xref: /openbsd-src/gnu/usr.bin/perl/t/uni/fold.t (revision 3d61058aa5c692477b6d18acfbbdb653a9930ff9) 
1use strict;
2use warnings;
3
4# re/fold_grind.t has more complex tests, but doesn't test every fold
5# This file also tests the fc() keyword.
6
7BEGIN {
8    chdir 't' if -d 't';
9    require './test.pl';
10    set_up_inc('../lib');
11    skip_all_without_unicode_tables();
12    skip_all_if_miniperl("miniperl, no Unicode::Normalize");
13    require Config; Config->import;
14    require './charset_tools.pl';
15    require './loc_tools.pl';   # Contains find_utf8_ctype_locale()
16}
17
18use feature 'unicode_strings';
19use Unicode::UCD qw(all_casefolds);
20
21binmode *STDOUT, ":utf8";
22
23our $TODO;
24
25
26plan("no_plan");
27# Read in the official case folding definitions.
28my $casefolds = all_casefolds();
29my @folds;
30my @CF;
31my @simple_folds;
32my %reverse_fold;
33use Unicode::UCD;
34use charnames();
35
36foreach my $decimal_code_point (sort { $a <=> $b } keys %$casefolds) {
37    # We only use simple folds in fc(), since the regex engine uses full case
38    # folding.
39
40    my $name = charnames::viacode($decimal_code_point);
41    my $type = $casefolds->{$decimal_code_point}{'status'};
42    my $code = $casefolds->{$decimal_code_point}{'code'};
43    my $simple = $casefolds->{$decimal_code_point}{'simple'};
44    my $full = $casefolds->{$decimal_code_point}{'full'};
45
46    if ($simple && $simple ne $full) { # If there is a distinction
47        push @simple_folds, [ $code, $simple, $type, $name ];
48    }
49
50    push @CF, [ $code, $full, $type, $name ];
51
52    # Get the inverse fold for single-char mappings.
53    $reverse_fold{pack "W*", hex $simple} = pack "W*", $decimal_code_point if $simple;
54}
55
56foreach my $test_ref ( @simple_folds ) {
57    use feature 'fc';
58    my ($code, $mapping, $type, $name) = @$test_ref;
59    my $c = pack("W*", hex $code);
60    utf8::upgrade($c);
61    my $f = pack("W*", map { hex } split " ", $mapping);
62
63    my $against = join "", "qq{", map("\\x{$_}", split " ", $mapping), "}";
64    {
65        isnt(fc($c), $f, "$code - $name - $mapping - $type - Full casefolding, fc(\\x{$code}) ne $against");
66        isnt("\F$c", $f, "$code - $name - $mapping - $type - Full casefolding, qq{\\F\\x{$code}} ne $against");
67    }
68}
69
70foreach my $test_ref (@CF) {
71    my ($code, $mapping, $type, $name) = @$test_ref;
72    my $c = pack("W*", hex $code);
73    utf8::upgrade($c);
74    my $f = pack("W*", map { hex } split " ", $mapping);
75    my $f_length = length $f;
76    foreach my $test (
77            qq[":$c:" =~ /:$c:/],
78            qq[":$c:" =~ /:$c:/i],
79            qq[":$c:" =~ /:[_$c]:/], # Place two chars in [] so doesn't get
80                                     # optimized to a non-charclass
81            qq[":$c:" =~ /:[_$c]:/i],
82            qq[":$c:" =~ /:$f:/i],
83            qq[":$f:" =~ /:$c:/i],
84    ) {
85        ok eval $test, "$code - $name - $mapping - $type - $test";
86    }
87
88    {
89        # fc() tests
90        my $against = join "", "qq{", map("\\x{$_}", split " ", $mapping), "}";
91        is(CORE::fc($c), $f,
92            "$code - $name - $mapping - $type - fc(\\x{$code}) eq $against");
93        is("\F$c", $f, "$code - $name - $mapping - $type - qq{\\F\\x{$code}} eq $against");
94
95        # And here we test bytes. For [A-Za-z0-9], the fold is the same as lc under
96        # bytes. For everything else, it's the bytes that formed the original string.
97        if ( $c =~ /[A-Za-z0-9]/ ) {
98            use bytes;
99            is(CORE::fc($c), lc($c), "$code - $name - fc and use bytes, ascii");
100        } else {
101            my $copy = "" . $c;
102            utf8::encode($copy);
103            is($copy, do { use bytes; CORE::fc($c) }, "$code - $name - fc and use bytes");
104        }
105    }
106    # Certain tests weren't convenient to put in the list above since they are
107    # TODO's in multi-character folds.
108    if ($f_length == 1) {
109
110        # The qq loses the utf8ness of ":$f:".  These tests are not about
111        # finding bugs in utf8ness, so make sure it's utf8.
112        my $test = qq[my \$s = ":$f:"; utf8::upgrade(\$s); \$s =~ /:[_$c]:/i];
113        ok eval $test, "$code - $name - $mapping - $type - $test";
114        $test = qq[":$c:" =~ /:[_$f]:/i];
115        ok eval $test, "$code - $name - $mapping - $type - $test";
116    }
117    else {
118
119        # There are two classes of multi-char folds that need more work.  For
120        # example,
121        #   ":ß:" =~ /:[_s]{2}:/i
122        #   ":ss:" =~ /:[_ß]:/i
123        #
124        # Some of the old tests for the second case happened to pass somewhat
125        # coincidentally.  But none would pass if changed to this.
