Commit | Line | Data |
---|---|---|
8ebc5c01 | 1 | #!./perl -wT |
2 | ||
3 | BEGIN { | |
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 CS |
9 | print "1..0\n"; |
10 | exit; | |
11 | } | |
2de3dbcc | 12 | $| = 1; |
8ebc5c01 | 13 | } |
14 | ||
15 | use strict; | |
16 | ||
284102e8 JH |
17 | my $debug = 1; |
18 | ||
db4b7445 A |
19 | use Dumpvalue; |
20 | ||
21 | my $dumper = Dumpvalue->new( | |
22 | tick => qq{"}, | |
23 | quoteHighBit => 0, | |
24 | unctrl => "quote" | |
25 | ); | |
6be75cd7 | 26 | sub debug { |
db4b7445 A |
27 | return unless $debug; |
28 | my($mess) = join "", @_; | |
29 | chop $mess; | |
30 | print $dumper->stringify($mess,1), "\n"; | |
6be75cd7 JH |
31 | } |
32 | ||
33 | sub debugf { | |
34 | printf @_ if $debug; | |
35 | } | |
36 | ||
8ebc5c01 | 37 | my $have_setlocale = 0; |
38 | eval { | |
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 | |
cd19b65c JH |
48 | # UWIN seems to loop after test 98, just skip for now |
49 | $have_setlocale = 0 if ($^O =~ /^uwin/); | |
50 | ||
906f284f | 51 | my $last = $have_setlocale ? &last : &last_without_setlocale; |
26d80d95 LC |
52 | |
53 | print "1..$last\n"; | |
8ebc5c01 | 54 | |
9a66ea41 | 55 | sub LC_ALL (); |
8ebc5c01 | 56 | |
0e053d1e | 57 | $a = 'abc %'; |
8ebc5c01 | 58 | |
59 | sub ok { | |
60 | my ($n, $result) = @_; | |
61 | ||
62 | print 'not ' unless ($result); | |
63 | print "ok $n\n"; | |
64 | } | |
65 | ||
66 | # First we'll do a lot of taint checking for locales. | |
67 | # This is the easiest to test, actually, as any locale, | |
68 | # even the default locale will taint under 'use locale'. | |
69 | ||
70 | sub is_tainted { # hello, camel two. | |
9f1b1f2d | 71 | no warnings 'uninitialized' ; |
8ebc5c01 | 72 | my $dummy; |
73 | not eval { $dummy = join("", @_), kill 0; 1 } | |
74 | } | |
75 | ||
76 | sub check_taint ($$) { | |
77 | ok $_[0], is_tainted($_[1]); | |
78 | } | |
79 | ||
80 | sub check_taint_not ($$) { | |
81 | ok $_[0], not is_tainted($_[1]); | |
82 | } | |
83 | ||
84 | use locale; # engage locale and therefore locale taint. | |
85 | ||
86 | check_taint_not 1, $a; | |
87 | ||
88 | check_taint 2, uc($a); | |
89 | check_taint 3, "\U$a"; | |
90 | check_taint 4, ucfirst($a); | |
91 | check_taint 5, "\u$a"; | |
92 | check_taint 6, lc($a); | |
93 | check_taint 7, "\L$a"; | |
94 | check_taint 8, lcfirst($a); | |
95 | check_taint 9, "\l$a"; | |
96 | ||
ff9121f8 JH |
97 | check_taint_not 10, sprintf('%e', 123.456); |
98 | check_taint_not 11, sprintf('%f', 123.456); | |
99 | check_taint_not 12, sprintf('%g', 123.456); | |
8ebc5c01 | 100 | check_taint_not 13, sprintf('%d', 123.456); |
101 | check_taint_not 14, sprintf('%x', 123.456); | |
102 | ||
103 | $_ = $a; # untaint $_ | |
104 | ||
105 | $_ = uc($a); # taint $_ | |
106 | ||
107 | check_taint 15, $_; | |
108 | ||
109 | /(\w)/; # taint $&, $`, $', $+, $1. | |
110 | check_taint 16, $&; | |
111 | check_taint 17, $`; | |
112 | check_taint 18, $'; | |
113 | check_taint 19, $+; | |
114 | check_taint 20, $1; | |
115 | check_taint_not 21, $2; | |
116 | ||
117 | /(.)/; # untaint $&, $`, $', $+, $1. | |
118 | check_taint_not 22, $&; | |
119 | check_taint_not 23, $`; | |
120 | check_taint_not 24, $'; | |
121 | check_taint_not 25, $+; | |
122 | check_taint_not 26, $1; | |
123 | check_taint_not 27, $2; | |
124 | ||
125 | /(\W)/; # taint $&, $`, $', $+, $1. | |
126 | check_taint 28, $&; | |
127 | check_taint 29, $`; | |
128 | check_taint 30, $'; | |
129 | check_taint 31, $+; | |
130 | check_taint 32, $1; | |
131 | check_taint_not 33, $2; | |
132 | ||
133 | /(\s)/; # taint $&, $`, $', $+, $1. | |
134 | check_taint 34, $&; | |
135 | check_taint 35, $`; | |
136 | check_taint 36, $'; | |
137 | check_taint 37, $+; | |
138 | check_taint 38, $1; | |
139 | check_taint_not 39, $2; | |
140 | ||
141 | /(\S)/; # taint $&, $`, $', $+, $1. | |
142 | check_taint 40, $&; | |
143 | check_taint 41, $`; | |
144 | check_taint 42, $'; | |
145 | check_taint 43, $+; | |
