4 require 'loc_tools.pl'; # Contains locales_enabled() and
5 # find_utf8_ctype_locale()
14 my $tab = " " x 4; # Indent subsidiary tests this much
16 use Unicode::UCD qw(search_invlist prop_invmap prop_invlist);
17 my ($charname_list, $charname_map, $format, $default) = prop_invmap("Name Alias");
22 # If there is a an abbreviation for the code point name, use it
23 my $name_index = search_invlist(\@{$charname_list}, $cp);
24 if (defined $name_index) {
25 my $synonyms = $charname_map->[$name_index];
27 my $pat = qr/: abbreviation/;
28 my @abbreviations = grep { $_ =~ $pat } @$synonyms;
30 return $abbreviations[0] =~ s/$pat//r;
35 # Otherwise, use the full name
37 return charnames::viacode($cp) // "No name";
40 sub truth($) { # Converts values so is() works
41 return (shift) ? 1 : 0;
46 if(locales_enabled('LC_ALL')) {
48 $base_locale = POSIX::setlocale( &POSIX::LC_ALL, "C");
49 if (defined $base_locale && $base_locale eq 'C') {
50 use locale; # make \w work right in non-ASCII lands
52 # Some locale implementations don't have the 128-255 characters all
53 # mean nothing. Skip the locale tests in that situation
54 for my $i (128 .. 255) {
55 if (chr(utf8::unicode_to_native($i)) =~ /[[:print:]]/) {
61 $utf8_locale = find_utf8_ctype_locale() if $base_locale;
65 sub get_display_locale_or_skip($$) {
67 # Helper function intimately tied to its callers. It knows the loop
68 # iterates with a locale of "", meaning don't use locale; $base_locale
69 # meaning to use a non-UTF-8 locale; and $utf8_locale.
71 # It checks to see if the current test should be skipped or executed,
72 # returning an empty list for the former, and for the latter:
73 # ( 'locale display name',
74 # bool of is this a UTF-8 locale )
76 # The display name is the empty string if not using locale. Functions
77 # with _LC in their name are skipped unless in locale, and functions
78 # without _LC are executed only outside locale. However, if no locales at
79 # all are on the system, the _LC functions are executed outside locale.
81 my ($locale, $suffix) = @_;
83 # The test should be skipped if the input is for a non-existent locale
84 return unless defined $locale;
86 # Here the input is defined, either a locale name or "". If the test is
87 # for not using locales, we want to do the test for non-LC functions,
88 # and skip it for LC ones (except if there are no locales on the system,
89 # we do it for LC ones as if they weren't LC).
91 return ("", 0) if $suffix !~ /LC/ || ! defined $base_locale;
95 # Here the input is for a real locale. We don't test the non-LC functions
97 return if $suffix !~ /LC/;
99 # Here is for a LC function and a real locale. The base locale is not
101 return (" ($locale locale)", 0) if $locale eq $base_locale;
103 # The only other possibility is that we have a UTF-8 locale
104 return (" ($locale)", 1);
107 sub try_malforming($$$)
109 # Determines if the tests for malformed UTF-8 should be done. When done,
110 # the .xs code creates malformations by pretending the length is shorter
111 # than it actually is. Some things can't be malformed, and sometimes this
112 # test knows that the current code doesn't look for a malformation under
113 # various circumstances.
115 my ($i, $function, $using_locale) = @_;
116 # $i is unicode code point;
118 # Single bytes can't be malformed
119 return 0 if $i < ((ord "A" == 65) ? 128 : 160);
121 # ASCII doesn't need to ever look beyond the first byte.
122 return 0 if $function eq "ASCII";
124 # No controls above 255, so the code doesn't look at those
125 return 0 if $i > 255 && $function eq "CNTRL";
127 # No non-ASCII digits below 256, except if using locales.
128 return 0 if $i < 256 && ! $using_locale && $function =~ /X?DIGIT/;
134 # name => Lookup-property name
137 alphanumeric => 'Alnum',
138 alpha => 'XPosixAlpha',
144 idfirst => '_Perl_IDStart',
145 idcont => '_Perl_IDCont',
146 lower => 'XPosixLower',
148 psxspc => 'XPosixSpace',
149 punct => 'XPosixPunct',
150 quotemeta => '_Perl_Quotemeta',
151 space => 'XPerlSpace',
152 vertws => 'VertSpace',
153 upper => 'XPosixUpper',
159 local $SIG{__WARN__} = sub { push @warnings, @_ };
161 my %utf8_param_code = (
163 "_safe, malformed" => 1,
164 "deprecated unsafe" => -1,
165 "deprecated mathoms" => -2,
168 foreach my $name (sort keys %properties, 'octal') {
170 if ($name eq 'octal') {
171 # Hand-roll an inversion list with 0-7 in it and nothing else.
