add TODO tests for slow our() declaration
[p5sagit/p5-mst-13.2.git] / lib / locale.t
CommitLineData
8ebc5c01 1#!./perl -wT
2
3BEGIN {
4 chdir 't' if -d 't';
20822f61 5 @INC = '../lib';
f9cbebe1 6 unshift @INC, '.';
b002077a 7 require Config; import Config;
97a0514d 8 if (!$Config{d_setlocale} || $Config{ccflags} =~ /\bD?NO_LOCALE\b/) {
b002077a 9 print "1..0\n";
10 exit;
11 }
2de3dbcc 12 $| = 1;
8ebc5c01 13}
14
15use strict;
16
284102e8 17my $debug = 1;
18
db4b7445 19use Dumpvalue;
20
21my $dumper = Dumpvalue->new(
22 tick => qq{"},
23 quoteHighBit => 0,
24 unctrl => "quote"
25 );
6be75cd7 26sub debug {
db4b7445 27 return unless $debug;
28 my($mess) = join "", @_;
29 chop $mess;
30 print $dumper->stringify($mess,1), "\n";
6be75cd7 31}
32
33sub debugf {
34 printf @_ if $debug;
35}
36
8ebc5c01 37my $have_setlocale = 0;
38eval {
39 require POSIX;
40 import POSIX ':locale_h';
41 $have_setlocale++;
42};
43
6dead956 44# Visual C's CRT goes silly on strings of the form "en_US.ISO8859-1"
f6c6487a 45# and mingw32 uses said silly CRT
2986a63f 46$have_setlocale = 0 if (($^O eq 'MSWin32' || $^O eq 'NetWare') && $Config{cc} =~ /^(cl|gcc)/i);
6dead956 47
906f284f 48my $last = $have_setlocale ? &last : &last_without_setlocale;
26d80d95 49
50print "1..$last\n";
8ebc5c01 51
097ee67d 52use vars qw(&LC_ALL);
8ebc5c01 53
0e053d1e 54$a = 'abc %';
8ebc5c01 55
56sub ok {
57 my ($n, $result) = @_;
58
59 print 'not ' unless ($result);
60 print "ok $n\n";
61}
62
63# First we'll do a lot of taint checking for locales.
64# This is the easiest to test, actually, as any locale,
65# even the default locale will taint under 'use locale'.
66
67sub is_tainted { # hello, camel two.
9f1b1f2d 68 no warnings 'uninitialized' ;
8ebc5c01 69 my $dummy;
70 not eval { $dummy = join("", @_), kill 0; 1 }
71}
72
73sub check_taint ($$) {
74 ok $_[0], is_tainted($_[1]);
75}
76
77sub check_taint_not ($$) {
78 ok $_[0], not is_tainted($_[1]);
79}
80
81use locale; # engage locale and therefore locale taint.
82
83check_taint_not 1, $a;
84
85check_taint 2, uc($a);
86check_taint 3, "\U$a";
87check_taint 4, ucfirst($a);
88check_taint 5, "\u$a";
89check_taint 6, lc($a);
90check_taint 7, "\L$a";
91check_taint 8, lcfirst($a);
92check_taint 9, "\l$a";
93
ff9121f8 94check_taint_not 10, sprintf('%e', 123.456);
95check_taint_not 11, sprintf('%f', 123.456);
96check_taint_not 12, sprintf('%g', 123.456);
8ebc5c01 97check_taint_not 13, sprintf('%d', 123.456);
98check_taint_not 14, sprintf('%x', 123.456);
99
100$_ = $a; # untaint $_
101
102$_ = uc($a); # taint $_
103
104check_taint 15, $_;
105
106/(\w)/; # taint $&, $`, $', $+, $1.
107check_taint 16, $&;
108check_taint 17, $`;
109check_taint 18, $';
110check_taint 19, $+;
111check_taint 20, $1;
112check_taint_not 21, $2;
113
114/(.)/; # untaint $&, $`, $', $+, $1.
115check_taint_not 22, $&;
116check_taint_not 23, $`;
117check_taint_not 24, $';
118check_taint_not 25, $+;
119check_taint_not 26, $1;
120check_taint_not 27, $2;
121
122/(\W)/; # taint $&, $`, $', $+, $1.
