# without using 'eval' as much as possible, which might cloud the issue, the
# crucial parts of the code are duplicated in a block for each pragma.
+# To make a TODO test, add the string 'TODO' to its %test_names value
+
binmode STDOUT, ':utf8';
binmode STDERR, ':utf8';
}
use strict;
+use feature 'fc';
+
+my $debug = $ENV{PERL_DEBUG_FULL_TEST} // 0;
-my $debug = 1;
+# Certain tests have been shown to be problematical for a few locales. Don't
+# fail them unless at least this percentage of the tested locales fail.
+# Some Windows machines are defective in every in every locale but the C,
+# calling \t printable; superscripts to be digits, etc. See
+# http://markmail.org/message/5jwam4xsx4amsdnv
+# (There aren't 1000 locales currently in existence, so 99.9 works)
+my $acceptable_fold_failure_percentage = $^O eq 'MSWin32' ? 99.9 : 5;
use Dumpvalue;
# UWIN seems to loop after taint tests, just skip for now
$have_setlocale = 0 if ($^O =~ /^uwin/);
-sub LC_ALL ();
-
$a = 'abc %';
my $test_num = 0;
not eval { $dummy = join("", @_), kill 0; 1 }
}
-sub check_taint ($) {
- ok is_tainted($_[0]), "verify that is tainted";
+sub check_taint ($;$) {
+ my $message_tail = $_[1] // "";
+ $message_tail = ": $message_tail" if $message_tail;
+ ok is_tainted($_[0]), "verify that is tainted$message_tail";
}
-sub check_taint_not ($) {
- ok((not is_tainted($_[0])), "verify that isn't tainted");
+sub check_taint_not ($;$) {
+ my $message_tail = $_[1] // "";
+ $message_tail = ": $message_tail" if $message_tail;
+ ok((not is_tainted($_[0])), "verify that isn't tainted$message_tail");
}
+"\tb\t" =~ /^m?(\s)(.*)\1$/;
+check_taint_not $&, "not tainted outside 'use locale'";
+;
+
use locale; # engage locale and therefore locale taint.
check_taint_not $a;
check_taint ucfirst($a);
check_taint "\u$a";
check_taint lc($a);
+check_taint fc($a);
check_taint "\L$a";
+check_taint "\F$a";
check_taint lcfirst($a);
check_taint "\l$a";
check_taint_not ucfirst($a);
check_taint_not "\u$a";
check_taint_not lc($a);
+ check_taint_not fc($a);
check_taint_not "\L$a";
+ check_taint_not "\F$a";
check_taint_not lcfirst($a);
check_taint_not "\l$a";
Hrvatski Croatian:hr:hr:2
Cymraeg Welsh:cy:cy:1 14 15
Czech:cs:cz:2
-Dansk Danish:dk:da:1 15
+Dansk Danish:da:dk:1 15
Nederlands Dutch:nl:be nl:1 15
English American British:en:au ca gb ie nz us uk zw:1 15 cp850
Esperanto:eo:eo:3
Greenlandic:kl:gl:4 6
Hebrew:iw:il:8 hebrew8
Hungarian:hu:hu:2
-Indonesian:in:id:1 15
+Indonesian:id:id:1 15
Gaeilge Irish:ga:IE:1 14 15
Italiano Italian:it:ch it:1 15
Nihongo Japanese:ja:jp:euc eucJP jp.EUC sjis
Macedonian:mk:mk:1 15
Maltese:mt:mt:3
Moldovan:mo:mo:2
-Norsk Norwegian:no no\@nynorsk:no:1 15
+Norsk Norwegian:no no\@nynorsk nb nn:no:1 15
Occitan:oc:es:1 15
Polski Polish:pl:pl:2
Rumanian:ro:ro:2
my @Locale;
my $Locale;
-my @Alnum_;
+my %posixes;
sub trylocale {
my $locale = shift;
return if grep { $locale eq $_ } @Locale;
- return unless setlocale(LC_ALL, $locale);
+ return unless setlocale(&POSIX::LC_ALL, $locale);
my $badutf8;
{
local $SIG{__WARN__} = sub {
trylocale($_);
}
close(LOCALES);
-} elsif ($^O eq 'openbsd' && -e '/usr/share/locale') {
+} elsif (($^O eq 'openbsd' || $^O eq 'bitrig' ) && -e '/usr/share/locale') {
# OpenBSD doesn't have a locale executable, so reading /usr/share/locale
# is much easier and faster than the last resort method.
}
}
-setlocale(LC_ALL, "C");
+setlocale(&POSIX::LC_ALL, "C");
if ($^O eq 'darwin') {
# Darwin 8/Mac OS X 10.4 and 10.5 have bad Basque locales: perl bug #35895,
my %Problem;
my %Okay;
my %Testing;
-my @Neoalpha; # Alnums that aren't in the C locale.
