
1261 my $ok5;
1262 my $ok6;
1263 my $ok7;
1264 my $ok8;
1265 my $ok9;
1266 my $ok10;
1267 my $ok11;
1268 my $ok12;
1269 my $ok13;
1500bd91 1270 my $ok14;
28acfe03
KW
1271 my $ok15;
1272 my $ok16;
66cbab2c
KW
1273
1274 my $c;
1275 my $d;
1276 my $e;
1277 my $f;
1278 my $g;
1279
1280 if (! $is_utf8_locale) {
71e5cbb3 1281 use locale;
6be75cd7 1282
71e5cbb3 1283 my ($x, $y) = (1.23, 1.23);
6be75cd7 1284
71e5cbb3
KW
1285 $a = "$x";
1286 printf ''; # printf used to reset locale to "C"
1287 $b = "$y";
1288 $ok1 = $a eq $b;
6be75cd7 1289
71e5cbb3
KW
1290 $c = "$x";
1291 my $z = sprintf ''; # sprintf used to reset locale to "C"
1292 $d = "$y";
1293 $ok2 = $c eq $d;
1294 {
66cbab2c 1295
71e5cbb3
KW
1296 use warnings;
1297 my $w = 0;
1298 local $SIG{__WARN__} =
1299 sub {
1300 print "# @_\n";
1301 $w++;
1302 };
6be75cd7 1303
71e5cbb3
KW
1304 # The == (among other ops) used to warn for locales
1305 # that had something else than "." as the radix character.
6be75cd7 1306
71e5cbb3
KW
1307 $ok3 = $c == 1.23;
1308 $ok4 = $c == $x;
1309 $ok5 = $c == $d;
1310 {
1311 no locale;
66cbab2c 1312
b79536ea 1313 $e = "$x";
71e5cbb3
KW
1314
1315 $ok6 = $e == 1.23;
1316 $ok7 = $e == $x;
1317 $ok8 = $e == $c;
1318 }
66cbab2c 1319
71e5cbb3
KW
1320 $f = "1.23";
1321 $g = 2.34;
66cbab2c 1322
71e5cbb3
KW
1323 $ok9 = $f == 1.23;
1324 $ok10 = $f == $x;
1325 $ok11 = $f == $c;
1326 $ok12 = abs(($f + $g) - 3.57) < 0.01;
1327 $ok13 = $w == 0;
28acfe03 1328 $ok14 = $ok15 = $ok16 = 1; # Skip for non-utf8 locales
71e5cbb3 1329 }
66cbab2c
KW
1330 }
1331 else {
1332 use locale ':not_characters';
1333
1334 my ($x, $y) = (1.23, 1.23);
1335 $a = "$x";
1336 printf ''; # printf used to reset locale to "C"
1337 $b = "$y";
1338 $ok1 = $a eq $b;
1339
1340 $c = "$x";
1341 my $z = sprintf ''; # sprintf used to reset locale to "C"
1342 $d = "$y";
1343 $ok2 = $c eq $d;
1344 {
1345 use warnings;
1346 my $w = 0;
1347 local $SIG{__WARN__} =
1348 sub {
1349 print "# @_\n";
1350 $w++;
1351 };
1352 $ok3 = $c == 1.23;
1353 $ok4 = $c == $x;
1354 $ok5 = $c == $d;
1355 {
1356 no locale;
b79536ea 1357 $e = "$x";
66cbab2c
KW
1358
1359 $ok6 = $e == 1.23;
1360 $ok7 = $e == $x;
1361 $ok8 = $e == $c;
1362 }
1363
1364 $f = "1.23";
1365 $g = 2.34;
1366
1367 $ok9 = $f == 1.23;
1368 $ok10 = $f == $x;
1369 $ok11 = $f == $c;
1370 $ok12 = abs(($f + $g) - 3.57) < 0.01;
1371 $ok13 = $w == 0;
1500bd91
KW
1372
1373 # Look for non-ASCII error messages, and verify that the first
1374 # such is in UTF-8 (the others almost certainly will be like the
1375 # first).