146 | check_taint 44, $1; | |
147 | check_taint_not 45, $2; | |
148 | ||
149 | $_ = $a; # untaint $_ | |
150 | ||
151 | check_taint_not 46, $_; | |
152 | ||
153 | /(b)/; # this must not taint | |
154 | check_taint_not 47, $&; | |
155 | check_taint_not 48, $`; | |
156 | check_taint_not 49, $'; | |
157 | check_taint_not 50, $+; | |
158 | check_taint_not 51, $1; | |
159 | check_taint_not 52, $2; | |
160 | ||
161 | $_ = $a; # untaint $_ | |
162 | ||
163 | check_taint_not 53, $_; | |
164 | ||
165 | $b = uc($a); # taint $b | |
166 | s/(.+)/$b/; # this must taint only the $_ | |
167 | ||
168 | check_taint 54, $_; | |
169 | check_taint_not 55, $&; | |
170 | check_taint_not 56, $`; | |
171 | check_taint_not 57, $'; | |
172 | check_taint_not 58, $+; | |
173 | check_taint_not 59, $1; | |
174 | check_taint_not 60, $2; | |
175 | ||
176 | $_ = $a; # untaint $_ | |
177 | ||
178 | s/(.+)/b/; # this must not taint | |
179 | check_taint_not 61, $_; | |
180 | check_taint_not 62, $&; | |
181 | check_taint_not 63, $`; | |
182 | check_taint_not 64, $'; | |
183 | check_taint_not 65, $+; | |
184 | check_taint_not 66, $1; | |
185 | check_taint_not 67, $2; | |
186 | ||
187 | $b = $a; # untaint $b | |
188 | ||
189 | ($b = $a) =~ s/\w/$&/; | |
190 | check_taint 68, $b; # $b should be tainted. | |
191 | check_taint_not 69, $a; # $a should be not. | |
192 | ||
193 | $_ = $a; # untaint $_ | |
194 | ||
195 | s/(\w)/\l$1/; # this must taint | |
196 | check_taint 70, $_; | |
197 | check_taint 71, $&; | |
198 | check_taint 72, $`; | |
199 | check_taint 73, $'; | |
200 | check_taint 74, $+; | |
201 | check_taint 75, $1; | |
202 | check_taint_not 76, $2; | |
203 | ||
204 | $_ = $a; # untaint $_ | |
205 | ||
206 | s/(\w)/\L$1/; # this must taint | |
207 | check_taint 77, $_; | |
208 | check_taint 78, $&; | |
209 | check_taint 79, $`; | |
210 | check_taint 80, $'; | |
211 | check_taint 81, $+; | |
212 | check_taint 82, $1; | |
213 | check_taint_not 83, $2; | |
214 | ||
215 | $_ = $a; # untaint $_ | |
216 | ||
217 | s/(\w)/\u$1/; # this must taint | |
218 | check_taint 84, $_; | |
219 | check_taint 85, $&; | |
220 | check_taint 86, $`; | |
221 | check_taint 87, $'; | |
222 | check_taint 88, $+; | |
223 | check_taint 89, $1; | |
224 | check_taint_not 90, $2; | |
225 | ||
226 | $_ = $a; # untaint $_ | |
227 | ||
228 | s/(\w)/\U$1/; # this must taint | |
229 | check_taint 91, $_; | |
230 | check_taint 92, $&; | |
231 | check_taint 93, $`; | |
232 | check_taint 94, $'; | |
233 | check_taint 95, $+; | |
234 | check_taint 96, $1; | |
235 | check_taint_not 97, $2; | |
236 | ||
237 | # After all this tainting $a should be cool. | |
238 | ||
239 | check_taint_not 98, $a; | |
240 | ||
906f284f NC |
241 | sub last_without_setlocale { 98 } |
242 | ||
8ebc5c01 | 243 | # I think we've seen quite enough of taint. |
244 | # Let us do some *real* locale work now, | |
284102e8 | 245 | # unless setlocale() is missing (i.e. minitest). |
8ebc5c01 | 246 | |
247 | exit unless $have_setlocale; | |
248 | ||
284102e8 JH |
249 | # Find locales. |
250 | ||
6be75cd7 JH |
251 | debug "# Scanning for locales...\n"; |
252 | ||
253 | # Note that it's okay that some languages have their native names | |
254 | # capitalized here even though that's not "right". They are lowercased | |
255 | # anyway later during the scanning process (and besides, some clueless | |
256 | # vendor might have them capitalized errorneously anyway). | |
257 | ||
284102e8 | 258 | my $locales = <<EOF; |
6be75cd7 | 259 | Afrikaans:af:za:1 15 |
284102e8 | 260 | Arabic:ar:dz eg sa:6 arabic8 |
6be75cd7 JH |
261 | Brezhoneg Breton:br:fr:1 15 |
262 | Bulgarski Bulgarian:bg:bg:5 | |
dd8482fc | 263 | Chinese:zh:cn tw:cn.EUC eucCN eucTW euc.CN euc.TW Big5 GB2312 tw.EUC |
6be75cd7 JH |
264 | Hrvatski Croatian:hr:hr:2 |
265 | Cymraeg Welsh:cy:cy:1 14 15 | |
284102e8 | 266 | Czech:cs:cz:2 |
6be75cd7 JH |
267 | Dansk Danish:dk:da:1 15 |
268 | Nederlands Dutch:nl:be nl:1 15 | |