123check_taint 28, $&;
124check_taint 29, $`;
125check_taint 30, $';
126check_taint 31, $+;
127check_taint 32, $1;
128check_taint_not 33, $2;
129
130/(\s)/; # taint $&, $`, $', $+, $1.
131check_taint 34, $&;
132check_taint 35, $`;
133check_taint 36, $';
134check_taint 37, $+;
135check_taint 38, $1;
136check_taint_not 39, $2;
137
138/(\S)/; # taint $&, $`, $', $+, $1.
139check_taint 40, $&;
140check_taint 41, $`;
141check_taint 42, $';
142check_taint 43, $+;
143check_taint 44, $1;
144check_taint_not 45, $2;
145
146$_ = $a; # untaint $_
147
148check_taint_not 46, $_;
149
150/(b)/; # this must not taint
151check_taint_not 47, $&;
152check_taint_not 48, $`;
153check_taint_not 49, $';
154check_taint_not 50, $+;
155check_taint_not 51, $1;
156check_taint_not 52, $2;
157
158$_ = $a; # untaint $_
159
160check_taint_not 53, $_;
161
162$b = uc($a); # taint $b
163s/(.+)/$b/; # this must taint only the $_
164
165check_taint 54, $_;
166check_taint_not 55, $&;
167check_taint_not 56, $`;
168check_taint_not 57, $';
169check_taint_not 58, $+;
170check_taint_not 59, $1;
171check_taint_not 60, $2;
172
173$_ = $a; # untaint $_
174
175s/(.+)/b/; # this must not taint
176check_taint_not 61, $_;
177check_taint_not 62, $&;
178check_taint_not 63, $`;
179check_taint_not 64, $';
180check_taint_not 65, $+;
181check_taint_not 66, $1;
182check_taint_not 67, $2;
183
184$b = $a; # untaint $b
185
186($b = $a) =~ s/\w/$&/;
187check_taint 68, $b; # $b should be tainted.
188check_taint_not 69, $a; # $a should be not.
189
190$_ = $a; # untaint $_
191
192s/(\w)/\l$1/; # this must taint
193check_taint 70, $_;
194check_taint 71, $&;
195check_taint 72, $`;
196check_taint 73, $';
197check_taint 74, $+;
198check_taint 75, $1;
199check_taint_not 76, $2;
200
201$_ = $a; # untaint $_
202
203s/(\w)/\L$1/; # this must taint
204check_taint 77, $_;
205check_taint 78, $&;
206check_taint 79, $`;
207check_taint 80, $';
208check_taint 81, $+;
209check_taint 82, $1;
210check_taint_not 83, $2;
211
212$_ = $a; # untaint $_
213
214s/(\w)/\u$1/; # this must taint
215check_taint 84, $_;
216check_taint 85, $&;
217check_taint 86, $`;
218check_taint 87, $';
219check_taint 88, $+;
220check_taint 89, $1;
221check_taint_not 90, $2;
222
223$_ = $a; # untaint $_
224
225s/(\w)/\U$1/; # this must taint
226check_taint 91, $_;
227check_taint 92, $&;
228check_taint 93, $`;
229check_taint 94, $';
230check_taint 95, $+;
231check_taint 96, $1;
232check_taint_not 97, $2;
233
234# After all this tainting $a should be cool.
235
236check_taint_not 98, $a;
237
906f284f 238sub last_without_setlocale { 98 }
239
8ebc5c01 240# I think we've seen quite enough of taint.
241# Let us do some *real* locale work now,
284102e8 242# unless setlocale() is missing (i.e. minitest).
8ebc5c01 243
244exit unless $have_setlocale;
245
284102e8 246# Find locales.
247
6be75cd7 248debug "# Scanning for locales...\n";
249
250# Note that it's okay that some languages have their native names
251# capitalized here even though that's not "right". They are lowercased
252# anyway later during the scanning process (and besides, some clueless
253# vendor might have them capitalized errorneously anyway).