1376 $ok14 = 1;
1377 foreach my $err (keys %!) {
1378 use Errno;
1379 $! = eval "&Errno::$err"; # Convert to strerror() output
1380 my $strerror = "$!";
1381 if ("$strerror" =~ /\P{ASCII}/) {
1382 my $utf8_strerror = $strerror;
1383 utf8::upgrade($utf8_strerror);
1384
1385 # If $! was already in UTF-8, the upgrade was a no-op;
1386 # otherwise they will be different byte strings.
1387 use bytes;
1388 $ok14 = $utf8_strerror eq $strerror;
1389 last;
1390 }
1391 }
28acfe03
KW
1392
1393 # Similarly, we verify that a non-ASCII radix is in UTF-8. This
1394 # also catches if there is a disparity between sprintf and
1395 # stringification.
1396
1397 my $string_g = "$g";
1398
1399 my $utf8_string_g = "$g";
1400 utf8::upgrade($utf8_string_g);
1401
1402 my $utf8_sprintf_g = sprintf("%g", $g);
1403 utf8::upgrade($utf8_sprintf_g);
1404 use bytes;
1405 $ok15 = $utf8_string_g eq $string_g;
1406 $ok16 = $utf8_sprintf_g eq $string_g;
66cbab2c
KW
1407 }
1408 }
1409
30032ef4 1410 report_result($Locale, ++$locales_test_number, $ok1);
66cbab2c
KW
1411 $test_names{$locales_test_number} = 'Verify that an intervening printf doesn\'t change assignment results';
1412 my $first_a_test = $locales_test_number;
1413
1414 debug "# $first_a_test..$locales_test_number: \$a = $a, \$b = $b, Locale = $Locale\n";
1415
30032ef4 1416 report_result($Locale, ++$locales_test_number, $ok2);
66cbab2c
KW
1417 $test_names{$locales_test_number} = 'Verify that an intervening sprintf doesn\'t change assignment results';
1418
1419 my $first_c_test = $locales_test_number;
1420
30032ef4 1421 report_result($Locale, ++$locales_test_number, $ok3);
71e5cbb3 1422 $test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a constant';
6be75cd7 1423
30032ef4 1424 report_result($Locale, ++$locales_test_number, $ok4);
71e5cbb3 1425 $test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a scalar';
66cbab2c 1426
30032ef4 1427 report_result($Locale, ++$locales_test_number, $ok5);
71e5cbb3 1428 $test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a scalar and an intervening sprintf';
66cbab2c 1429
71e5cbb3 1430 debug "# $first_c_test..$locales_test_number: \$c = $c, \$d = $d, Locale = $Locale\n";
66cbab2c 1431
30032ef4 1432 report_result($Locale, ++$locales_test_number, $ok6);
b79536ea 1433 $test_names{$locales_test_number} = 'Verify that can assign stringified under inner no-locale block';
71e5cbb3 1434 my $first_e_test = $locales_test_number;
6be75cd7 1435
30032ef4 1436 report_result($Locale, ++$locales_test_number, $ok7);
71e5cbb3 1437 $test_names{$locales_test_number} = 'Verify that "==" with a scalar still works in inner no locale';
66cbab2c 1438
30032ef4 1439 report_result($Locale, ++$locales_test_number, $ok8);
71e5cbb3 1440 $test_names{$locales_test_number} = 'Verify that "==" with a scalar and an intervening sprintf still works in inner no locale';
c4093d7d 1441
71e5cbb3 1442 debug "# $first_e_test..$locales_test_number: \$e = $e, no locale\n";
2a680da6 1443
30032ef4 1444 report_result($Locale, ++$locales_test_number, $ok9);
71e5cbb3
KW
1445 $test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix still works when doing "==" with a constant';
1446 my $first_f_test = $locales_test_number;
6be75cd7 1447
30032ef4 1448 report_result($Locale, ++$locales_test_number, $ok10);
71e5cbb3 1449 $test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix still works when doing "==" with a scalar';
66cbab2c 1450
30032ef4 1451 report_result($Locale, ++$locales_test_number, $ok11);
71e5cbb3 1452 $test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix still works when doing "==" with a scalar and an intervening sprintf';
906f284f 1453
30032ef4 1454 report_result($Locale, ++$locales_test_number, $ok12);
71e5cbb3 1455 $test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix can participate in an addition and function call as numeric';
c4093d7d 1456
30032ef4 1457 report_result($Locale, ++$locales_test_number, $ok13);
71e5cbb3 1458 $test_names{$locales_test_number} = 'Verify that don\'t get warning under "==" even if radix is not a dot';
c4093d7d 1459
30032ef4 1460 report_result($Locale, ++$locales_test_number, $ok14);
1500bd91
KW
1461 $test_names{$locales_test_number} = 'Verify that non-ASCII UTF-8 error messages are in UTF-8';
1462
30032ef4 1463 report_result($Locale, ++$locales_test_number, $ok15);
28acfe03
KW
1464 $test_names{$locales_test_number} = 'Verify that a number with a UTF-8 radix has a UTF-8 stringification';
1465
30032ef4 1466 report_result($Locale, ++$locales_test_number, $ok16);
28acfe03
KW
1467 $test_names{$locales_test_number} = 'Verify that a sprintf of a number with a UTF-8 radix yields UTF-8';
1468
71e5cbb3 1469 debug "# $first_f_test..$locales_test_number: \$f = $f, \$g = $g, back to locale = $Locale\n";
906f284f 1470
26d80d95
LC
1471 # Does taking lc separately differ from taking
1472 # the lc "in-line"? (This was the bug 19990704.002, change #3568.)