126        #   ":SS:" =~ /:[_ß]:/i
127        #
128        # As the capital SS doesn't get folded.  When those pass, it means
129        # that the code has been changed to take into account folding in the
130        # string, and all should pass, capitalized or not (this wouldn't be
131        # true for [^complemented character classes], for which the fold case
132        # is better, but these aren't used in this .t currently.  So, what is
133        # done is to essentially upper-case the string for this class (but use
134        # the reverse fold not uc(), as that is more correct)
135        my $u;
136        for my $i (0 .. $f_length - 1) {
137            my $cur_char = substr($f, $i, 1);
138            $u .= $reverse_fold{$cur_char} || $cur_char;
139        }
140        my $test;
141
142        # A multi-char fold should not match just one char;
143        # e.g., ":ß:" !~ /:[_s]:/i
144        $test = qq[":$c:" !~ /:[_$f]:/i];
145        ok eval $test, "$code - $name - $mapping - $type - $test";
146
147        TODO: { # e.g., ":ß:" =~ /:[_s]{2}:/i
148            local $TODO = 'Multi-char fold in [character class]';
149
150            $test = qq[":$c:" =~ /:[_$f]{$f_length}:/i];
151            ok eval $test, "$code - $name - $mapping - $type - $test";
152        }
153
154        # e.g., ":SS:" =~ /:[_ß]:/i now pass, so TODO has been removed, but
155        # since they use '$u', they are left out of the main loop
156        $test = qq[ my \$s = ":$u:"; utf8::upgrade(\$s); \$s =~ /:[_$c]:/i];
157        ok eval $test, "$code - $name - $mapping - $type - $test";
158
159        my $bracketed_f = ($f =~ s/(.)/[$1]/gr);
160        $test = qq[":$c:" =~ /:$bracketed_f:/iu];
161        ok eval $test, "$code - $name - $mapping - $type - $test";
162
163        my @f_chars = ($f =~ / (.) (.) (.?) /x);
164        my $every_other_bracketed_f = "[$f_chars[0]]$f_chars[1]";
165        $every_other_bracketed_f .= "[$f_chars[2]]" if $f_chars[2];
166        $test = qq[":$c:" =~ /:$every_other_bracketed_f:/iu];
167        ok eval $test, "$code - $name - $mapping - $type - $test";
168
169        my $other_every_bracketed_f = "$f_chars[0]";
170        $other_every_bracketed_f .= "[$f_chars[1]]";
171        $other_every_bracketed_f .= "$f_chars[2]" if $f_chars[2];
172        $test = qq[":$c:" =~ /:$other_every_bracketed_f:/iu];
173        ok eval $test, "$code - $name - $mapping - $type - $test";
174    }
175}
176
177{
178    use utf8;
179    use feature qw(fc);
180    # These three come from the ICU project's test suite, more especifically
181    # http://icu.sourcearchive.com/documentation/4.4~rc1-1/strcase_8cpp-source.html
182
183    my $s = "A\N{U+00df}\N{U+00b5}\N{U+fb03}\N{U+1040C}\N{U+0130}\N{U+0131}";
184    #\N{LATIN CAPITAL LETTER A}\N{LATIN SMALL LETTER SHARP S}\N{MICRO SIGN}\N{LATIN SMALL LIGATURE FFI}\N{DESERET CAPITAL LETTER AY}\N{LATIN CAPITAL LETTER I WITH DOT ABOVE}\N{LATIN SMALL LETTER DOTLESS I}
185
186    my $f = "ass\N{U+03bc}ffi\N{U+10434}i\N{U+0307}\N{U+0131}";
187    #\N{LATIN SMALL LETTER A}\N{LATIN SMALL LETTER S}\N{LATIN SMALL LETTER S}\N{GREEK SMALL LETTER MU}\N{LATIN SMALL LETTER F}\N{LATIN SMALL LETTER F}\N{LATIN SMALL LETTER I}\N{DESERET SMALL LETTER AY}\N{LATIN SMALL LETTER I}\N{COMBINING DOT ABOVE}\N{LATIN SMALL LETTER DOTLESS I}
188
189    is(fc($s), $f, "ICU's casefold test passes");
190    is("\F$s", $f, "ICU's casefold test passes");
191
192    is( fc("aBİIıϐßffi��"), "abi̇iıβssffi��" );
193    is( "\FaBİIıϐßffi��", "abi̇iıβssffi��" );
194#    TODO: {
195#        local $::TODO = "turkic special cases";
196#        is( fc "aBİIıϐßffi��", "abiııβssffi��" );
197#    }
198
199    # The next batch come from http://www.devdaily.com/java/jwarehouse/lucene/contrib/icu/src/test/org/apache/lucene/analysis/icu/TestICUFoldingFilter.java.shtml
200    # Except the article got most casings wrong. Or maybe Lucene does.
201
202    is( fc("This is a test"), "this is a test" );
203    is( fc("Ruß"), "russ"    );
204    is( fc("ΜΆΪΟΣ"), "μάϊοσ" );
205    is( fc("Μάϊος"), "μάϊοσ" );
206    is( fc("��"), "��"       );
207    is( fc("r" . uni_to_native("\xe9") . "sum" . uni_to_native("\xe9")),
208           "r" . uni_to_native("\xe9") . "sum" . uni_to_native("\xe9") );
209    is( fc("re\x{0301}sume\x{0301}"), "re\x{301}sume\x{301}" );
210    is( fc("ELİF"), "eli\x{307}f" );
211    is( fc("eli\x{307}f"), "eli\x{307}f");
212
213    # This batch comes from
214    # http://www.java2s.com/Open-Source/Java-Document/Internationalization-Localization/icu4j/com/ibm/icu/dev/test/lang/UCharacterCaseTest.java.htm
215    # Which uses ICU as the backend.