172 push @invlist, ord "0", ord "8";
175 my $property = $properties{$name};
176 @invlist = prop_invlist($property, '_perl_core_internal_ok');
179 # An empty return could mean an unknown property, or merely that
180 # it is empty. Call in scalar context to differentiate
181 if (! prop_invlist($property, '_perl_core_internal_ok')) {
182 fail("No inversion list found for $property");
188 # Include all the Latin1 code points, plus 0x100.
189 my @code_points = (0 .. 256);
191 # Then include the next few boundaries above those from this property
192 my $above_latins = 0;
193 foreach my $range_start (@invlist) {
194 next if $range_start < 257;
195 push @code_points, $range_start - 1, $range_start;
197 last if $above_latins > 5;
200 # This makes sure we are using the Perl definition of idfirst and idcont,
201 # and not the Unicode. There are a few differences.
202 push @code_points, ord "\N{ESTIMATED SYMBOL}" if $name =~ /^id(first|cont)/;
203 if ($name eq "idcont") { # And some that are continuation but not start
204 push @code_points, ord("\N{GREEK ANO TELEIA}"),
205 ord("\N{COMBINING GRAVE ACCENT}");
208 # And finally one non-Unicode code point.
209 push @code_points, 0x110000; # Above Unicode, no prop should match
210 no warnings 'non_unicode';
212 for my $j (@code_points) {
213 my $i = utf8::native_to_unicode($j);
214 my $function = uc($name);
216 is (@warnings, 0, "Got no unexpected warnings in previous iteration")
217 or diag("@warnings");
220 my $matches = search_invlist(\@invlist, $j);
221 if (! defined $matches) {
225 $matches = truth(! ($matches % 2));
229 my $char_name = get_charname($j);
230 my $display_name = sprintf "\\x{%02X, %s}", $j, $char_name;
231 my $display_call = "is${function}( $display_name )";
233 foreach my $suffix ("", "_A", "_L1", "_LC", "_uni", "_uvchr",
234 "_LC_uvchr", "_utf8", "_LC_utf8")
237 # Not all possible macros have been defined
238 if ($name eq 'vertws') {
240 # vertws is always all of Unicode
241 next if $suffix !~ / ^ _ ( uni | uvchr | utf8 ) $ /x;
243 elsif ($name eq 'alnum') {
245 # ALNUM_A, ALNUM_L1, and ALNUM_uvchr are not defined as these
246 # suffixes were added later, after WORDCHAR was created to be
247 # a clearer synonym for ALNUM
248 next if $suffix eq '_A'
250 || $suffix eq '_uvchr';
252 elsif ($name eq 'octal') {
253 next if $suffix ne "" && $suffix ne '_A' && $suffix ne '_L1';
255 elsif ($name eq 'quotemeta') {
256 # There is only one macro for this, and is defined only for
258 next if $suffix ne ""
261 foreach my $locale ("", $base_locale, $utf8_locale) {
263 my ($display_locale, $locale_is_utf8)
264 = get_display_locale_or_skip($locale, $suffix);
265 next unless defined $display_locale;
267 use if $locale, "locale";
268 POSIX::setlocale( &POSIX::LC_ALL, $locale) if $locale;
270 if ($suffix !~ /utf8/) { # _utf8 has to handled specially
272 = "is${function}$suffix( $display_name )$display_locale";
273 $ret = truth eval "test_is${function}$suffix($j)";
274 if (is ($@, "", "$display_call didn't give error")) {
275 my $truth = $matches;
278 # The single byte functions are false for
282 if $suffix=~ / ^ ( _A | _L [1C] )? $ /x;
285 && $name ne 'quotemeta')
288 # The no-suffix and _A functions are false
289 # for non-ASCII. So are _LC functions on a
291 $truth = 0 if $suffix eq "_A"
294 && ! $locale_is_utf8);
298 is ($ret, $truth, "${tab}And correctly returns $truth");
301 else { # _utf8 suffix