254
284102e8 255my $locales = <<EOF;
6be75cd7 256Afrikaans:af:za:1 15
284102e8 257Arabic:ar:dz eg sa:6 arabic8
6be75cd7 258Brezhoneg Breton:br:fr:1 15
259Bulgarski Bulgarian:bg:bg:5
dd8482fc 260Chinese:zh:cn tw:cn.EUC eucCN eucTW euc.CN euc.TW Big5 GB2312 tw.EUC
6be75cd7 261Hrvatski Croatian:hr:hr:2
262Cymraeg Welsh:cy:cy:1 14 15
284102e8 263Czech:cs:cz:2
6be75cd7 264Dansk Danish:dk:da:1 15
265Nederlands Dutch:nl:be nl:1 15
dd8482fc 266English American British:en:au ca gb ie nz us uk zw:1 15 cp850
6be75cd7 267Esperanto:eo:eo:3
268Eesti Estonian:et:ee:4 6 13
269Suomi Finnish:fi:fi:1 15
270Flamish::fl:1 15
6be75cd7 271Deutsch German:de:at be ch de lu:1 15
272Euskaraz Basque:eu:es fr:1 15
6be75cd7 273Galego Galician:gl:es:1 15
274Ellada Greek:el:gr:7 g8
6be75cd7 275Frysk:fy:nl:1 15
276Greenlandic:kl:gl:4 6
284102e8 277Hebrew:iw:il:8 hebrew8
278Hungarian:hu:hu:2
6be75cd7 279Indonesian:in:id:1 15
280Gaeilge Irish:ga:IE:1 14 15
281Italiano Italian:it:ch it:1 15
282Nihongo Japanese:ja:jp:euc eucJP jp.EUC sjis
284102e8 283Korean:ko:kr:
6be75cd7 284Latine Latin:la:va:1 15
285Latvian:lv:lv:4 6 13
286Lithuanian:lt:lt:4 6 13
287Macedonian:mk:mk:1 15
288Maltese:mt:mt:3
dd8482fc 289Moldovan:mo:mo:2
290Norsk Norwegian:no no\@nynorsk:no:1 15
6be75cd7 291Occitan:oc:es:1 15
292Polski Polish:pl:pl:2
284102e8 293Rumanian:ro:ro:2
a528dad0 294Russki Russian:ru:ru su ua:5 koi8 koi8r KOI8-R koi8u cp1251 cp866
6be75cd7 295Serbski Serbian:sr:yu:5
284102e8 296Slovak:sk:sk:2
6be75cd7 297Slovene Slovenian:sl:si:2
d43ce814 298Sqhip Albanian:sq:sq:1 15
299Svenska Swedish:sv:fi se:1 15
6be75cd7 300Thai:th:th:11 tis620
284102e8 301Turkish:tr:tr:9 turkish8
dd8482fc 302Yiddish:yi::1 15
284102e8 303EOF
304
ee50adbe 305if ($^O eq 'os390') {
dd8482fc 306 # These cause heartburn. Broken locales?
ee50adbe 307 $locales =~ s/Svenska Swedish:sv:fi se:1 15\n//;
308 $locales =~ s/Thai:th:th:11 tis620\n//;
309}
310
ef4a39e5 311sub in_utf8 () { $^H & 0x08 || (${^OPEN} || "") =~ /:utf8/ }
f9cbebe1 312
313if (in_utf8) {
8a6cb2cb 314 require "lib/locale/utf8";
f9cbebe1 315} else {
8a6cb2cb 316 require "lib/locale/latin1";
f9cbebe1 317}
318
284102e8 319my @Locale;
320my $Locale;
321my @Alnum_;
322
0e053d1e 323my @utf8locale;
324my %utf8skip;
325
284102e8 326sub getalnum_ {
8ebc5c01 327 sort grep /\w/, map { chr } 0..255
328}
329
284102e8 330sub trylocale {
331 my $locale = shift;
332 if (setlocale(LC_ALL, $locale)) {
333 push @Locale, $locale;
334 }
335}
8ebc5c01 336
284102e8 337sub decode_encodings {
338 my @enc;
8ebc5c01 339
284102e8 340 foreach (split(/ /, shift)) {
341 if (/^(\d+)$/) {
342 push @enc, "ISO8859-$1";
343 push @enc, "iso8859$1"; # HP
344 if ($1 eq '1') {
345 push @enc, "roman8"; # HP
346 }
347 } else {
348 push @enc, $_;
dd8482fc 349 push @enc, "$_.UTF-8";
8ebc5c01 350 }
351 }
ee50adbe 352 if ($^O eq 'os390') {
353 push @enc, qw(IBM-037 IBM-819 IBM-1047);
354 }
8ebc5c01 355
284102e8 356 return @enc;
8ebc5c01 357}
358
284102e8 359trylocale("C");
360trylocale("POSIX");
361foreach (0..15) {
362 trylocale("ISO8859-$_");
284102e8 363 trylocale("iso8859$_");
097ee67d 364 trylocale("iso8859-$_");
365 trylocale("iso_8859_$_");
366 trylocale("isolatin$_");
367 trylocale("isolatin-$_");
368 trylocale("iso_latin_$_");
8ebc5c01 369}
370
645e49ed 371# Sanitize the environment so that we can run the external 'locale'
372# program without the taint mode getting grumpy.
cce5967e 373
374# $ENV{PATH} is special in VMS.