1473 # The bug was in the caching of the 'o'-magic.
66cbab2c 1474 if (! $is_utf8_locale) {
2a680da6 1475 use locale;
6be75cd7 1476
2a680da6
JH
1477 sub lcA {
1478 my $lc0 = lc $_[0];
1479 my $lc1 = lc $_[1];
1480 return $lc0 cmp $lc1;
1481 }
6be75cd7 1482
2a680da6
JH
1483 sub lcB {
1484 return lc($_[0]) cmp lc($_[1]);
1485 }
6be75cd7 1486
2a680da6
JH
1487 my $x = "ab";
1488 my $y = "aa";
1489 my $z = "AB";
6be75cd7 1490
30032ef4 1491 report_result($Locale, ++$locales_test_number,
2a680da6
JH
1492 lcA($x, $y) == 1 && lcB($x, $y) == 1 ||
1493 lcA($x, $z) == 0 && lcB($x, $z) == 0);
6be75cd7 1494 }
66cbab2c
KW
1495 else {
1496 use locale ':not_characters';
1497
1498 sub lcC {
1499 my $lc0 = lc $_[0];
1500 my $lc1 = lc $_[1];
1501 return $lc0 cmp $lc1;
1502 }
1503
1504 sub lcD {
1505 return lc($_[0]) cmp lc($_[1]);
1506 }
1507
1508 my $x = "ab";
1509 my $y = "aa";
1510 my $z = "AB";
1511
30032ef4 1512 report_result($Locale, ++$locales_test_number,
66cbab2c
KW
1513 lcC($x, $y) == 1 && lcD($x, $y) == 1 ||
1514 lcC($x, $z) == 0 && lcD($x, $z) == 0);
1515 }
1516 $test_names{$locales_test_number} = 'Verify "lc(foo) cmp lc(bar)" is the same as using intermediaries for the cmp';
d8093b23 1517
26d80d95
LC
1518 # Does lc of an UPPER (if different from the UPPER) match
1519 # case-insensitively the UPPER, and does the UPPER match
1520 # case-insensitively the lc of the UPPER. And vice versa.
3ba0e062 1521 {
ef4a39e5
JH
1522 use locale;
1523 no utf8;
1524 my $re = qr/[\[\(\{\*\+\?\|\^\$\\]/;
1525
1526 my @f = ();
c4093d7d 1527 ++$locales_test_number;
c08acc4c 1528 $test_names{$locales_test_number} = 'Verify case insensitive matching works';
f78d9f29 1529 foreach my $x (sort keys %UPPER) {
66cbab2c 1530 if (! $is_utf8_locale) {
71e5cbb3
KW
1531 my $y = lc $x;
1532 next unless uc $y eq $x;
1533 print "# UPPER $x lc $y ",
faf0c248
KW
1534 $x =~ /$y/i ? 1 : 0, " ",
1535 $y =~ /$x/i ? 1 : 0, "\n" if 0;
71e5cbb3
KW
1536 #
1537 # If $x and $y contain regular expression characters
1538 # AND THEY lowercase (/i) to regular expression characters,
1539 # regcomp() will be mightily confused. No, the \Q doesn't
1540 # help here (maybe regex engine internal lowercasing
1541 # is done after the \Q?) An example of this happening is
1542 # the bg_BG (Bulgarian) locale under EBCDIC (OS/390 USS):
1543 # the chr(173) (the "[") is the lowercase of the chr(235).