303 utf8::upgrade($char);
304 $char = quotemeta $char if $char eq '\\' || $char eq "'";
310 { # The C-locale _LC function returns FALSE for Latin1
318 foreach my $utf8_param("_safe",
323 my $utf8_param_code = $utf8_param_code{$utf8_param};
324 my $expect_error = $utf8_param_code > 0;
325 next if $expect_error
326 && ! try_malforming($i, $function,
329 my $display_call = "is${function}$suffix( $display_name"
330 . ", $utf8_param )$display_locale";
331 $ret = truth eval "test_is${function}$suffix('$char',"
332 . " $utf8_param_code)";
335 "expected and got error in $display_call");
336 like($@, qr/Malformed UTF-8 character/,
337 "${tab}And got expected message");
338 if (is (@warnings, 1,
339 "${tab}Got a single warning besides"))
342 qr/Malformed UTF-8 character.*short/,
343 "${tab}Got expected warning");
350 elsif (is ($@, "", "$display_call didn't give error")) {
352 "${tab}And correctly returned $truth");
353 if ($utf8_param_code < 0) {
355 my $unique_function = "is" . $function . $suffix;
356 if (! $seen{$unique_function}++) {
357 $warnings_ok = is(@warnings, 1,
358 "${tab}This is first call to"
359 . " $unique_function; Got a single"
362 $warnings_ok = like($warnings[0],
363 qr/starting in Perl .* will require an additional parameter/,
364 "${tab}The warning was the expected"
365 . " deprecation one");
369 $warnings_ok = is(@warnings, 0,
370 "${tab}This subsequent call to"
371 . " $unique_function did not warn");
373 $warnings_ok or diag("@warnings");
384 my %to_properties = (
385 FOLD => 'Case_Folding',
386 LOWER => 'Lowercase_Mapping',
387 TITLE => 'Titlecase_Mapping',
388 UPPER => 'Uppercase_Mapping',
392 foreach my $name (sort keys %to_properties) {
393 my $property = $to_properties{$name};
394 my ($list_ref, $map_ref, $format, $missing)
395 = prop_invmap($property, );
396 if (! $list_ref || ! $map_ref) {
397 fail("No inversion map found for $property");
400 if ($format !~ / ^ a l? $ /x) {
401 fail("Unexpected inversion map format ('$format') found for $property");
405 # Include all the Latin1 code points, plus 0x100.
406 my @code_points = (0 .. 256);
408 # Then include the next few multi-char folds above those from this
409 # property, and include the next few single folds as well
410 my $above_latins = 0;
412 for my $i (0 .. @{$list_ref} - 1) {
413 my $range_start = $list_ref->[$i];
414 next if $range_start < 257;
415 if (ref $map_ref->[$i] && $multi_char < 5) {
416 push @code_points, $range_start - 1
417 if $code_points[-1] != $range_start - 1;
418 push @code_points, $range_start;
421 elsif ($above_latins < 5) {
422 push @code_points, $range_start - 1
423 if $code_points[-1] != $range_start - 1;
424 push @code_points, $range_start;
427 last if $above_latins >= 5 && $multi_char >= 5;
430 # And finally one non-Unicode code point.
431 push @code_points, 0x110000; # Above Unicode, no prop should match
432 no warnings 'non_unicode';
434 # $j is native; $i unicode.
435 for my $j (@code_points) {
436 my $i = utf8::native_to_unicode($j);
437 my $function = $name;
439 my $index = search_invlist(\@{$list_ref}, $j);
442 my $char_name = get_charname($j);
443 my $display_name = sprintf "\\N{U+%02X, %s}", $j, $char_name;
445 foreach my $suffix ("", "_L1", "_LC") {
447 # This is the only macro defined for L1
448 next if $suffix eq "_L1" && $function ne "LOWER";
451 foreach my $locale ("", $base_locale, $utf8_locale) {
453 # titlecase is not defined in locales.