375delete $ENV{PATH} if $^O ne 'VMS' or $Config{d_setenv};
376
377# Other subversive stuff.
378delete @ENV{qw(IFS CDPATH ENV BASH_ENV)};
dd8482fc 379
21477fb4 380if (-x "/usr/bin/locale" && open(LOCALES, "/usr/bin/locale -a 2>/dev/null|")) {
dd8482fc 381 while (<LOCALES>) {
382 chomp;
383 trylocale($_);
284102e8 384 }
dd8482fc 385 close(LOCALES);
a6259068 386} elsif ($^O eq 'VMS' && defined($ENV{'SYS$I18N_LOCALE'}) && -d 'SYS$I18N_LOCALE') {
387# The SYS$I18N_LOCALE logical name search list was not present on
388# VAX VMS V5.5-12, but was on AXP && VAX VMS V6.2 as well as later versions.
389 opendir(LOCALES, "SYS\$I18N_LOCALE:");
390 while ($_ = readdir(LOCALES)) {
391 chomp;
392 trylocale($_);
393 }
394 close(LOCALES);
dd8482fc 395} else {
396
397 # This is going to be slow.
398
399 foreach my $locale (split(/\n/, $locales)) {
400 my ($locale_name, $language_codes, $country_codes, $encodings) =
401 split(/:/, $locale);
402 my @enc = decode_encodings($encodings);
403 foreach my $loc (split(/ /, $locale_name)) {
404 trylocale($loc);
284102e8 405 foreach my $enc (@enc) {
dd8482fc 406 trylocale("$loc.$enc");
284102e8 407 }
dd8482fc 408 $loc = lc $loc;
284102e8 409 foreach my $enc (@enc) {
dd8482fc 410 trylocale("$loc.$enc");
411 }
412 }
413 foreach my $lang (split(/ /, $language_codes)) {
414 trylocale($lang);
415 foreach my $country (split(/ /, $country_codes)) {
416 my $lc = "${lang}_${country}";
417 trylocale($lc);
418 foreach my $enc (@enc) {
419 trylocale("$lc.$enc");
420 }
421 my $lC = "${lang}_\U${country}";
422 trylocale($lC);
423 foreach my $enc (@enc) {
424 trylocale("$lC.$enc");
425 }
284102e8 426 }
427 }
428 }
429}
4599a1de 430
d43ce814 431setlocale(LC_ALL, "C");
432
4599a1de 433@Locale = sort @Locale;
434
887ef7ed 435debug "# Locales =\n";
436for ( @Locale ) {
437 debug "# $_\n";
438}
8ebc5c01 439
284102e8 440my %Problem;
2a680da6 441my %Okay;
442my %Testing;
097ee67d 443my @Neoalpha;
d8093b23 444my %Neoalpha;
284102e8 445
2a680da6 446sub tryneoalpha {
447 my ($Locale, $i, $test) = @_;
2a680da6 448 unless ($test) {
449 $Problem{$i}{$Locale} = 1;
450 debug "# failed $i with locale '$Locale'\n";
451 } else {
452 push @{$Okay{$i}}, $Locale;
453 }
454}
455
284102e8 456foreach $Locale (@Locale) {
457 debug "# Locale = $Locale\n";
458 @Alnum_ = getalnum_();
db4b7445 459 debug "# w = ", join("",@Alnum_), "\n";
284102e8 460
461 unless (setlocale(LC_ALL, $Locale)) {
462 foreach (99..103) {
463 $Problem{$_}{$Locale} = -1;
8ebc5c01 464 }
284102e8 465 next;
8ebc5c01 466 }
8ebc5c01 467
284102e8 468 # Sieve the uppercase and the lowercase.