216
217    my @folding_mixed = (
218        uni_to_native("\x{61}\x{42}\x{130}\x{49}\x{131}\x{3d0}\x{df}\x{fb03}"),
219        "A" . uni_to_native("\x{df}\x{b5}\x{fb03}\x{1040C}\x{130}\x{131}"),
220    );
221
222    my @folding_default = (
223        uni_to_native("\x{61}\x{62}\x{69}\x{307}\x{69}\x{131}\x{3b2}\x{73}\x{73}\x{66}\x{66}\x{69}"),
224        "ass\x{3bc}ffi\x{10434}i\x{307}\x{131}"
225    );
226
227    my @folding_exclude_turkic = (
228        uni_to_native("\x{61}\x{62}\x{69}\x{131}\x{131}\x{3b2}\x{73}\x{73}\x{66}\x{66}\x{69}"),
229                         "ass\x{3bc}ffi\x{10434}i\x{131}",
230    );
231
232    is( fc($folding_mixed[1]), $folding_default[1] );
233
234    is( fc($folding_mixed[0]), $folding_default[0] );
235
236}
237
238{
239    use utf8;
240    # Table stolen from tchrist's mail in
241    # http://bugs.python.org/file23051/casing-tests.py
242    # and http://98.245.80.27/tcpc/OSCON2011/case-test.python3
243    # For reference, it's a longer version of what he posted here:
244    # http://stackoverflow.com/questions/6991038/case-insensitive-storage-and-unicode-compatibility
245
246    #Couple of repeats because I'm lazy, not tchrist's fault.
247
248    #This should probably go in t/op/lc.t
249
250    my @test_table = (
251# ORIG LC_SIMPLE TC_SIMPLE UC_SIMPLE LC_FULL TC_FULL UC_FULL FC_SIMPLE FC_TURKIC FC_FULL
252[ 'þǽr rihtes', 'þǽr rihtes', 'Þǽr Rihtes', 'ÞǼR RIHTES', 'þǽr rihtes', 'Þǽr Rihtes', 'ÞǼR RIHTES', 'þǽr rihtes', 'þǽr rihtes', 'þǽr rihtes',  ],
253[ 'duȝeðlice', 'duȝeðlice', 'Duȝeðlice', 'DUȜEÐLICE', 'duȝeðlice', 'Duȝeðlice', 'DUȜEÐLICE', 'duȝeðlice', 'duȝeðlice', 'duȝeðlice',  ],
254[ 'Ævar Arnfjörð Bjarmason', 'ævar arnfjörð bjarmason', 'Ævar Arnfjörð Bjarmason', 'ÆVAR ARNFJÖRÐ BJARMASON', 'ævar arnfjörð bjarmason', 'Ævar Arnfjörð Bjarmason', 'ÆVAR ARNFJÖRÐ BJARMASON', 'ævar arnfjörð bjarmason', 'ævar arnfjörð bjarmason', 'ævar arnfjörð bjarmason',  ],
255[ 'Кириллица', 'кириллица', 'Кириллица', 'КИРИЛЛИЦА', 'кириллица', 'Кириллица', 'КИРИЛЛИЦА', 'кириллица', 'кириллица', 'кириллица',  ],
256[ 'ij', 'ij', 'IJ', 'IJ', 'ij', 'IJ', 'IJ', 'ij', 'ij', 'ij',  ],
257[ 'Van Dijke', 'van dijke', 'Van Dijke', 'VAN DIJKE', 'van dijke', 'Van Dijke', 'VAN DIJKE', 'van dijke', 'van dijke', 'van dijke',  ],
258[ 'VAN DIJKE', 'van dijke', 'Van Dijke', 'VAN DIJKE', 'van dijke', 'Van Dijke', 'VAN DIJKE', 'van dijke', 'van dijke', 'van dijke',  ],
259[ 'efficient', 'efficient', 'Efficient', 'EffiCIENT', 'efficient', 'Efficient', 'EFFICIENT', 'efficient', 'efficient', 'efficient',  ],
260[ 'flour', 'flour', 'flour', 'flOUR', 'flour', 'Flour', 'FLOUR', 'flour', 'flour', 'flour',  ],
261[ 'flour and water', 'flour and water', 'flour And Water', 'flOUR AND WATER', 'flour and water', 'Flour And Water', 'FLOUR AND WATER', 'flour and water', 'flour and water', 'flour and water',  ],
262[ 'dzur', 'dzur', 'Dzur', 'DZUR', 'dzur', 'Dzur', 'DZUR', 'dzur', 'dzur', 'dzur',  ],
263[ 'Dzur', 'dzur', 'Dzur', 'DZUR', 'dzur', 'Dzur', 'DZUR', 'dzur', 'dzur', 'dzur',  ],
264[ 'DZUR', 'dzur', 'Dzur', 'DZUR', 'dzur', 'Dzur', 'DZUR', 'dzur', 'dzur', 'dzur',  ],