454 next if $name eq 'TITLE' && $suffix eq "_LC";
456 my ($display_locale, $locale_is_utf8)
457 = get_display_locale_or_skip($locale, $suffix);
458 next unless defined $display_locale;
460 skip("to${name}_LC does not work for LATIN SMALL LETTER SHARP S"
461 . "$display_locale", 1)
462 if $i == 0xDF && $name =~ / FOLD | UPPER /x
463 && $suffix eq "_LC" && $locale_is_utf8;
465 use if $locale, "locale";
466 POSIX::setlocale( &POSIX::LC_ALL, $locale) if $locale;
468 my $display_call = "to${function}$suffix("
469 . " $display_name )$display_locale";
470 $ret = eval "test_to${function}$suffix($j)";
471 if (is ($@, "", "$display_call didn't give error")) {
478 || ($suffix eq "_LC" && ! $locale_is_utf8)))
482 elsif ($map_ref->[$index] != $missing) {
483 $should_be = $map_ref->[$index] + $j - $list_ref->[$index]
489 is ($ret, $should_be,
490 sprintf("${tab}And correctly returned 0x%02X",
496 # The _uni, uvchr, and _utf8 functions return both the ordinal of the
497 # first code point of the result, and the result in utf8. The .xs
498 # tests return these in an array, in [0] and [1] respectively, with
499 # [2] the length of the utf8 in bytes.
500 my $utf8_should_be = "";
501 my $first_ord_should_be;
502 if (ref $map_ref->[$index]) { # A multi-char result
503 for my $j (0 .. @{$map_ref->[$index]} - 1) {
504 $utf8_should_be .= chr $map_ref->[$index][$j];
507 $first_ord_should_be = $map_ref->[$index][0];
509 else { # A single-char result
510 $first_ord_should_be = ($map_ref->[$index] != $missing)
511 ? $map_ref->[$index] + $j
512 - $list_ref->[$index]
514 $utf8_should_be = chr $first_ord_should_be;
516 utf8::upgrade($utf8_should_be);
519 foreach my $suffix ('_uni', '_uvchr') {
522 my $display_call = "to${function}$suffix( $display_name )";
523 $ret = eval "test_to${function}$suffix($j)";
524 if (is ($@, "", "$display_call didn't give error")) {
525 is ($ret->[0], $first_ord_should_be,
526 sprintf("${tab}And correctly returned 0x%02X",
527 $first_ord_should_be));
528 is ($ret->[1], $utf8_should_be, "${tab}Got correct utf8");
530 is ($ret->[2], length $utf8_should_be,
531 "${tab}Got correct number of bytes for utf8 length");
537 utf8::upgrade($char);
538 $char = quotemeta $char if $char eq '\\' || $char eq "'";
539 foreach my $utf8_param("_safe",
542 "deprecated mathoms",
546 next if $utf8_param eq 'deprecated mathoms'
547 && $Config{'ccflags'} =~ /-DNO_MATHOMS/;
549 my $utf8_param_code = $utf8_param_code{$utf8_param};
550 my $expect_error = $utf8_param_code > 0;
552 # Skip if can't malform (because is a UTF-8 invariant)
553 next if $expect_error && $i < ((ord "A" == 65) ? 128 : 160);
555 my $display_call = "to${function}_utf8($display_name, $utf8_param )";
556 $ret = eval "test_to${function}_utf8('$char', $utf8_param_code)";
558 isnt ($@, "", "expected and got error in $display_call");
559 like($@, qr/Malformed UTF-8 character/,
560 "${tab}And got expected message");
563 elsif (is ($@, "", "$display_call didn't give error")) {
564 is ($ret->[0], $first_ord_should_be,
565 sprintf("${tab}And correctly returned 0x%02X",
566 $first_ord_should_be));
567 is ($ret->[1], $utf8_should_be, "${tab}Got correct utf8");
569 is ($ret->[2], length $utf8_should_be,
570 "${tab}Got correct number of bytes for utf8 length");
571 if ($utf8_param_code < 0) {
573 if (! $seen{"${function}_utf8$utf8_param"}++) {
574 $warnings_ok = is(@warnings, 1,
575 "${tab}Got a single warning");
578 if ($utf8_param_code == -2) {
579 my $lc_func = lc $function;
581 = qr/starting in Perl .* to_utf8_$lc_func\(\) will be removed/;
585 = qr/starting in Perl .* will require an additional parameter/;
587 $warnings_ok = like($warnings[0], $expected,
588 "${tab}Got expected deprecation warning");
592 $warnings_ok = is(@warnings, 0,
593 "${tab}Deprecation warned only the one time");
595 $warnings_ok or diag("@warnings");
603 # This is primarily to make sure that no non-Unicode warnings get generated
604 is(scalar @warnings, 0, "No unexpected warnings were generated in the tests")