469
097ee67d 470 my %UPPER = ();
471 my %lower = ();
472 my %BoThCaSe = ();
284102e8 473 for (@Alnum_) {
474 if (/[^\d_]/) { # skip digits and the _
475 if (uc($_) eq $_) {
476 $UPPER{$_} = $_;
477 }
478 if (lc($_) eq $_) {
479 $lower{$_} = $_;
480 }
481 }
482 }
483 foreach (keys %UPPER) {
097ee67d 484 $BoThCaSe{$_}++ if exists $lower{$_};
284102e8 485 }
486 foreach (keys %lower) {
097ee67d 487 $BoThCaSe{$_}++ if exists $UPPER{$_};
284102e8 488 }
097ee67d 489 foreach (keys %BoThCaSe) {
284102e8 490 delete $UPPER{$_};
491 delete $lower{$_};
492 }
493
db4b7445 494 debug "# UPPER = ", join("", sort keys %UPPER ), "\n";
495 debug "# lower = ", join("", sort keys %lower ), "\n";
496 debug "# BoThCaSe = ", join("", sort keys %BoThCaSe), "\n";
284102e8 497
498 # Find the alphabets that are not alphabets in the default locale.
8ebc5c01 499
284102e8 500 {
501 no locale;
8ebc5c01 502
284102e8 503 @Neoalpha = ();
504 for (keys %UPPER, keys %lower) {
505 push(@Neoalpha, $_) if (/\W/);
d8093b23 506 $Neoalpha{$_} = $_;
284102e8 507 }
8ebc5c01 508 }
8ebc5c01 509
284102e8 510 @Neoalpha = sort @Neoalpha;
8ebc5c01 511
db4b7445 512 debug "# Neoalpha = ", join("",@Neoalpha), "\n";
8ebc5c01 513
284102e8 514 if (@Neoalpha == 0) {
515 # If we have no Neoalphas the remaining tests are no-ops.
6be75cd7 516 debug "# no Neoalpha, skipping tests 99..102 for locale '$Locale'\n";
a88c3d7c 517 foreach (99..102) {
518 push @{$Okay{$_}}, $Locale;
519 }
6be75cd7 520 } else {
8ebc5c01 521
6be75cd7 522 # Test \w.
284102e8 523
ef4a39e5 524 my $word = join('', @Neoalpha);
8ebc5c01 525
ef4a39e5 526 if ($Locale =~ /utf-?8/i) {
527 debug "# unknown whether locale and Unicode have the same \\w, skipping test 99 for locale '$Locale'\n";
528 push @{$Okay{99}}, $Locale;
529 } else {
530 if ($word =~ /^(\w+)$/) {
531 tryneoalpha($Locale, 99, 1);
532 } else {
533 tryneoalpha($Locale, 99, 0);
534 }
8ac0d9e6 535 }
ef4a39e5 536
2a680da6 537 # Cross-check the whole 8-bit character set.
8ebc5c01 538
6be75cd7 539 for (map { chr } 0..255) {
2a680da6 540 tryneoalpha($Locale, 100,
541 (/\w/ xor /\W/) ||
542 (/\d/ xor /\D/) ||
543 (/\s/ xor /\S/));
284102e8 544 }
8ebc5c01 545
6be75cd7 546 # Test for read-only scalars' locale vs non-locale comparisons.
284102e8 547
284102e8 548 {
6be75cd7 549 no locale;
550 $a = "qwerty";
551 {
552 use locale;
2a680da6 553 tryneoalpha($Locale, 101, ($a cmp "qwerty") == 0);
8ebc5c01 554 }
555 }
8ebc5c01 556
6be75cd7 557 {
558 my ($from, $to, $lesser, $greater,
559 @test, %test, $test, $yes, $no, $sign);
560
561 for (0..9) {
562 # Select a slice.
563 $from = int(($_*@Alnum_)/10);
564 $to = $from + int(@Alnum_/10);
565 $to = $#Alnum_ if ($to > $#Alnum_);
566 $lesser = join('', @Alnum_[$from..$to]);
567 # Select a slice one character on.
568 $from++; $to++;
569 $to = $#Alnum_ if ($to > $#Alnum_);
570 $greater = join('', @Alnum_[$from..$to]);
571 ($yes, $no, $sign) = ($lesser lt $greater
572 ? (" ", "not ", 1)
573 : ("not ", " ", -1));
574 # all these tests should FAIL (return 0).
575 # Exact lt or gt cannot be tested because
576 # in some locales, say, eacute and E may test equal.