265[ 'dzur mountain', 'dzur mountain', 'Dzur Mountain', 'DZUR MOUNTAIN', 'dzur mountain', 'Dzur Mountain', 'DZUR MOUNTAIN', 'dzur mountain', 'dzur mountain', 'dzur mountain',  ],
266[ 'Dzur Mountain', 'dzur mountain', 'Dzur Mountain', 'DZUR MOUNTAIN', 'dzur mountain', 'Dzur Mountain', 'DZUR MOUNTAIN', 'dzur mountain', 'dzur mountain', 'dzur mountain',  ],
267[ 'DZUR MOUNTAIN', 'dzur mountain', 'Dzur Mountain', 'DZUR MOUNTAIN', 'dzur mountain', 'Dzur Mountain', 'DZUR MOUNTAIN', 'dzur mountain', 'dzur mountaın', 'dzur mountain',  ],
268[ 'poſt', 'poſt', 'Poſt', 'POST', 'poſt', 'Poſt', 'POST', 'post', 'post', 'post',  ],
269[ 'poſt', 'poſt', 'Poſt', 'POſt', 'poſt', 'Poſt', 'POST', 'poſt', 'post', 'post',  ],
270[ 'ſtop', 'ſtop', 'ſtop', 'ſtOP', 'ſtop', 'Stop', 'STOP', 'ſtop', 'stop', 'stop',  ],
271[ 'tschüß', 'tschüß', 'Tschüß', 'TSCHÜß', 'tschüß', 'Tschüß', 'TSCHÜSS', 'tschüß', 'tschüss', 'tschüss',  ],
272[ 'TSCHÜẞ', 'tschüß', 'Tschüß', 'TSCHÜẞ', 'tschüß', 'Tschüß', 'TSCHÜẞ', 'tschüß', 'tschüss', 'tschüss',  ],
273[ 'weiß', 'weiß', 'Weiß', 'WEIß', 'weiß', 'Weiß', 'WEISS', 'weiß', 'weiss', 'weiss',  ],
274[ 'WEIẞ', 'weiß', 'Weiß', 'WEIẞ', 'weiß', 'Weiß', 'WEIẞ', 'weiß', 'weıss', 'weiss',  ],
275[ 'ẞIEW', 'ßiew', 'ẞiew', 'ẞIEW', 'ßiew', 'ẞiew', 'ẞIEW', 'ßiew', 'ssıew', 'ssiew',  ],
276[ 'ᾲ', 'ᾲ', 'Ὰͅ', 'ᾺΙ', 'ᾲ', 'Ὰͅ', 'ᾺΙ', 'ὰι', 'ὰι', 'ὰι',  ],
277[ 'Ὰι', 'ὰι', 'Ὰι', 'ᾺΙ', 'ὰι', 'Ὰι', 'ᾺΙ', 'ὰι', 'ὰι', 'ὰι',  ],
278[ 'ᾺΙ', 'ὰι', 'Ὰι', 'ᾺΙ', 'ὰι', 'Ὰι', 'ᾺΙ', 'ὰι', 'ὰι', 'ὰι',  ],
279[ 'ᾲ', 'ᾲ', 'ᾲ', 'ᾲ', 'ᾲ', 'Ὰͅ', 'ᾺΙ', 'ᾲ', 'ὰι', 'ὰι',  ],
280[ 'Ὰͅ', 'ᾲ', 'Ὰͅ', 'ᾺΙ', 'ᾲ', 'Ὰͅ', 'ᾺΙ', 'ὰι', 'ὰι', 'ὰι',  ],
281[ 'ᾺΙ', 'ὰι', 'Ὰι', 'ᾺΙ', 'ὰι', 'Ὰι', 'ᾺΙ', 'ὰι', 'ὰι', 'ὰι',  ],
282[ 'ᾲ στο διάολο', 'ᾲ στο διάολο', 'ᾲ Στο Διάολο', 'ᾲ ΣΤΟ ΔΙΆΟΛΟ', 'ᾲ στο διάολο', 'Ὰͅ Στο Διάολο', 'ᾺΙ ΣΤΟ ΔΙΆΟΛΟ', 'ᾲ στο διάολο', 'ὰι στο διάολο', 'ὰι στο διάολο',  ],
283[ 'ᾲ στο διάολο', 'ᾲ στο διάολο', 'Ὰͅ Στο Διάολο', 'ᾺΙ ΣΤΟ ΔΙΆΟΛΟ', 'ᾲ στο διάολο', 'Ὰͅ Στο Διάολο', 'ᾺΙ ΣΤΟ ΔΙΆΟΛΟ', 'ὰι στο διάολο', 'ὰι στο διάολο', 'ὰι στο διάολο',  ],
284[ '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������',  ],
285[ '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������',  ],
286[ '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������', '��������������',  ],
287[ 'henry ⅷ', 'henry ⅷ', 'Henry Ⅷ', 'HENRY Ⅷ', 'henry ⅷ', 'Henry Ⅷ', 'HENRY Ⅷ', 'henry ⅷ', 'henry ⅷ', 'henry ⅷ',  ],
288[ 'Henry Ⅷ', 'henry ⅷ', 'Henry Ⅷ', 'HENRY Ⅷ', 'henry ⅷ', 'Henry Ⅷ', 'HENRY Ⅷ', 'henry ⅷ', 'henry ⅷ', 'henry ⅷ',  ],
289[ 'HENRY Ⅷ', 'henry ⅷ', 'Henry Ⅷ', 'HENRY Ⅷ', 'henry ⅷ', 'Henry Ⅷ', 'HENRY Ⅷ', 'henry ⅷ', 'henry ⅷ', 'henry ⅷ',  ],
290[ 'i work at ⓚ', 'i work at ⓚ', 'I Work At Ⓚ', 'I WORK AT Ⓚ', 'i work at ⓚ', 'I Work At Ⓚ', 'I WORK AT Ⓚ', 'i work at ⓚ', 'i work at ⓚ', 'i work at ⓚ',  ],
291[ 'I Work At Ⓚ', 'i work at ⓚ', 'I Work At Ⓚ', 'I WORK AT Ⓚ', 'i work at ⓚ', 'I Work At Ⓚ', 'I WORK AT Ⓚ', 'i work at ⓚ', 'ı work at ⓚ', 'i work at ⓚ',  ],
292[ 'I WORK AT Ⓚ', 'i work at ⓚ', 'I Work At Ⓚ', 'I WORK AT Ⓚ', 'i work at ⓚ', 'I Work At Ⓚ', 'I WORK AT Ⓚ', 'i work at ⓚ', 'ı work at ⓚ', 'i work at ⓚ',  ],