dd8482fc | 269 | English American British:en:au ca gb ie nz us uk zw:1 15 cp850 |
6be75cd7 JH |
270 | Esperanto:eo:eo:3 |
271 | Eesti Estonian:et:ee:4 6 13 | |
272 | Suomi Finnish:fi:fi:1 15 | |
273 | Flamish::fl:1 15 | |
6be75cd7 JH |
274 | Deutsch German:de:at be ch de lu:1 15 |
275 | Euskaraz Basque:eu:es fr:1 15 | |
6be75cd7 JH |
276 | Galego Galician:gl:es:1 15 |
277 | Ellada Greek:el:gr:7 g8 | |
6be75cd7 JH |
278 | Frysk:fy:nl:1 15 |
279 | Greenlandic:kl:gl:4 6 | |
284102e8 JH |
280 | Hebrew:iw:il:8 hebrew8 |
281 | Hungarian:hu:hu:2 | |
6be75cd7 JH |
282 | Indonesian:in:id:1 15 |
283 | Gaeilge Irish:ga:IE:1 14 15 | |
284 | Italiano Italian:it:ch it:1 15 | |
285 | Nihongo Japanese:ja:jp:euc eucJP jp.EUC sjis | |
284102e8 | 286 | Korean:ko:kr: |
6be75cd7 JH |
287 | Latine Latin:la:va:1 15 |
288 | Latvian:lv:lv:4 6 13 | |
289 | Lithuanian:lt:lt:4 6 13 | |
290 | Macedonian:mk:mk:1 15 | |
291 | Maltese:mt:mt:3 | |
dd8482fc JH |
292 | Moldovan:mo:mo:2 |
293 | Norsk Norwegian:no no\@nynorsk:no:1 15 | |
6be75cd7 JH |
294 | Occitan:oc:es:1 15 |
295 | Polski Polish:pl:pl:2 | |
284102e8 | 296 | Rumanian:ro:ro:2 |
a528dad0 | 297 | Russki Russian:ru:ru su ua:5 koi8 koi8r KOI8-R koi8u cp1251 cp866 |
6be75cd7 | 298 | Serbski Serbian:sr:yu:5 |
284102e8 | 299 | Slovak:sk:sk:2 |
6be75cd7 | 300 | Slovene Slovenian:sl:si:2 |
d43ce814 JH |
301 | Sqhip Albanian:sq:sq:1 15 |
302 | Svenska Swedish:sv:fi se:1 15 | |
6be75cd7 | 303 | Thai:th:th:11 tis620 |
284102e8 | 304 | Turkish:tr:tr:9 turkish8 |
dd8482fc | 305 | Yiddish:yi::1 15 |
284102e8 JH |
306 | EOF |
307 | ||
ee50adbe | 308 | if ($^O eq 'os390') { |
dd8482fc | 309 | # These cause heartburn. Broken locales? |
ee50adbe PP |
310 | $locales =~ s/Svenska Swedish:sv:fi se:1 15\n//; |
311 | $locales =~ s/Thai:th:th:11 tis620\n//; | |
312 | } | |
313 | ||
ef4a39e5 | 314 | sub in_utf8 () { $^H & 0x08 || (${^OPEN} || "") =~ /:utf8/ } |
f9cbebe1 JH |
315 | |
316 | if (in_utf8) { | |
8a6cb2cb | 317 | require "lib/locale/utf8"; |
f9cbebe1 | 318 | } else { |
8a6cb2cb | 319 | require "lib/locale/latin1"; |
f9cbebe1 JH |
320 | } |
321 | ||
284102e8 JH |
322 | my @Locale; |
323 | my $Locale; | |
324 | my @Alnum_; | |
325 | ||
0e053d1e JH |
326 | my @utf8locale; |
327 | my %utf8skip; | |
328 | ||
284102e8 | 329 | sub getalnum_ { |
8ebc5c01 | 330 | sort grep /\w/, map { chr } 0..255 |
331 | } | |
332 | ||
284102e8 JH |
333 | sub trylocale { |
334 | my $locale = shift; | |
335 | if (setlocale(LC_ALL, $locale)) { | |
336 | push @Locale, $locale; | |
337 | } | |
338 | } | |
8ebc5c01 | 339 | |
284102e8 JH |
340 | sub decode_encodings { |
341 | my @enc; | |
8ebc5c01 | 342 | |
284102e8 JH |
343 | foreach (split(/ /, shift)) { |
344 | if (/^(\d+)$/) { | |
345 | push @enc, "ISO8859-$1"; | |
346 | push @enc, "iso8859$1"; # HP | |
347 | if ($1 eq '1') { | |
348 | push @enc, "roman8"; # HP | |
349 | } | |
350 | } else { | |
351 | push @enc, $_; | |
dd8482fc | 352 | push @enc, "$_.UTF-8"; |
8ebc5c01 | 353 | } |
354 | } | |
ee50adbe PP |
355 | if ($^O eq 'os390') { |
356 | push @enc, qw(IBM-037 IBM-819 IBM-1047); | |
357 | } | |
8ebc5c01 | 358 | |
284102e8 | 359 | return @enc; |
8ebc5c01 | 360 | } |
361 | ||
284102e8 JH |
362 | trylocale("C"); |
363 | trylocale("POSIX"); | |
364 | foreach (0..15) { | |
365 | trylocale("ISO8859-$_"); | |
284102e8 | 366 | trylocale("iso8859$_"); |
097ee67d JH |
367 | trylocale("iso8859-$_"); |
368 | trylocale("iso_8859_$_"); | |
369 | trylocale("isolatin$_"); | |
370 | trylocale("isolatin-$_"); | |
371 | trylocale("iso_latin_$_"); | |
8ebc5c01 | 372 | } |
373 | ||
645e49ed JH |
374 | # Sanitize the environment so that we can run the external 'locale' |
375 | # program without the taint mode getting grumpy. | |
cce5967e JH |
376 | |
377 | # $ENV{PATH} is special in VMS. | |
378 | delete $ENV{PATH} if $^O ne 'VMS' or $Config{d_setenv}; | |
379 | ||
380 | # Other subversive stuff. | |
381 | delete @ENV{qw(IFS CDPATH ENV BASH_ENV)}; | |