1544 #
1545 # Similarly losing EBCDIC locales include cs_cz, cs_CZ,
1546 # el_gr, el_GR, en_us.IBM-037 (!), en_US.IBM-037 (!),
1547 # et_ee, et_EE, hr_hr, hr_HR, hu_hu, hu_HU, lt_LT,
1548 # mk_mk, mk_MK, nl_nl.IBM-037, nl_NL.IBM-037,
1549 # pl_pl, pl_PL, ro_ro, ro_RO, ru_ru, ru_RU,
1550 # sk_sk, sk_SK, sl_si, sl_SI, tr_tr, tr_TR.
1551 #
1552 # Similar things can happen even under (bastardised)
1553 # non-EBCDIC locales: in many European countries before the
1554 # advent of ISO 8859-x nationally customised versions of
1555 # ISO 646 were devised, reusing certain punctuation
1556 # characters for modified characters needed by the
1557 # country/language. For example, the "|" might have
1558 # stood for U+00F6 or LATIN SMALL LETTER O WITH DIAERESIS.
1559 #
1560 if ($x =~ $re || $y =~ $re) {
1561 print "# Regex characters in '$x' or '$y', skipping test $locales_test_number for locale '$Locale'\n";
1562 next;
1563 }
1564 # With utf8 both will fail since the locale concept
1565 # of upper/lower does not work well in Unicode.
1566 push @f, $x unless $x =~ /$y/i == $y =~ /$x/i;
26c1569f
KW
1567
1568 # fc is not a locale concept, so Perl uses lc for it.
1569 push @f, $x unless lc $x eq fc $x;
66cbab2c
KW
1570 }
1571 else {
1572 use locale ':not_characters';
1573 my $y = lc $x;
1574 next unless uc $y eq $x;
1575 print "# UPPER $x lc $y ",
faf0c248
KW
1576 $x =~ /$y/i ? 1 : 0, " ",
1577 $y =~ /$x/i ? 1 : 0, "\n" if 0;
66cbab2c
KW
1578
1579 # Here, we can fully test things, unlike plain 'use locale',
1580 # because this form does work well with Unicode
1581 push @f, $x unless $x =~ /$y/i && $y =~ /$x/i;
26c1569f
KW
1582
1583 # The places where Unicode's lc is different from fc are
1584 # skipped here by virtue of the 'next unless uc...' line above
1585 push @f, $x unless lc $x eq fc $x;
66cbab2c 1586 }
c00ff1c7 1587 }
ef4a39e5 1588
f78d9f29 1589 foreach my $x (sort keys %lower) {
66cbab2c 1590 if (! $is_utf8_locale) {
71e5cbb3
KW
1591 my $y = uc $x;
1592 next unless lc $y eq $x;
1593 print "# lower $x uc $y ",
faf0c248
KW
1594 $x =~ /$y/i ? 1 : 0, " ",
1595 $y =~ /$x/i ? 1 : 0, "\n" if 0;
71e5cbb3
KW
1596 if ($x =~ $re || $y =~ $re) { # See above.
1597 print "# Regex characters in '$x' or '$y', skipping test $locales_test_number for locale '$Locale'\n";
1598 next;
1599 }
1600 # With utf8 both will fail since the locale concept
1601 # of upper/lower does not work well in Unicode.