+my @Added_alpha; # Alphas that aren't in the C locale.
my %test_names;
-sub tryneoalpha {
- my ($Locale, $i, $test) = @_;
- unless ($test) {
+sub display_characters {
+ # This returns a display string denoting the input parameter @_, each
+ # entry of which is a single character in the range 0-255. The first part
+ # of the output is a string of the characters in @_ that are ASCII
+ # graphics, and hence unambiguously displayable. They are given by code
+ # point order. The second part is the remaining code points, the ordinals
+ # of which are each displayed as 2-digit hex. Blanks are inserted so as
+ # to keep anything from the first part looking like a 2-digit hex number.
+
+ no locale;
+ my @chars = sort { ord $a <=> ord $b } @_;
+ my $output = "";
+ my $range_start;
+ my $start_class;
+ push @chars, chr(258); # This sentinel simplifies the loop termination
+ # logic
+ foreach my $i (0 .. @chars - 1) {
+ my $char = $chars[$i];
+ my $range_end;
+ my $class;
+
+ # We avoid using [:posix:] classes, as these are being tested in this
+ # file. Each equivalence class below is for things that can appear in
+ # a range; those that can't be in a range have class -1. 0 for those
+ # which should be output in hex; and >0 for the other ranges
+ if ($char =~ /[A-Z]/) {
+ $class = 2;
+ }
+ elsif ($char =~ /[a-z]/) {
+ $class = 3;
+ }
+ elsif ($char =~ /[0-9]/) {
+ $class = 4;
+ }
+ # Uncomment to get literal punctuation displayed instead of hex
+ #elsif ($char =~ /[[\]!"#\$\%&\'()*+,.\/:\\;<=>?\@\^_`{|}~-]/) {
+ # $class = -1; # Punct never appears in a range
+ #}
+ else {
+ $class = 0; # Output in hex
+ }
+
+ if (! defined $range_start) {
+ if ($class < 0) {
+ $output .= " " . $char;
+ }
+ else {
+ $range_start = ord $char;
+ $start_class = $class;
+ }
+ } # A range ends if not consecutive, or the class-type changes
+ elsif (ord $char != ($range_end = ord($chars[$i-1])) + 1
+ || $class != $start_class)
+ {
+
+ # Here, the current character is not in the range. This means the
+ # previous character must have been. Output the range up through
+ # that one.
+ my $range_length = $range_end - $range_start + 1;
+ if ($start_class > 0) {
+ $output .= " " . chr($range_start);
+ $output .= "-" . chr($range_end) if $range_length > 1;
+ }
+ else {
+ $output .= sprintf(" %02X", $range_start);
+ $output .= sprintf("-%02X", $range_end) if $range_length > 1;
+ }
+
+ # Handle the new current character, as potentially beginning a new
+ # range
+ undef $range_start;
+ redo;
+ }
+ }
+
+ $output =~ s/^ //;
+ return $output;
+}
+
+sub report_result {
+ my ($Locale, $i, $pass_fail, $message) = @_;
+ $message //= "";
+ $message = " ($message)" if $message;
+ unless ($pass_fail) {
$Problem{$i}{$Locale} = 1;
- debug "# failed $i with locale '$Locale'\n";
+ debug "# failed $i ($test_names{$i}) with locale '$Locale'$message\n";
} else {
push @{$Okay{$i}}, $Locale;
}
}
+sub report_multi_result {
+ my ($Locale, $i, $results_ref) = @_;
+
+ # $results_ref points to an array, each element of which is a character that was
+ # in error for this test numbered '$i'. If empty, the test passed
+
+ my $message = "";
+ if (@$results_ref) {
+ $message = join " ", "for", display_characters(@$results_ref);
+ }
+ report_result($Locale, $i, @$results_ref == 0, $message);
+}
+
my $first_locales_test_number = $final_without_setlocale + 1;
my $locales_test_number;
my $not_necessarily_a_problem_test_number;
+my $first_casing_test_number;
+my $final_casing_test_number;
my %setlocale_failed; # List of locales that setlocale() didn't work on
foreach $Locale (@Locale) {
$locales_test_number = $first_locales_test_number - 1;
+ debug "#\n";
debug "# Locale = $Locale\n";
- unless (setlocale(LC_ALL, $Locale)) {
+ unless (setlocale(&POSIX::LC_ALL, $Locale)) {
$setlocale_failed{$Locale} = $Locale;
next;
}
# documented deficiencies. Non- UTF-8 locales are tested only under plain
# 'use locale', as otherwise we would have to convert everything in them
# to Unicode.
+ # The locale name doesn't necessarily have to have "utf8" in it to be a
+ # UTF-8 locale, but it works mostly.