dd8482fc | 382 | |
21477fb4 | 383 | if (-x "/usr/bin/locale" && open(LOCALES, "/usr/bin/locale -a 2>/dev/null|")) { |
dd8482fc | 384 | while (<LOCALES>) { |
d281a6ac NC |
385 | # It seems that /usr/bin/locale steadfastly outputs 8 bit data, which |
386 | # ain't great when we're running this testPERL_UNICODE= so that utf8 | |
387 | # locales will cause all IO hadles to default to (assume) utf8 | |
388 | next unless utf8::valid($_); | |
dd8482fc JH |
389 | chomp; |
390 | trylocale($_); | |
284102e8 | 391 | } |
dd8482fc | 392 | close(LOCALES); |
a6259068 PP |
393 | } elsif ($^O eq 'VMS' && defined($ENV{'SYS$I18N_LOCALE'}) && -d 'SYS$I18N_LOCALE') { |
394 | # The SYS$I18N_LOCALE logical name search list was not present on | |
395 | # VAX VMS V5.5-12, but was on AXP && VAX VMS V6.2 as well as later versions. | |
396 | opendir(LOCALES, "SYS\$I18N_LOCALE:"); | |
397 | while ($_ = readdir(LOCALES)) { | |
398 | chomp; | |
399 | trylocale($_); | |
400 | } | |
401 | close(LOCALES); | |
87e33296 SP |
402 | } elsif ($^O eq 'openbsd' && -e '/usr/share/locale') { |
403 | ||
404 | # OpenBSD doesn't have a locale executable, so reading /usr/share/locale | |
405 | # is much easier and faster than the last resort method. | |
406 | ||
407 | opendir(LOCALES, '/usr/share/locale'); | |
408 | while ($_ = readdir(LOCALES)) { | |
409 | chomp; | |
410 | trylocale($_); | |
411 | } | |
412 | close(LOCALES); | |
dd8482fc JH |
413 | } else { |
414 | ||
415 | # This is going to be slow. | |
416 | ||
417 | foreach my $locale (split(/\n/, $locales)) { | |
418 | my ($locale_name, $language_codes, $country_codes, $encodings) = | |
419 | split(/:/, $locale); | |
420 | my @enc = decode_encodings($encodings); | |
421 | foreach my $loc (split(/ /, $locale_name)) { | |
422 | trylocale($loc); | |
284102e8 | 423 | foreach my $enc (@enc) { |
dd8482fc | 424 | trylocale("$loc.$enc"); |
284102e8 | 425 | } |
dd8482fc | 426 | $loc = lc $loc; |
284102e8 | 427 | foreach my $enc (@enc) { |
dd8482fc JH |
428 | trylocale("$loc.$enc"); |
429 | } | |
430 | } | |
431 | foreach my $lang (split(/ /, $language_codes)) { | |
432 | trylocale($lang); | |
433 | foreach my $country (split(/ /, $country_codes)) { | |
434 | my $lc = "${lang}_${country}"; | |
435 | trylocale($lc); | |
436 | foreach my $enc (@enc) { | |
437 | trylocale("$lc.$enc"); | |
438 | } | |
439 | my $lC = "${lang}_\U${country}"; | |
440 | trylocale($lC); | |
441 | foreach my $enc (@enc) { | |
442 | trylocale("$lC.$enc"); | |
443 | } | |
284102e8 JH |
444 | } |
445 | } | |
446 | } | |
447 | } | |
4599a1de | 448 | |
d43ce814 JH |
449 | setlocale(LC_ALL, "C"); |
450 | ||
86f50d7d | 451 | if ($^O eq 'darwin') { |
766639d9 | 452 | # Darwin 8/Mac OS X 10.4 has bad Basque locales: perl bug #35895, |
86f50d7d | 453 | # Apple bug ID# 4139653. It also has a problem in Byelorussian. |
1d74b28c | 454 | if ($Config{osvers} ge '8' and $Config{osvers} lt '9') { |
86f50d7d DD |
455 | debug "# Skipping eu_ES, be_BY locales -- buggy in Darwin\n"; |
456 | @Locale = grep ! m/^(eu_ES|be_BY.CP1131$)/, @Locale; | |
457 | } | |
458 | } | |
459 | ||
4599a1de JH |
460 | @Locale = sort @Locale; |
461 | ||
887ef7ed PP |
462 | debug "# Locales =\n"; |
463 | for ( @Locale ) { | |
464 | debug "# $_\n"; | |
465 | } | |
8ebc5c01 | 466 | |
284102e8 | 467 | my %Problem; |
2a680da6 JH |
468 | my %Okay; |
469 | my %Testing; | |
097ee67d | 470 | my @Neoalpha; |
d8093b23 | 471 | my %Neoalpha; |
284102e8 | 472 | |
2a680da6 JH |
473 | sub tryneoalpha { |
474 | my ($Locale, $i, $test) = @_; | |
2a680da6 JH |
475 | unless ($test) { |
476 | $Problem{$i}{$Locale} = 1; | |
477 | debug "# failed $i with locale '$Locale'\n"; | |
478 | } else { | |
479 | push @{$Okay{$i}}, $Locale; | |
480 | } | |
481 | } | |
482 | ||
284102e8 JH |
483 | foreach $Locale (@Locale) { |
484 | debug "# Locale = $Locale\n"; | |
485 | @Alnum_ = getalnum_(); | |
db4b7445 | 486 | debug "# w = ", join("",@Alnum_), "\n"; |
284102e8 JH |