293[ 'istambul', 'istambul', 'Istambul', 'ISTAMBUL', 'istambul', 'Istambul', 'ISTAMBUL', 'istambul', 'istambul', 'istambul',  ],
294[ 'i̇stanbul', 'i̇stanbul', 'İstanbul', 'İSTANBUL', 'i̇stanbul', 'İstanbul', 'İSTANBUL', 'i̇stanbul', 'i̇stanbul', 'i̇stanbul',  ],
295[ 'İstanbul', 'i̇stanbul', 'İstanbul', 'İSTANBUL', 'i̇stanbul', 'İstanbul', 'İSTANBUL', 'i̇stanbul', 'ı̇stanbul', 'i̇stanbul',  ],
296[ 'İSTANBUL', 'istanbul', 'İstanbul', 'İSTANBUL', 'i̇stanbul', 'İstanbul', 'İSTANBUL', 'İstanbul', 'istanbul', 'i̇stanbul',  ],
297[ 'στιγμας', 'στιγμας', 'Στιγμας', 'ΣΤΙΓΜΑΣ', 'στιγμας', 'Στιγμας', 'ΣΤΙΓΜΑΣ', 'στιγμασ', 'στιγμασ', 'στιγμασ',  ],
298[ 'στιγμασ', 'στιγμασ', 'Στιγμασ', 'ΣΤΙΓΜΑΣ', 'στιγμασ', 'Στιγμασ', 'ΣΤΙΓΜΑΣ', 'στιγμασ', 'στιγμασ', 'στιγμασ',  ],
299[ 'ΣΤΙΓΜΑΣ', 'στιγμασ', 'Στιγμασ', 'ΣΤΙΓΜΑΣ', 'στιγμασ', 'Στιγμασ', 'ΣΤΙΓΜΑΣ', 'στιγμασ', 'στιγμασ', 'στιγμασ',  ],
300[ 'ʀᴀʀᴇ', 'ʀᴀʀᴇ', 'Ʀᴀʀᴇ', 'ƦᴀƦᴇ', 'ʀᴀʀᴇ', 'Ʀᴀʀᴇ', 'ƦᴀƦᴇ', 'ʀᴀʀᴇ', 'ʀᴀʀᴇ', 'ʀᴀʀᴇ',  ],
301[ 'Ʀᴀʀᴇ', 'ʀᴀʀᴇ', 'Ʀᴀʀᴇ', 'ƦᴀƦᴇ', 'ʀᴀʀᴇ', 'Ʀᴀʀᴇ', 'ƦᴀƦᴇ', 'ʀᴀʀᴇ', 'ʀᴀʀᴇ', 'ʀᴀʀᴇ',  ],
302[ 'ƦᴀƦᴇ', 'ʀᴀʀᴇ', 'Ʀᴀʀᴇ', 'ƦᴀƦᴇ', 'ʀᴀʀᴇ', 'Ʀᴀʀᴇ', 'ƦᴀƦᴇ', 'ʀᴀʀᴇ', 'ʀᴀʀᴇ', 'ʀᴀʀᴇ',  ],
303[ 'Ԧԧ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'ԧԧ', 'ԧԧ',  ],
304[ 'ԧԧ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'ԧԧ', 'ԧԧ',  ],
305[ 'Ԧԧ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'ԧԧ', 'ԧԧ',  ],
306[ 'ԦԦ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'Ԧԧ', 'ԦԦ', 'ԧԧ', 'ԧԧ', 'ԧԧ',  ],
307[ "þǽr rihtes", "þǽr rihtes", "Þǽr Rihtes", "ÞǼR RIHTES", "þǽr rihtes", "Þǽr Rihtes", "ÞǼR RIHTES", "þǽr rihtes", "þǽr rihtes", "þǽr rihtes",  ],
308[ "duȝeðlice", "duȝeðlice", "Duȝeðlice", "DUȜEÐLICE", "duȝeðlice", "Duȝeðlice", "DUȜEÐLICE", "duȝeðlice", "duȝeðlice", "duȝeðlice",  ],
309[ "Van Dijke", "van dijke", "Van Dijke", "VAN DIJKE", "van dijke", "Van Dijke", "VAN DIJKE", "van dijke", "van dijke", "van dijke",  ],
310[ "fi", "fi", "fi", "fi", "fi", "Fi", "FI", "fi", "fi", "fi",  ],
311[ "filesystem", "filesystem", "filesystem", "fiLESYSTEM", "filesystem", "Filesystem", "FILESYSTEM", "filesystem", "filesystem", "filesystem",  ],
312[ "efficient", "efficient", "Efficient", "EffiCIENT", "efficient", "Efficient", "EFFICIENT", "efficient", "efficient", "efficient",  ],
313[ "flour and water", "flour and water", "flour And Water", "flOUR AND WATER", "flour and water", "Flour And Water", "FLOUR AND WATER", "flour and water", "flour and water", "flour and water",  ],
314[ "dz", "dz", "Dz", "DZ", "dz", "Dz", "DZ", "dz", "dz", "dz",  ],
315[ "dzur mountain", "dzur mountain", "Dzur Mountain", "DZUR MOUNTAIN", "dzur mountain", "Dzur Mountain", "DZUR MOUNTAIN", "dzur mountain", "dzur mountain", "dzur mountain",  ],
316[ "poſt", "poſt", "Poſt", "POST", "poſt", "Poſt", "POST", "post", "post", "post",  ],
317[ "poſt", "poſt", "Poſt", "POſt", "poſt", "Poſt", "POST", "poſt", "post", "post",  ],
318[ "ſtop", "ſtop", "ſtop", "ſtOP", "ſtop", "Stop", "STOP", "ſtop", "stop", "stop",  ],
319[ "tschüß", "tschüß", "Tschüß", "TSCHÜß", "tschüß", "Tschüß", "TSCHÜSS", "tschüß", "tschüss", "tschüss",  ],