1602 push @f, $x unless $x =~ /$y/i == $y =~ /$x/i;
26c1569f
KW
1603
1604 push @f, $x unless lc $x eq fc $x;
66cbab2c
KW
1605 }
1606 else {
1607 use locale ':not_characters';
1608 my $y = uc $x;
1609 next unless lc $y eq $x;
1610 print "# lower $x uc $y ",
faf0c248
KW
1611 $x =~ /$y/i ? 1 : 0, " ",
1612 $y =~ /$x/i ? 1 : 0, "\n" if 0;
66cbab2c 1613 push @f, $x unless $x =~ /$y/i && $y =~ /$x/i;
26c1569f
KW
1614
1615 push @f, $x unless lc $x eq fc $x;
66cbab2c 1616 }
c00ff1c7 1617 }
7c844d17 1618 report_multi_result($Locale, $locales_test_number, \@f);
d8093b23 1619 }
78787052
JL
1620
1621 # [perl #109318]
1622 {
1623 my @f = ();
1624 ++$locales_test_number;
1625 $test_names{$locales_test_number} = 'Verify atof with locale radix and negative exponent';
1626
1627 my $radix = POSIX::localeconv()->{decimal_point};
1628 my @nums = (
1629 "3.14e+9", "3${radix}14e+9", "3.14e-9", "3${radix}14e-9",
1630 "-3.14e+9", "-3${radix}14e+9", "-3.14e-9", "-3${radix}14e-9",
1631 );
1632
1633 if (! $is_utf8_locale) {
1634 use locale;
1635 for my $num (@nums) {
1636 push @f, $num
1637 unless sprintf("%g", $num) =~ /3.+14/;
1638 }
1639 }
1640 else {
1641 use locale ':not_characters';
1642 for my $num (@nums) {
1643 push @f, $num
1644 unless sprintf("%g", $num) =~ /3.+14/;
1645 }
1646 }
1647
30032ef4 1648 report_result($Locale, $locales_test_number, @f == 0);
78787052
JL
1649 if (@f) {
1650 print "# failed $locales_test_number locale '$Locale' numbers @f\n"
1651 }
1652 }
8ebc5c01 1653}
284102e8 1654
c4093d7d 1655my $final_locales_test_number = $locales_test_number;
6cf0b567 1656
2a680da6
JH
1657# Recount the errors.
1658
c4093d7d
KW
1659foreach ($first_locales_test_number..$final_locales_test_number) {
1660 if (%setlocale_failed) {
1661 print "not ";
1662 }
1663 elsif ($Problem{$_} || !defined $Okay{$_} || !@{$Okay{$_}}) {
1664 if (defined $not_necessarily_a_problem_test_number
1665 && $_ == $not_necessarily_a_problem_test_number)
1666 {
1667 print "# The failure of test $not_necessarily_a_problem_test_number is not necessarily fatal.\n";
b4e009be 1668 print "# It usually indicates a problem in the environment,\n";
284102e8
JH
1669 print "# not in Perl itself.\n";
1670 }
6d5d702a
KW
1671 if ($Okay{$_} && ($_ >= $first_casing_test_number
1672 && $_ <= $final_casing_test_number))
1673 {
0a974e2d
KW
1674 # Round to nearest .1%
1675 my $percent_fail = (int(.5 + (1000 * scalar(keys $Problem{$_})
1676 / scalar(@Locale))))
1677 / 10;
6d5d702a
KW
1678 if ($percent_fail < $acceptable_fold_failure_percentage) {
1679 $test_names{$_} .= 'TODO';
1680 print "# ", 100 - $percent_fail, "% of locales pass the following test, so it is likely that the failures\n";
1681 print "# are errors in the locale definitions. The test is marked TODO, as the\n";
1682 print "# problem is not likely to be Perl's\n";
1683 }
1684 }
108a305e
KW
1685 unless ($debug) {
1686 print "#\nFor more details, rerun, with environment variable PERL_DEBUG_FULL_TEST=1\n";
1687 }
284102e8 1688 print "not ";
8ebc5c01 1689 }
c4093d7d 1690 print "ok $_";
6c2e653d
KW
1691 if (defined $test_names{$_}) {
1692 # If TODO is in the test name, make it thus
1693 my $todo = $test_names{$_} =~ s/TODO\s*//;
1694 print " $test_names{$_}";
1695 print " # TODO" if $todo;
1696 }
c4093d7d 1697 print "\n";
8ebc5c01 1698}
fb73857a 1699
2a680da6
JH
1700# Give final advice.