487 | |
488 | unless (setlocale(LC_ALL, $Locale)) { | |
489 | foreach (99..103) { | |
490 | $Problem{$_}{$Locale} = -1; | |
8ebc5c01 | 491 | } |
284102e8 | 492 | next; |
8ebc5c01 | 493 | } |
8ebc5c01 | 494 | |
284102e8 JH |
495 | # Sieve the uppercase and the lowercase. |
496 | ||
097ee67d JH |
497 | my %UPPER = (); |
498 | my %lower = (); | |
499 | my %BoThCaSe = (); | |
284102e8 JH |
500 | for (@Alnum_) { |
501 | if (/[^\d_]/) { # skip digits and the _ | |
502 | if (uc($_) eq $_) { | |
503 | $UPPER{$_} = $_; | |
504 | } | |
505 | if (lc($_) eq $_) { | |
506 | $lower{$_} = $_; | |
507 | } | |
508 | } | |
509 | } | |
510 | foreach (keys %UPPER) { | |
097ee67d | 511 | $BoThCaSe{$_}++ if exists $lower{$_}; |
284102e8 JH |
512 | } |
513 | foreach (keys %lower) { | |
097ee67d | 514 | $BoThCaSe{$_}++ if exists $UPPER{$_}; |
284102e8 | 515 | } |
097ee67d | 516 | foreach (keys %BoThCaSe) { |
284102e8 JH |
517 | delete $UPPER{$_}; |
518 | delete $lower{$_}; | |
519 | } | |
520 | ||
db4b7445 A |
521 | debug "# UPPER = ", join("", sort keys %UPPER ), "\n"; |
522 | debug "# lower = ", join("", sort keys %lower ), "\n"; | |
523 | debug "# BoThCaSe = ", join("", sort keys %BoThCaSe), "\n"; | |
284102e8 JH |
524 | |
525 | # Find the alphabets that are not alphabets in the default locale. | |
8ebc5c01 | 526 | |
284102e8 JH |
527 | { |
528 | no locale; | |
8ebc5c01 | 529 | |
284102e8 JH |
530 | @Neoalpha = (); |
531 | for (keys %UPPER, keys %lower) { | |
532 | push(@Neoalpha, $_) if (/\W/); | |
d8093b23 | 533 | $Neoalpha{$_} = $_; |
284102e8 | 534 | } |
8ebc5c01 | 535 | } |
8ebc5c01 | 536 | |
284102e8 | 537 | @Neoalpha = sort @Neoalpha; |
8ebc5c01 | 538 | |
db4b7445 | 539 | debug "# Neoalpha = ", join("",@Neoalpha), "\n"; |
8ebc5c01 | 540 | |
284102e8 JH |
541 | if (@Neoalpha == 0) { |
542 | # If we have no Neoalphas the remaining tests are no-ops. | |
6be75cd7 | 543 | debug "# no Neoalpha, skipping tests 99..102 for locale '$Locale'\n"; |
a88c3d7c GS |
544 | foreach (99..102) { |
545 | push @{$Okay{$_}}, $Locale; | |
546 | } | |
6be75cd7 | 547 | } else { |
8ebc5c01 | 548 | |
6be75cd7 | 549 | # Test \w. |
284102e8 | 550 | |
ef4a39e5 | 551 | my $word = join('', @Neoalpha); |
8ebc5c01 | 552 | |
5fef21a3 JH |
553 | my $badutf8; |
554 | { | |
555 | local $SIG{__WARN__} = sub { | |
556 | $badutf8 = $_[0] =~ /Malformed UTF-8/; | |
557 | }; | |
558 | $Locale =~ /utf-?8/i; | |
559 | } | |
560 | ||
561 | if ($badutf8) { | |
562 | debug "# Locale name contains bad UTF-8, skipping test 99 for locale '$Locale'\n"; | |
563 | } elsif ($Locale =~ /utf-?8/i) { | |
ef4a39e5 JH |
564 | debug "# unknown whether locale and Unicode have the same \\w, skipping test 99 for locale '$Locale'\n"; |
565 | push @{$Okay{99}}, $Locale; | |
566 | } else { | |
567 | if ($word =~ /^(\w+)$/) { | |
568 | tryneoalpha($Locale, 99, 1); | |
569 | } else { | |
570 | tryneoalpha($Locale, 99, 0); | |
571 | } | |
8ac0d9e6 | 572 | } |
ef4a39e5 | 573 | |
2a680da6 | 574 | # Cross-check the whole 8-bit character set. |
8ebc5c01 | 575 | |
6be75cd7 | 576 | for (map { chr } 0..255) { |
2a680da6 JH |
577 | tryneoalpha($Locale, 100, |
578 | (/\w/ xor /\W/) || | |
579 | (/\d/ xor /\D/) || | |
580 | (/\s/ xor /\S/)); | |
284102e8 | 581 | } |
8ebc5c01 | 582 | |
6be75cd7 | 583 | # Test for read-only scalars' locale vs non-locale comparisons. |
284102e8 | 584 | |
284102e8 | 585 | { |
6be75cd7 JH |
586 | no locale; |
587 | $a = "qwerty"; | |
588 | { | |
589 | use locale; | |
2a680da6 | 590 | tryneoalpha($Locale, 101, ($a cmp "qwerty") == 0); |
8ebc5c01 | 591 | } |
592 | } | |
8ebc5c01 | 593 | |
6be75cd7 JH |
594 | { |
595 | my ($from, $to, $lesser, $greater, | |
596 | @test, %test, $test, $yes, $no, $sign); | |
597 | ||
598 | for (0..9) { | |
599 | # Select a slice. | |
600 | $from = int(($_*@Alnum_)/10); | |
601 | $to = $from + int(@Alnum_/10); | |
602 | $to = $#Alnum_ if ($to > $#Alnum_); | |
603 | $lesser = join('', @Alnum_[$from..$to]); | |