320[ "TSCHÜẞ", "tschüß", "Tschüß", "TSCHÜẞ", "tschüß", "Tschüß", "TSCHÜẞ", "tschüß", "tschüss", "tschüss",  ],
321[ "rußland", "rußland", "Rußland", "RUßLAND", "rußland", "Rußland", "RUSSLAND", "rußland", "russland", "russland",  ],
322[ "RUẞLAND", "rußland", "Rußland", "RUẞLAND", "rußland", "Rußland", "RUẞLAND", "rußland", "russland", "russland",  ],
323[ "weiß", "weiß", "Weiß", "WEIß", "weiß", "Weiß", "WEISS", "weiß", "weiss", "weiss",  ],
324[ "WEIẞ", "weiß", "Weiß", "WEIẞ", "weiß", "Weiß", "WEIẞ", "weiß", "weıss", "weiss",  ],
325[ "ẞIEW", "ßiew", "ẞiew", "ẞIEW", "ßiew", "ẞiew", "ẞIEW", "ßiew", "ssıew", "ssiew",  ],
326[ "ͅ", "ͅ", "Ι", "Ι", "ͅ", "Ι", "Ι", "ι", "ι", "ι",  ],
327[ "ᾲ", "ᾲ", "Ὰͅ", "ᾺΙ", "ᾲ", "Ὰͅ", "ᾺΙ", "ὰι", "ὰι", "ὰι",  ],
328[ "Ὰι", "ὰι", "Ὰι", "ᾺΙ", "ὰι", "Ὰι", "ᾺΙ", "ὰι", "ὰι", "ὰι",  ],
329[ "ᾺΙ", "ὰι", "Ὰι", "ᾺΙ", "ὰι", "Ὰι", "ᾺΙ", "ὰι", "ὰι", "ὰι",  ],
330[ "ᾲ", "ᾲ", "ᾲ", "ᾲ", "ᾲ", "Ὰͅ", "ᾺΙ", "ᾲ", "ὰι", "ὰι",  ],
331[ "Ὰͅ", "ᾲ", "Ὰͅ", "ᾺΙ", "ᾲ", "Ὰͅ", "ᾺΙ", "ὰι", "ὰι", "ὰι",  ],
332[ "ᾺΙ", "ὰι", "Ὰι", "ᾺΙ", "ὰι", "Ὰι", "ᾺΙ", "ὰι", "ὰι", "ὰι",  ],
333[ "ᾲ στο διάολο", "ᾲ στο διάολο", "ᾲ Στο Διάολο", "ᾲ ΣΤΟ ΔΙΆΟΛΟ", "ᾲ στο διάολο", "Ὰͅ Στο Διάολο", "ᾺΙ ΣΤΟ ΔΙΆΟΛΟ", "ᾲ στο διάολο", "ὰι στο διάολο", "ὰι στο διάολο",  ],
334[ "ᾲ στο διάολο", "ᾲ στο διάολο", "Ὰͅ Στο Διάολο", "ᾺΙ ΣΤΟ ΔΙΆΟΛΟ", "ᾲ στο διάολο", "Ὰͅ Στο Διάολο", "ᾺΙ ΣΤΟ ΔΙΆΟΛΟ", "ὰι στο διάολο", "ὰι στο διάολο", "ὰι στο διάολο",  ],
335[ "ⅷ", "ⅷ", "Ⅷ", "Ⅷ", "ⅷ", "Ⅷ", "Ⅷ", "ⅷ", "ⅷ", "ⅷ",  ],
336[ "henry ⅷ", "henry ⅷ", "Henry Ⅷ", "HENRY Ⅷ", "henry ⅷ", "Henry Ⅷ", "HENRY Ⅷ", "henry ⅷ", "henry ⅷ", "henry ⅷ",  ],
337[ "ⓚ", "ⓚ", "Ⓚ", "Ⓚ", "ⓚ", "Ⓚ", "Ⓚ", "ⓚ", "ⓚ", "ⓚ",  ],
338[ "i work at ⓚ", "i work at ⓚ", "I Work At Ⓚ", "I WORK AT Ⓚ", "i work at ⓚ", "I Work At Ⓚ", "I WORK AT Ⓚ", "i work at ⓚ", "i work at ⓚ", "i work at ⓚ",  ],
339[ "istambul", "istambul", "Istambul", "ISTAMBUL", "istambul", "Istambul", "ISTAMBUL", "istambul", "istambul", "istambul",  ],
340[ "i̇stanbul", "i̇stanbul", "İstanbul", "İSTANBUL", "i̇stanbul", "İstanbul", "İSTANBUL", "i̇stanbul", "i̇stanbul", "i̇stanbul",  ],
341[ "İstanbul", "i̇stanbul", "İstanbul", "İSTANBUL", "i̇stanbul", "İstanbul", "İSTANBUL", "i̇stanbul", "ı̇stanbul", "i̇stanbul",  ],
342[ "İSTANBUL", "istanbul", "İstanbul", "İSTANBUL", "i̇stanbul", "İstanbul", "İSTANBUL", "İstanbul", "istanbul", "i̇stanbul",  ],
343[ "στιγμας", "στιγμας", "Στιγμας", "ΣΤΙΓΜΑΣ", "στιγμας", "Στιγμας", "ΣΤΙΓΜΑΣ", "στιγμασ", "στιγμασ", "στιγμασ",  ],
344[ "στιγμασ", "στιγμασ", "Στιγμασ", "ΣΤΙΓΜΑΣ", "στιγμασ", "Στιγμασ", "ΣΤΙΓΜΑΣ", "στιγμασ", "στιγμασ", "στιγμασ",  ],
345[ "ΣΤΙΓΜΑΣ", "στιγμασ", "Στιγμασ", "ΣΤΙΓΜΑΣ", "στιγμασ", "Στιγμασ", "ΣΤΙΓΜΑΣ", "στιγμασ", "στιγμασ", "στιγμασ",  ],
346[ "ʀᴀʀᴇ", "ʀᴀʀᴇ", "Ʀᴀʀᴇ", "ƦᴀƦᴇ", "ʀᴀʀᴇ", "Ʀᴀʀᴇ", "ƦᴀƦᴇ", "ʀᴀʀᴇ", "ʀᴀʀᴇ", "ʀᴀʀᴇ",  ],
347[ "��������������", "��������������", "��������������", "��������������", "��������������", "��������������", "��������������", "��������������", "��������������", "��������������",  ],
348[ "Ԧԧ", "ԧԧ", "Ԧԧ", "ԦԦ", "ԧԧ", "Ԧԧ", "ԦԦ", "ԧԧ", "ԧԧ", "ԧԧ",  ],