1701
284102e8
JH
1702my $didwarn = 0;
1703
c4093d7d 1704foreach ($first_locales_test_number..$final_locales_test_number) {
284102e8
JH
1705 if ($Problem{$_}) {
1706 my @f = sort keys %{ $Problem{$_} };
1707 my $f = join(" ", @f);
1708 $f =~ s/(.{50,60}) /$1\n#\t/g;
2a680da6
JH
1709 print
1710 "#\n",
1711 "# The locale ", (@f == 1 ? "definition" : "definitions"), "\n#\n",
284102e8
JH
1712 "#\t", $f, "\n#\n",
1713 "# on your system may have errors because the locale test $_\n",
baae13cb 1714 "# \"$test_names{$_}\"\n",
284102e8
JH
1715 "# failed in ", (@f == 1 ? "that locale" : "those locales"),
1716 ".\n";
2a680da6 1717 print <<EOW;
284102e8
JH
1718#
1719# If your users are not using these locales you are safe for the moment,
1720# but please report this failure first to perlbug\@perl.com using the
1721# perlbug script (as described in the INSTALL file) so that the exact
1722# details of the failures can be sorted out first and then your operating
1723# system supplier can be alerted about these anomalies.
1724#
1725EOW
1726 $didwarn = 1;
fb73857a
PP
1727 }
1728}
774d564b 1729
26d80d95 1730# Tell which locales were okay and which were not.
2a680da6 1731
284102e8 1732if ($didwarn) {
26d80d95 1733 my (@s, @F);
71e5cbb3 1734
284102e8
JH
1735 foreach my $l (@Locale) {
1736 my $p = 0;
c4093d7d
KW
1737 if ($setlocale_failed{$l}) {
1738 $p++;
1739 }
1740 else {
1f5852c9
KW
1741 foreach my $t
1742 ($first_locales_test_number..$final_locales_test_number)
1743 {
1744 $p++ if $Problem{$t}{$l};
1745 }
c4093d7d 1746 }
284102e8 1747 push @s, $l if $p == 0;
9445c837 1748 push @F, $l unless $p == 0;
8ebc5c01 1749 }
71e5cbb3 1750
68d47915
CK
1751 if (@s) {
1752 my $s = join(" ", @s);
1753 $s =~ s/(.{50,60}) /$1\n#\t/g;
1754
1755 warn
1756 "# The following locales\n#\n",
1757 "#\t", $s, "\n#\n",
1758 "# tested okay.\n#\n",
1759 } else {
26d80d95
LC
1760 warn "# None of your locales were fully okay.\n";
1761 }
1762
1763 if (@F) {
1764 my $F = join(" ", @F);
1765 $F =~ s/(.{50,60}) /$1\n#\t/g;
1766
1767 warn
1768 "# The following locales\n#\n",
0e053d1e 1769 "#\t", $F, "\n#\n",
26d80d95
LC
1770 "# had problems.\n#\n",
1771 } else {
1772 warn "# None of your locales were broken.\n";
68d47915 1773 }
8ebc5c01 1774}
90248788 1775
c4093d7d 1776$test_num = $final_locales_test_number;
c213d471 1777
fbd840df
KW
1778{ # perl #115808
1779 use warnings;
1780 my $warned = 0;
1781 local $SIG{__WARN__} = sub {
1782 $warned = $_[0] =~ /uninitialized/;
1783 };
1784 my $z = "y" . setlocale(&POSIX::LC_ALL, "xyzzy");
1785 ok($warned, "variable set to setlocale(BAD LOCALE) is considered uninitialized");
1786}
1787
094a2f8c 1788# Test that tainting and case changing works on utf8 strings. These tests are
1f5852c9
KW
1789# placed last to avoid disturbing the hard-coded test numbers that existed at
1790# the time these were added above this in this file.
0099bb8d
KW
1791# This also tests that locale overrides unicode_strings in the same scope for
1792# non-utf8 strings.
094a2f8c
KW
1793setlocale(LC_ALL, "C");
1794{
1795 use locale;
0099bb8d 1796 use feature 'unicode_strings';
094a2f8c 1797
26c1569f 1798 foreach my $function ("uc", "ucfirst", "lc", "lcfirst", "fc") {
094a2f8c
KW
1799 my @list; # List of code points to test for $function
1800
1801 # Used to calculate the changed case for ASCII characters by using the
1802 # ord, instead of using one of the functions under test.
1803 my $ascii_case_change_delta;
1804 my $above_latin1_case_change_delta; # Same for the specific ords > 255
1805 # that we use
1806
1807 # We test an ASCII character, which should change case and be tainted;
1808 # a Latin1 character, which shouldn't change case under this C locale,
1809 # and is tainted.