604 | # Select a slice one character on. | |
605 | $from++; $to++; | |
606 | $to = $#Alnum_ if ($to > $#Alnum_); | |
607 | $greater = join('', @Alnum_[$from..$to]); | |
608 | ($yes, $no, $sign) = ($lesser lt $greater | |
609 | ? (" ", "not ", 1) | |
610 | : ("not ", " ", -1)); | |
611 | # all these tests should FAIL (return 0). | |
612 | # Exact lt or gt cannot be tested because | |
613 | # in some locales, say, eacute and E may test equal. | |
614 | @test = | |
615 | ( | |
616 | $no.' ($lesser le $greater)', # 1 | |
617 | 'not ($lesser ne $greater)', # 2 | |
618 | ' ($lesser eq $greater)', # 3 | |
619 | $yes.' ($lesser ge $greater)', # 4 | |
620 | $yes.' ($lesser ge $greater)', # 5 | |
621 | $yes.' ($greater le $lesser )', # 7 | |
622 | 'not ($greater ne $lesser )', # 8 | |
623 | ' ($greater eq $lesser )', # 9 | |
624 | $no.' ($greater ge $lesser )', # 10 | |
0e053d1e | 625 | 'not (($lesser cmp $greater) == -($sign))' # 11 |
6be75cd7 JH |
626 | ); |
627 | @test{@test} = 0 x @test; | |
628 | $test = 0; | |
284102e8 | 629 | for my $ti (@test) { |
6be75cd7 JH |
630 | $test{$ti} = eval $ti; |
631 | $test ||= $test{$ti} | |
284102e8 | 632 | } |
2a680da6 | 633 | tryneoalpha($Locale, 102, $test == 0); |
6be75cd7 | 634 | if ($test) { |
6be75cd7 JH |
635 | debug "# lesser = '$lesser'\n"; |
636 | debug "# greater = '$greater'\n"; | |
637 | debug "# lesser cmp greater = ", | |
638 | $lesser cmp $greater, "\n"; | |
639 | debug "# greater cmp lesser = ", | |
640 | $greater cmp $lesser, "\n"; | |
641 | debug "# (greater) from = $from, to = $to\n"; | |
642 | for my $ti (@test) { | |
643 | debugf("# %-40s %-4s", $ti, | |
644 | $test{$ti} ? 'FAIL' : 'ok'); | |
645 | if ($ti =~ /\(\.*(\$.+ +cmp +\$[^\)]+)\.*\)/) { | |
646 | debugf("(%s == %4d)", $1, eval $1); | |
647 | } | |
648 | debug "\n#"; | |
649 | } | |
284102e8 | 650 | |
6be75cd7 JH |
651 | last; |
652 | } | |
284102e8 | 653 | } |
8ebc5c01 | 654 | } |
655 | } | |
6be75cd7 JH |
656 | |
657 | use locale; | |
658 | ||
659 | my ($x, $y) = (1.23, 1.23); | |
660 | ||
0e053d1e | 661 | $a = "$x"; |
6be75cd7 | 662 | printf ''; # printf used to reset locale to "C" |
0e053d1e | 663 | $b = "$y"; |
6be75cd7 | 664 | |
2a680da6 JH |
665 | debug "# 103..107: a = $a, b = $b, Locale = $Locale\n"; |
666 | ||
667 | tryneoalpha($Locale, 103, $a eq $b); | |
6be75cd7 JH |
668 | |
669 | my $c = "$x"; | |
670 | my $z = sprintf ''; # sprintf used to reset locale to "C" | |
671 | my $d = "$y"; | |
672 | ||
2a680da6 | 673 | debug "# 104..107: c = $c, d = $d, Locale = $Locale\n"; |
6be75cd7 | 674 | |
2a680da6 | 675 | tryneoalpha($Locale, 104, $c eq $d); |
6be75cd7 | 676 | |
2a680da6 | 677 | { |
9f1b1f2d | 678 | use warnings; |
2a680da6 | 679 | my $w = 0; |
0e053d1e JH |
680 | local $SIG{__WARN__} = |
681 | sub { | |
906f284f | 682 | print "# @_\n"; |
0e053d1e JH |
683 | $w++; |
684 | }; | |
6be75cd7 | 685 | |
0e053d1e JH |
686 | # The == (among other ops) used to warn for locales |
687 | # that had something else than "." as the radix character. | |
6be75cd7 | 688 | |
2a680da6 | 689 | tryneoalpha($Locale, 105, $c == 1.23); |
6be75cd7 | 690 | |
2a680da6 | 691 | tryneoalpha($Locale, 106, $c == $x); |
6be75cd7 | 692 | |
2a680da6 | 693 | tryneoalpha($Locale, 107, $c == $d); |
6be75cd7 | 694 | |
2a680da6 | 695 | { |
2de3dbcc | 696 | # no locale; # XXX did this ever work correctly? |
6be75cd7 | 697 | |
2a680da6 | 698 | my $e = "$x"; |
6be75cd7 | 699 | |
2a680da6 | 700 | debug "# 108..110: e = $e, Locale = $Locale\n"; |
6be75cd7 | 701 | |
2a680da6 | 702 | tryneoalpha($Locale, 108, $e == 1.23); |
6be75cd7 | 703 | |
2a680da6 JH |
704 | tryneoalpha($Locale, 109, $e == $x); |
705 | ||
706 | tryneoalpha($Locale, 110, $e == $c); | |
6be75cd7 | 707 | } |
2a680da6 | 708 | |
2a680da6 | 709 | my $f = "1.23"; |
906f284f | 710 | my $g = 2.34; |
2a680da6 | 711 | |
906f284f | 712 | debug "# 111..115: f = $f, g = $g, locale = $Locale\n"; |
2a680da6 | 713 | |