577 @test =
578 (
579 $no.' ($lesser le $greater)', # 1
580 'not ($lesser ne $greater)', # 2
581 ' ($lesser eq $greater)', # 3
582 $yes.' ($lesser ge $greater)', # 4
583 $yes.' ($lesser ge $greater)', # 5
584 $yes.' ($greater le $lesser )', # 7
585 'not ($greater ne $lesser )', # 8
586 ' ($greater eq $lesser )', # 9
587 $no.' ($greater ge $lesser )', # 10
0e053d1e 588 'not (($lesser cmp $greater) == -($sign))' # 11
6be75cd7 589 );
590 @test{@test} = 0 x @test;
591 $test = 0;
284102e8 592 for my $ti (@test) {
6be75cd7 593 $test{$ti} = eval $ti;
594 $test ||= $test{$ti}
284102e8 595 }
2a680da6 596 tryneoalpha($Locale, 102, $test == 0);
6be75cd7 597 if ($test) {
6be75cd7 598 debug "# lesser = '$lesser'\n";
599 debug "# greater = '$greater'\n";
600 debug "# lesser cmp greater = ",
601 $lesser cmp $greater, "\n";
602 debug "# greater cmp lesser = ",
603 $greater cmp $lesser, "\n";
604 debug "# (greater) from = $from, to = $to\n";
605 for my $ti (@test) {
606 debugf("# %-40s %-4s", $ti,
607 $test{$ti} ? 'FAIL' : 'ok');
608 if ($ti =~ /\(\.*(\$.+ +cmp +\$[^\)]+)\.*\)/) {
609 debugf("(%s == %4d)", $1, eval $1);
610 }
611 debug "\n#";
612 }
284102e8 613
6be75cd7 614 last;
615 }
284102e8 616 }
8ebc5c01 617 }
618 }
6be75cd7 619
620 use locale;
621
622 my ($x, $y) = (1.23, 1.23);
623
0e053d1e 624 $a = "$x";
6be75cd7 625 printf ''; # printf used to reset locale to "C"
0e053d1e 626 $b = "$y";
6be75cd7 627
2a680da6 628 debug "# 103..107: a = $a, b = $b, Locale = $Locale\n";
629
630 tryneoalpha($Locale, 103, $a eq $b);
6be75cd7 631
632 my $c = "$x";
633 my $z = sprintf ''; # sprintf used to reset locale to "C"
634 my $d = "$y";
635
2a680da6 636 debug "# 104..107: c = $c, d = $d, Locale = $Locale\n";
6be75cd7 637
2a680da6 638 tryneoalpha($Locale, 104, $c eq $d);
6be75cd7 639
2a680da6 640 {
9f1b1f2d 641 use warnings;
2a680da6 642 my $w = 0;
0e053d1e 643 local $SIG{__WARN__} =
644 sub {
906f284f 645 print "# @_\n";
0e053d1e 646 $w++;
647 };
6be75cd7 648
0e053d1e 649 # The == (among other ops) used to warn for locales
650 # that had something else than "." as the radix character.
6be75cd7 651
2a680da6 652 tryneoalpha($Locale, 105, $c == 1.23);
6be75cd7 653
2a680da6 654 tryneoalpha($Locale, 106, $c == $x);
6be75cd7 655
2a680da6 656 tryneoalpha($Locale, 107, $c == $d);
6be75cd7 657
2a680da6 658 {
2de3dbcc 659# no locale; # XXX did this ever work correctly?
6be75cd7 660
2a680da6 661 my $e = "$x";
6be75cd7 662
2a680da6 663 debug "# 108..110: e = $e, Locale = $Locale\n";
6be75cd7 664
2a680da6 665 tryneoalpha($Locale, 108, $e == 1.23);
6be75cd7 666
2a680da6 667 tryneoalpha($Locale, 109, $e == $x);
668
669 tryneoalpha($Locale, 110, $e == $c);
6be75cd7 670 }
2a680da6 671
2a680da6 672 my $f = "1.23";
906f284f 673 my $g = 2.34;
2a680da6 674
906f284f 675 debug "# 111..115: f = $f, g = $g, locale = $Locale\n";
2a680da6 676
906f284f 677 tryneoalpha($Locale, 111, $f == 1.23);
6be75cd7 678
906f284f 679 tryneoalpha($Locale, 112, $f == $x);
2a680da6 680
906f284f 681 tryneoalpha($Locale, 113, $f == $c);
682
683 tryneoalpha($Locale, 114, abs(($f + $g) - 3.57) < 0.01);
684
685 tryneoalpha($Locale, 115, $w == 0);
6be75cd7 686 }
687
26d80d95 688 # Does taking lc separately differ from taking
689 # the lc "in-line"? (This was the bug 19990704.002, change #3568.)