349[ "ﬓﬔﬕﬖﬗ", "ﬓﬔﬕﬖﬗ", "ﬓﬔﬕﬖﬗ", "ﬓﬔﬕﬖﬗ", "ﬓﬔﬕﬖﬗ", "Մնﬔﬕﬖﬗ", "ՄՆՄԵՄԻՎՆՄԽ", "ﬓﬔﬕﬖﬗ", "մնմեմիվնմխ", "մնմեմիվնմխ",  ],
350[ "ʼn groot", "ʼn groot", "ʼn Groot", "ʼn GROOT", "ʼn groot", "ʼN Groot", "ʼN GROOT", "ʼn groot", "ʼn groot", "ʼn groot",  ],
351[ "ẚ", "ẚ", "ẚ", "ẚ", "ẚ", "Aʾ", "Aʾ", "ẚ", "aʾ", "aʾ",  ],
352[ "ff", "ff", "ff", "ff", "ff", "Ff", "FF", "ff", "ff", "ff",  ],
353[ "ǰ", "ǰ", "ǰ", "ǰ", "ǰ", "J̌", "J̌", "ǰ", "ǰ", "ǰ",  ],
354[ "550 nm or Å", "550 nm or å", "550 Nm Or Å", "550 NM OR Å", "550 nm or å", "550 Nm Or Å", "550 NM OR Å", "550 nm or å", "550 nm or å", "550 nm or å",  ],
355);
356
357    use feature qw(fc);
358
359    for (@test_table) {
360        my ($simple_lc, $simple_tc, $simple_uc, $simple_fc) = @{$_}[1, 2, 3, 7];
361        my ($orig, $lower, $titlecase, $upper, $fc_turkic, $fc_full) = @{$_}[0,4,5,6,8,9];
362
363        if ($orig =~ /(\P{Assigned})/) {   # So can fail gracefully in earlier
364                                           # Unicode versions
365            fail(sprintf "because U+%04X is unassigned", ord($1));
366            next;
367        }
368        is( fc($orig), $fc_full, "fc('$orig') returns '$fc_full'" );
369        is( "\F$orig", $fc_full, '\F works' );
370        is( lc($orig), $lower,   "lc('$orig') returns '$lower'" );
371        is( "\L$orig", $lower,   '\L works' );
372        is( uc($orig), $upper,   "uc('$orig') returns '$upper'" );
373        is( "\U$orig", $upper,   '\U works' );
374    }
375}
376
377{
378    use feature qw(fc);
379    package Eeyup  { use overload q{""} => sub { main::uni_to_native("\x{df}")   }, fallback => 1 }
380    package Uunope { use overload q{""} => sub { "\x{30cb}" }, fallback => 1 }
381    package Undef  { use overload q{""} => sub {   undef    }, fallback => 1 }
382
383    my $obj = bless {}, "Eeyup";
384    is(fc($obj), "ss", "fc() works on overloaded objects returning latin-1");
385    $obj = bless {}, "Eeyup";
386    is("\F$obj", "ss", '\F works on overloaded objects returning latin-1');
387
388    $obj = bless {}, "Uunope";
389    is(fc($obj), "\x{30cb}", "fc() works on overloaded objects returning UTF-8");
390    $obj = bless {}, "Uunope";
391    is("\F$obj", "\x{30cb}", '\F works on overloaded objects returning UTF-8');
392
393    $obj = bless {}, "Undef";
394    my $warnings;
395    {
396        no warnings;
397        use warnings "uninitialized";
398        local $SIG{__WARN__} = sub { $warnings++; like(shift, qr/Use of uninitialized value (?:\$obj )?in fc/) };
399        fc(undef);
400        fc($obj);
401    }
402    is( $warnings, 2, "correct number of warnings" );
403
404    my $fetched = 0;
405    package Derpy { sub TIESCALAR { bless {}, shift } sub FETCH { $fetched++; main::uni_to_native("\x{df}") } }
406
407    tie my $x, "Derpy";
408
409    is( fc($x), "ss", "fc() works on tied values" );
410    is( $fetched, 1, "and only calls the magic once" );
411
412}
413
414{
415    use feature qw( fc );
416    my $troublesome1 = uni_to_native("\xdf") x 11; #SvLEN should be 12, SvCUR should be 11
417                                    #So this should force fc() to grow the string.