906f284f | 714 | tryneoalpha($Locale, 111, $f == 1.23); |
6be75cd7 | 715 | |
906f284f | 716 | tryneoalpha($Locale, 112, $f == $x); |
2a680da6 | 717 | |
906f284f NC |
718 | tryneoalpha($Locale, 113, $f == $c); |
719 | ||
720 | tryneoalpha($Locale, 114, abs(($f + $g) - 3.57) < 0.01); | |
721 | ||
722 | tryneoalpha($Locale, 115, $w == 0); | |
6be75cd7 JH |
723 | } |
724 | ||
26d80d95 LC |
725 | # Does taking lc separately differ from taking |
726 | # the lc "in-line"? (This was the bug 19990704.002, change #3568.) | |
727 | # The bug was in the caching of the 'o'-magic. | |
2a680da6 JH |
728 | { |
729 | use locale; | |
6be75cd7 | 730 | |
2a680da6 JH |
731 | sub lcA { |
732 | my $lc0 = lc $_[0]; | |
733 | my $lc1 = lc $_[1]; | |
734 | return $lc0 cmp $lc1; | |
735 | } | |
6be75cd7 | 736 | |
2a680da6 JH |
737 | sub lcB { |
738 | return lc($_[0]) cmp lc($_[1]); | |
739 | } | |
6be75cd7 | 740 | |
2a680da6 JH |
741 | my $x = "ab"; |
742 | my $y = "aa"; | |
743 | my $z = "AB"; | |
6be75cd7 | 744 | |
906f284f | 745 | tryneoalpha($Locale, 116, |
2a680da6 JH |
746 | lcA($x, $y) == 1 && lcB($x, $y) == 1 || |
747 | lcA($x, $z) == 0 && lcB($x, $z) == 0); | |
6be75cd7 | 748 | } |
d8093b23 | 749 | |
26d80d95 LC |
750 | # Does lc of an UPPER (if different from the UPPER) match |
751 | # case-insensitively the UPPER, and does the UPPER match | |
752 | # case-insensitively the lc of the UPPER. And vice versa. | |
3ba0e062 | 753 | { |
ef4a39e5 JH |
754 | use locale; |
755 | no utf8; | |
756 | my $re = qr/[\[\(\{\*\+\?\|\^\$\\]/; | |
757 | ||
758 | my @f = (); | |
759 | foreach my $x (keys %UPPER) { | |
760 | my $y = lc $x; | |
761 | next unless uc $y eq $x; | |
762 | print "# UPPER $x lc $y ", | |
763 | $x =~ /$y/i ? 1 : 0, " ", | |
764 | $y =~ /$x/i ? 1 : 0, "\n" if 0; | |
83d38f53 | 765 | # |
ef4a39e5 JH |
766 | # If $x and $y contain regular expression characters |
767 | # AND THEY lowercase (/i) to regular expression characters, | |
768 | # regcomp() will be mightily confused. No, the \Q doesn't | |
769 | # help here (maybe regex engine internal lowercasing | |
770 | # is done after the \Q?) An example of this happening is | |
771 | # the bg_BG (Bulgarian) locale under EBCDIC (OS/390 USS): | |
772 | # the chr(173) (the "[") is the lowercase of the chr(235). | |
83d38f53 | 773 | # |
ef4a39e5 JH |
774 | # Similarly losing EBCDIC locales include cs_cz, cs_CZ, |
775 | # el_gr, el_GR, en_us.IBM-037 (!), en_US.IBM-037 (!), | |
776 | # et_ee, et_EE, hr_hr, hr_HR, hu_hu, hu_HU, lt_LT, | |
777 | # mk_mk, mk_MK, nl_nl.IBM-037, nl_NL.IBM-037, | |
778 | # pl_pl, pl_PL, ro_ro, ro_RO, ru_ru, ru_RU, | |
779 | # sk_sk, sk_SK, sl_si, sl_SI, tr_tr, tr_TR. | |
83d38f53 JH |
780 | # |
781 | # Similar things can happen even under (bastardised) | |
782 | # non-EBCDIC locales: in many European countries before the | |
783 | # advent of ISO 8859-x nationally customised versions of | |
784 | # ISO 646 were devised, reusing certain punctuation | |
785 | # characters for modified characters needed by the | |
786 | # country/language. For example, the "|" might have | |
787 | # stood for U+00F6 or LATIN SMALL LETTER O WITH DIAERESIS. | |
788 | # | |
ef4a39e5 JH |
789 | if ($x =~ $re || $y =~ $re) { |
790 | print "# Regex characters in '$x' or '$y', skipping test 117 for locale '$Locale'\n"; | |
791 | next; | |
8ac0d9e6 | 792 | } |
ef4a39e5 JH |
793 | # With utf8 both will fail since the locale concept |
794 | # of upper/lower does not work well in Unicode. | |
795 | push @f, $x unless $x =~ /$y/i == $y =~ /$x/i; | |
796 | ||
8ac0d9e6 JH |
797 | foreach my $x (keys %lower) { |
798 | my $y = uc $x; | |
799 | next unless lc $y eq $x; | |
4c53e876 | 800 | print "# lower $x uc $y ", |
ef4a39e5 JH |
801 | $x =~ /$y/i ? 1 : 0, " ", |
802 | $y =~ /$x/i ? 1 : 0, "\n" if 0; | |
047e14fb JH |
803 | if ($x =~ $re || $y =~ $re) { # See above. |
804 | print "# Regex characters in '$x' or '$y', skipping test 117 for locale '$Locale'\n"; | |
805 | next; | |