690 # The bug was in the caching of the 'o'-magic.
2a680da6 691 {
692 use locale;
6be75cd7 693
2a680da6 694 sub lcA {
695 my $lc0 = lc $_[0];
696 my $lc1 = lc $_[1];
697 return $lc0 cmp $lc1;
698 }
6be75cd7 699
2a680da6 700 sub lcB {
701 return lc($_[0]) cmp lc($_[1]);
702 }
6be75cd7 703
2a680da6 704 my $x = "ab";
705 my $y = "aa";
706 my $z = "AB";
6be75cd7 707
906f284f 708 tryneoalpha($Locale, 116,
2a680da6 709 lcA($x, $y) == 1 && lcB($x, $y) == 1 ||
710 lcA($x, $z) == 0 && lcB($x, $z) == 0);
6be75cd7 711 }
d8093b23 712
26d80d95 713 # Does lc of an UPPER (if different from the UPPER) match
714 # case-insensitively the UPPER, and does the UPPER match
715 # case-insensitively the lc of the UPPER. And vice versa.
3ba0e062 716 {
ef4a39e5 717 use locale;
718 no utf8;
719 my $re = qr/[\[\(\{\*\+\?\|\^\$\\]/;
720
721 my @f = ();
722 foreach my $x (keys %UPPER) {
723 my $y = lc $x;
724 next unless uc $y eq $x;
725 print "# UPPER $x lc $y ",
726 $x =~ /$y/i ? 1 : 0, " ",
727 $y =~ /$x/i ? 1 : 0, "\n" if 0;
83d38f53 728 #
ef4a39e5 729 # If $x and $y contain regular expression characters
730 # AND THEY lowercase (/i) to regular expression characters,
731 # regcomp() will be mightily confused. No, the \Q doesn't
732 # help here (maybe regex engine internal lowercasing
733 # is done after the \Q?) An example of this happening is
734 # the bg_BG (Bulgarian) locale under EBCDIC (OS/390 USS):
735 # the chr(173) (the "[") is the lowercase of the chr(235).
83d38f53 736 #
ef4a39e5 737 # Similarly losing EBCDIC locales include cs_cz, cs_CZ,
738 # el_gr, el_GR, en_us.IBM-037 (!), en_US.IBM-037 (!),
739 # et_ee, et_EE, hr_hr, hr_HR, hu_hu, hu_HU, lt_LT,
740 # mk_mk, mk_MK, nl_nl.IBM-037, nl_NL.IBM-037,
741 # pl_pl, pl_PL, ro_ro, ro_RO, ru_ru, ru_RU,
742 # sk_sk, sk_SK, sl_si, sl_SI, tr_tr, tr_TR.
83d38f53 743 #
744 # Similar things can happen even under (bastardised)
745 # non-EBCDIC locales: in many European countries before the
746 # advent of ISO 8859-x nationally customised versions of
747 # ISO 646 were devised, reusing certain punctuation
748 # characters for modified characters needed by the
749 # country/language. For example, the "|" might have
750 # stood for U+00F6 or LATIN SMALL LETTER O WITH DIAERESIS.
751 #
ef4a39e5 752 if ($x =~ $re || $y =~ $re) {
753 print "# Regex characters in '$x' or '$y', skipping test 117 for locale '$Locale'\n";
754 next;
8ac0d9e6 755 }
ef4a39e5 756 # With utf8 both will fail since the locale concept
757 # of upper/lower does not work well in Unicode.
758 push @f, $x unless $x =~ /$y/i == $y =~ /$x/i;
759
8ac0d9e6 760 foreach my $x (keys %lower) {
761 my $y = uc $x;
762 next unless lc $y eq $x;
4c53e876 763 print "# lower $x uc $y ",
ef4a39e5 764 $x =~ /$y/i ? 1 : 0, " ",
765 $y =~ /$x/i ? 1 : 0, "\n" if 0;
047e14fb 766 if ($x =~ $re || $y =~ $re) { # See above.