418
419    is( fc($troublesome1), "ss" x 11, "fc() grows the string" );
420
421    my $troublesome2 = "abcdef:" . uni_to_native("\x{df}")
422                     . ":fjksjs"; #SvLEN should be 16, SvCUR should be 15
423    is( fc($troublesome2), "abcdef:ss:fjksjs", "fc() expands \\x{DF} in the middle of a string that needs to grow" );
424
425    my $troublesome3 = ":" . uni_to_native("\x{df}") . ":";
426    is( fc($troublesome3), ":ss:", "fc() expands \\x{DF} in the middle of a string" );
427
428
429    my $troublesome4 = uni_to_native("\x{B5}"); #\N{MICRON SIGN} is latin-1, but its foldcase is in UTF-8
430
431    is( fc($troublesome4), "\x{3BC}", "fc() for a latin-1 \x{B5} returns UTF-8" );
432    ok( !utf8::is_utf8($troublesome4), "fc() doesn't upgrade the original string" );
433
434
435    my $troublesome5 = uni_to_native("\x{C9}") . "abda"
436                     . uni_to_native("\x{B5}") . "aaf"
437                     . uni_to_native("\x{C8}");  # Up until foldcasing \x{B5}, the string
438                                                    # was in Latin-1. This tests that the
439                                                    # results don't have illegal UTF-8
440                                                    # (i.e. leftover latin-1) in them
441
442    is( fc($troublesome5), uni_to_native("\x{E9}") . "abda\x{3BC}aaf"
443                         . uni_to_native("\x{E8}") );
444}
445
446
447SKIP: {
448    use feature qw( fc unicode_strings );
449
450    skip "locales not available", 256 unless locales_enabled('LC_ALL');
451
452    setlocale(&POSIX::LC_ALL, "C");
453
454    # This tests both code paths in pp_fc
455
456    for (0..0xff) {
457        my $latin1 = chr;
458        my $utf8   = $latin1;
459        utf8::downgrade($latin1); #No-op, but doesn't hurt
460        utf8::upgrade($utf8);
461        is(fc($latin1), fc($utf8), "fc() gives the same results for \\x{$_} in Latin-1 and UTF-8 under unicode_strings");
462        SKIP: {
463            skip 'Locales not available', 2 unless locales_enabled('LC_CTYPE');
464            use locale;
465            is(fc($latin1), lc($latin1), "use locale; fc(qq{\\x{$_}}), lc(qq{\\x{$_}}) when qq{\\x{$_}} is in latin-1");
466            is(fc($utf8), lc($utf8), "use locale; fc(qq{\\x{$_}}), lc(qq{\\x{$_}}) when qq{\\x{$_}} is in latin-1");
467        }
468        {
469            no feature 'unicode_strings';
470            is(fc($latin1), lc($latin1), "under nothing, fc() for <256 is the same as lc");
471        }
472    }
473}
474
475my $utf8_locale = find_utf8_ctype_locale();
476
477{
478    use feature qw( fc );
479    use locale;
480    no warnings 'locale';   # Would otherwise warn
481    is(fc("\x{1E9E}"), fc("\x{17F}\x{17F}"), 'fc("\x{1E9E}") eq fc("\x{17F}\x{17F}")');
482    use warnings 'locale';
483    SKIP: {
484        skip 'Can\'t find a UTF-8 locale', 1 unless defined $utf8_locale;
485        setlocale(&LC_CTYPE, $utf8_locale);
486        is(fc("\x{1E9E}"), "ss", 'fc("\x{1E9E}") eq "ss" in a UTF-8 locale)');
487    }
488}
489
490SKIP: {
491    skip 'Can\'t find a UTF-8 locale', 256 unless defined $utf8_locale;
492
493    use feature qw( fc unicode_strings );
494
495    # Get the official fc values outside locale.
496    no locale;
497    my @unicode_fc;
498    for (0..0xff) {
499        push @unicode_fc, fc(chr);
500    }
501
502    # These should match the UTF-8 locale values
503    setlocale(&LC_CTYPE, $utf8_locale);
504    use locale;
505    for (0..0xff) {
506        is(fc(chr), $unicode_fc[$_], "In a UTF-8 locale, fc(chr $_) is the same as official Unicode");
507    }
508}
509
510
511my $num_tests = curr_test() - 1;
512
513plan($num_tests);
514