806 | } | |
4c53e876 JH |
807 | # With utf8 both will fail since the locale concept |
808 | # of upper/lower does not work well in Unicode. | |
809 | push @f, $x unless $x =~ /$y/i == $y =~ /$x/i; | |
8ac0d9e6 | 810 | } |
906f284f | 811 | tryneoalpha($Locale, 117, @f == 0); |
8ac0d9e6 | 812 | if (@f) { |
906f284f | 813 | print "# failed 117 locale '$Locale' characters @f\n" |
8ac0d9e6 | 814 | } |
26d80d95 | 815 | } |
d8093b23 | 816 | } |
8ebc5c01 | 817 | } |
284102e8 | 818 | |
2a680da6 JH |
819 | # Recount the errors. |
820 | ||
906f284f | 821 | foreach (&last_without_setlocale()+1..$last) { |
2a680da6 | 822 | if ($Problem{$_} || !defined $Okay{$_} || !@{$Okay{$_}}) { |
097ee67d JH |
823 | if ($_ == 102) { |
824 | print "# The failure of test 102 is not necessarily fatal.\n"; | |
b4e009be | 825 | print "# It usually indicates a problem in the environment,\n"; |
284102e8 JH |
826 | print "# not in Perl itself.\n"; |
827 | } | |
828 | print "not "; | |
8ebc5c01 | 829 | } |
284102e8 | 830 | print "ok $_\n"; |
8ebc5c01 | 831 | } |
fb73857a | 832 | |
2a680da6 JH |
833 | # Give final advice. |
834 | ||
284102e8 JH |
835 | my $didwarn = 0; |
836 | ||
26d80d95 | 837 | foreach (99..$last) { |
284102e8 JH |
838 | if ($Problem{$_}) { |
839 | my @f = sort keys %{ $Problem{$_} }; | |
840 | my $f = join(" ", @f); | |
841 | $f =~ s/(.{50,60}) /$1\n#\t/g; | |
2a680da6 JH |
842 | |
843 | "#\n", | |
844 | "# The locale ", (@f == 1 ? "definition" : "definitions"), "\n#\n", | |
284102e8 JH |
845 | "#\t", $f, "\n#\n", |
846 | "# on your system may have errors because the locale test $_\n", | |
847 | "# failed in ", (@f == 1 ? "that locale" : "those locales"), | |
848 | ".\n"; | |
2a680da6 | 849 | print <<EOW; |
284102e8 JH |
850 | # |
851 | # If your users are not using these locales you are safe for the moment, | |
852 | # but please report this failure first to perlbug\@perl.com using the | |
853 | # perlbug script (as described in the INSTALL file) so that the exact | |
854 | # details of the failures can be sorted out first and then your operating | |
855 | # system supplier can be alerted about these anomalies. | |
856 | # | |
857 | EOW | |
858 | $didwarn = 1; | |
fb73857a | 859 | } |
860 | } | |
774d564b | 861 | |
26d80d95 | 862 | # Tell which locales were okay and which were not. |
2a680da6 | 863 | |
284102e8 | 864 | if ($didwarn) { |
26d80d95 | 865 | my (@s, @F); |
284102e8 JH |
866 | |
867 | foreach my $l (@Locale) { | |
868 | my $p = 0; | |
26d80d95 | 869 | foreach my $t (102..$last) { |
284102e8 | 870 | $p++ if $Problem{$t}{$l}; |
8ebc5c01 | 871 | } |
284102e8 | 872 | push @s, $l if $p == 0; |
26d80d95 | 873 | push @F, $l unless $p == 0; |
8ebc5c01 | 874 | } |
284102e8 | 875 | |
68d47915 CK |
876 | if (@s) { |
877 | my $s = join(" ", @s); | |
878 | $s =~ s/(.{50,60}) /$1\n#\t/g; | |
879 | ||
880 | warn | |
881 | "# The following locales\n#\n", | |
882 | "#\t", $s, "\n#\n", | |
883 | "# tested okay.\n#\n", | |
884 | } else { | |
26d80d95 LC |
885 | warn "# None of your locales were fully okay.\n"; |
886 | } | |
887 | ||
888 | if (@F) { | |
889 | my $F = join(" ", @F); | |
890 | $F =~ s/(.{50,60}) /$1\n#\t/g; | |
891 | ||
892 | warn | |
893 | "# The following locales\n#\n", | |
0e053d1e | 894 | "#\t", $F, "\n#\n", |
26d80d95 LC |
895 | "# had problems.\n#\n", |
896 | } else { | |
897 | warn "# None of your locales were broken.\n"; | |
68d47915 | 898 | } |
0e053d1e | 899 | |
c406fbf4 JH |
900 | if (@utf8locale) { |
901 | my $S = join(" ", @utf8locale); | |
902 | $S =~ s/(.{50,60}) /$1\n#\t/g; | |
903 | ||
904 | warn "#\n# The following locales\n#\n", | |
905 | "#\t", $S, "\n#\n", | |
906 | "# were skipped for the tests ", | |
907 | join(" ", sort {$a<=>$b} keys %utf8skip), "\n", | |
908 | "# because UTF-8 and locales do not work together in Perl.\n#\n"; | |
909 | } | |
8ebc5c01 | 910 | } |
90248788 | 911 | |
906f284f NC |
912 | sub last { 117 } |
913 | ||
90248788 | 914 | # eof |