767 print "# Regex characters in '$x' or '$y', skipping test 117 for locale '$Locale'\n";
768 next;
769 }
4c53e876 770 # With utf8 both will fail since the locale concept
771 # of upper/lower does not work well in Unicode.
772 push @f, $x unless $x =~ /$y/i == $y =~ /$x/i;
8ac0d9e6 773 }
906f284f 774 tryneoalpha($Locale, 117, @f == 0);
8ac0d9e6 775 if (@f) {
906f284f 776 print "# failed 117 locale '$Locale' characters @f\n"
8ac0d9e6 777 }
26d80d95 778 }
d8093b23 779 }
8ebc5c01 780}
284102e8 781
2a680da6 782# Recount the errors.
783
906f284f 784foreach (&last_without_setlocale()+1..$last) {
2a680da6 785 if ($Problem{$_} || !defined $Okay{$_} || !@{$Okay{$_}}) {
097ee67d 786 if ($_ == 102) {
787 print "# The failure of test 102 is not necessarily fatal.\n";
b4e009be 788 print "# It usually indicates a problem in the environment,\n";
284102e8 789 print "# not in Perl itself.\n";
790 }
791 print "not ";
8ebc5c01 792 }
284102e8 793 print "ok $_\n";
8ebc5c01 794}
fb73857a 795
2a680da6 796# Give final advice.
797
284102e8 798my $didwarn = 0;
799
26d80d95 800foreach (99..$last) {
284102e8 801 if ($Problem{$_}) {
802 my @f = sort keys %{ $Problem{$_} };
803 my $f = join(" ", @f);
804 $f =~ s/(.{50,60}) /$1\n#\t/g;
2a680da6 805 print
806 "#\n",
807 "# The locale ", (@f == 1 ? "definition" : "definitions"), "\n#\n",
284102e8 808 "#\t", $f, "\n#\n",
809 "# on your system may have errors because the locale test $_\n",
810 "# failed in ", (@f == 1 ? "that locale" : "those locales"),
811 ".\n";
2a680da6 812 print <<EOW;
284102e8 813#
814# If your users are not using these locales you are safe for the moment,
815# but please report this failure first to perlbug\@perl.com using the
816# perlbug script (as described in the INSTALL file) so that the exact
817# details of the failures can be sorted out first and then your operating
818# system supplier can be alerted about these anomalies.
819#
820EOW
821 $didwarn = 1;
fb73857a 822 }
823}
774d564b 824
26d80d95 825# Tell which locales were okay and which were not.
2a680da6 826
284102e8 827if ($didwarn) {
26d80d95 828 my (@s, @F);
284102e8 829
830 foreach my $l (@Locale) {
831 my $p = 0;
26d80d95 832 foreach my $t (102..$last) {
284102e8 833 $p++ if $Problem{$t}{$l};
8ebc5c01 834 }
284102e8 835 push @s, $l if $p == 0;
26d80d95 836 push @F, $l unless $p == 0;
8ebc5c01 837 }
284102e8 838
68d47915 839 if (@s) {
840 my $s = join(" ", @s);
841 $s =~ s/(.{50,60}) /$1\n#\t/g;
842
843 warn
844 "# The following locales\n#\n",
845 "#\t", $s, "\n#\n",
846 "# tested okay.\n#\n",
847 } else {
26d80d95 848 warn "# None of your locales were fully okay.\n";
849 }
850
851 if (@F) {
852 my $F = join(" ", @F);
853 $F =~ s/(.{50,60}) /$1\n#\t/g;
854
855 warn
856 "# The following locales\n#\n",
0e053d1e 857 "#\t", $F, "\n#\n",
26d80d95 858 "# had problems.\n#\n",
859 } else {
860 warn "# None of your locales were broken.\n";
68d47915 861 }
0e053d1e 862
c406fbf4 863 if (@utf8locale) {
864 my $S = join(" ", @utf8locale);
865 $S =~ s/(.{50,60}) /$1\n#\t/g;
866
867 warn "#\n# The following locales\n#\n",
868 "#\t", $S, "\n#\n",
869 "# were skipped for the tests ",
870 join(" ", sort {$a<=>$b} keys %utf8skip), "\n",
871 "# because UTF-8 and locales do not work together in Perl.\n#\n";
872 }
8ebc5c01 873}
90248788 874
906f284f 875sub last { 117 }
876
90248788 877# eof