1810 # an above-Latin1 character that when the case is changed would cross
1811 # the 255/256 boundary, so doesn't change case and isn't tainted
1812 # (the \x{149} is one of these, but changes into 2 characters, the
1813 # first one of which doesn't cross the boundary.
1814 # the final one in each list is an above-Latin1 character whose case
1815 # does change, and shouldn't be tainted. The code below uses its
1816 # position in its list as a marker to indicate that it, unlike the
1817 # other code points above ASCII, has a successful case change
1818 if ($function =~ /^u/) {
094a2f8c
KW
1819 @list = ("", "a", "\xe0", "\xff", "\x{fb00}", "\x{149}", "\x{101}");
1820 $ascii_case_change_delta = -32;
1821 $above_latin1_case_change_delta = -1;
1822 }
1823 else {
1ca267a5 1824 @list = ("", "A", "\xC0", "\x{17F}", "\x{100}");
094a2f8c
KW
1825 $ascii_case_change_delta = +32;
1826 $above_latin1_case_change_delta = +1;
1827 }
66cbab2c 1828 foreach my $is_utf8_locale (0 .. 1) {
71e5cbb3
KW
1829 foreach my $j (0 .. $#list) {
1830 my $char = $list[$j];
0099bb8d
KW
1831
1832 for my $encoded_in_utf8 (0 .. 1) {
faf0c248
KW
1833 my $should_be;
1834 my $changed;
1835 if (! $is_utf8_locale) {
1836 $should_be = ($j == $#list)
1837 ? chr(ord($char) + $above_latin1_case_change_delta)
1838 : (length $char == 0 || ord($char) > 127)
1839 ? $char
1840 : chr(ord($char) + $ascii_case_change_delta);
1841
1842 # This monstrosity is in order to avoid using an eval,
1843 # which might perturb the results
1844 $changed = ($function eq "uc")
1845 ? uc($char)
1846 : ($function eq "ucfirst")
1847 ? ucfirst($char)
1848 : ($function eq "lc")
1849 ? lc($char)
1850 : ($function eq "lcfirst")
1851 ? lcfirst($char)
26c1569f
KW
1852 : ($function eq "fc")
1853 ? fc($char)
faf0c248
KW
1854 : die("Unexpected function \"$function\"");
1855 }
1856 else {
1857 {
1858 no locale;
71e5cbb3 1859
faf0c248
KW
1860 # For utf8-locales the case changing functions
1861 # should work just like they do outside of locale.
1862 # Can use eval here because not testing it when
1863 # not in locale.
1864 $should_be = eval "$function('$char')";
1865 die "Unexpected eval error $@ from 'eval \"$function('$char')\"'" if $@;
71e5cbb3 1866
faf0c248
KW
1867 }
1868 use locale ':not_characters';
1869 $changed = ($function eq "uc")
1870 ? uc($char)
1871 : ($function eq "ucfirst")
1872 ? ucfirst($char)
1873 : ($function eq "lc")
1874 ? lc($char)
1875 : ($function eq "lcfirst")
1876 ? lcfirst($char)
26c1569f
KW
1877 : ($function eq "fc")
1878 ? fc($char)
faf0c248 1879 : die("Unexpected function \"$function\"");
71e5cbb3 1880 }
faf0c248
KW
1881 ok($changed eq $should_be,
1882 "$function(\"$char\") in C locale "
1883 . (($is_utf8_locale)
1884 ? "(use locale ':not_characters'"
1885 : "(use locale")
1886 . (($encoded_in_utf8)
1887 ? "; encoded in utf8)"
1888 : "; not encoded in utf8)")
1889 . " should be \"$should_be\", got \"$changed\"");
1890
1891 # Tainting shouldn't happen for utf8 locales, empty
1892 # strings, or those characters above 255.
1893 (! $is_utf8_locale && length($char) > 0 && ord($char) < 256)
1894 ? check_taint($changed)
1895 : check_taint_not($changed);
1896
1897 # Use UTF-8 next time through the loop
1898 utf8::upgrade($char);
0099bb8d 1899 }
66cbab2c 1900 }
094a2f8c
KW
1901 }
1902 }
1903}
1904
fdf053ee 1905print "1..$test_num\n";
906f284f 1906
90248788 1907# eof