my $is_utf8_locale = $Locale =~ /UTF-?8/i;
- my %UPPER = ();
- my %lower = ();
- my %BoThCaSe = ();
+ my %UPPER = (); # All alpha X for which uc(X) == X and lc(X) != X
+ my %lower = (); # All alpha X for which lc(X) == X and uc(X) != X
+ my %BoThCaSe = (); # All alpha X for which uc(X) == lc(X) == X
if (! $is_utf8_locale) {
use locale;
- @Alnum_ = sort grep /\w/, map { chr } 0..255;
-
- debug "# w = ", join("",@Alnum_), "\n";
+ @{$posixes{'word'}} = grep /\w/, map { chr } 0..255;
+ @{$posixes{'digit'}} = grep /\d/, map { chr } 0..255;
+ @{$posixes{'space'}} = grep /\s/, map { chr } 0..255;
+ @{$posixes{'alpha'}} = grep /[[:alpha:]]/, map {chr } 0..255;
+ @{$posixes{'alnum'}} = grep /[[:alnum:]]/, map {chr } 0..255;
+ @{$posixes{'ascii'}} = grep /[[:ascii:]]/, map {chr } 0..255;
+ @{$posixes{'blank'}} = grep /[[:blank:]]/, map {chr } 0..255;
+ @{$posixes{'cntrl'}} = grep /[[:cntrl:]]/, map {chr } 0..255;
+ @{$posixes{'graph'}} = grep /[[:graph:]]/, map {chr } 0..255;
+ @{$posixes{'lower'}} = grep /[[:lower:]]/, map {chr } 0..255;
+ @{$posixes{'print'}} = grep /[[:print:]]/, map {chr } 0..255;
+ @{$posixes{'punct'}} = grep /[[:punct:]]/, map {chr } 0..255;
+ @{$posixes{'upper'}} = grep /[[:upper:]]/, map {chr } 0..255;
+ @{$posixes{'xdigit'}} = grep /[[:xdigit:]]/, map {chr } 0..255;
+ @{$posixes{'cased'}} = grep /[[:upper:]]/i, map {chr } 0..255;
# Sieve the uppercase and the lowercase.
- for (@Alnum_) {
+ for (@{$posixes{'word'}}) {
if (/[^\d_]/) { # skip digits and the _
if (uc($_) eq $_) {
$UPPER{$_} = $_;
}
else {
use locale ':not_characters';
- @Alnum_ = sort grep /\w/, map { chr } 0..255;
- debug "# w = ", join("",@Alnum_), "\n";
- for (@Alnum_) {
+ @{$posixes{'word'}} = grep /\w/, map { chr } 0..255;
+ @{$posixes{'digit'}} = grep /\d/, map { chr } 0..255;
+ @{$posixes{'space'}} = grep /\s/, map { chr } 0..255;
+ @{$posixes{'alpha'}} = grep /[[:alpha:]]/, map {chr } 0..255;
+ @{$posixes{'alnum'}} = grep /[[:alnum:]]/, map {chr } 0..255;
+ @{$posixes{'ascii'}} = grep /[[:ascii:]]/, map {chr } 0..255;
+ @{$posixes{'blank'}} = grep /[[:blank:]]/, map {chr } 0..255;
+ @{$posixes{'cntrl'}} = grep /[[:cntrl:]]/, map {chr } 0..255;
+ @{$posixes{'graph'}} = grep /[[:graph:]]/, map {chr } 0..255;
+ @{$posixes{'lower'}} = grep /[[:lower:]]/, map {chr } 0..255;
+ @{$posixes{'print'}} = grep /[[:print:]]/, map {chr } 0..255;
+ @{$posixes{'punct'}} = grep /[[:punct:]]/, map {chr } 0..255;
+ @{$posixes{'upper'}} = grep /[[:upper:]]/, map {chr } 0..255;
+ @{$posixes{'xdigit'}} = grep /[[:xdigit:]]/, map {chr } 0..255;
+ @{$posixes{'cased'}} = grep /[[:upper:]]/i, map {chr } 0..255;
+ for (@{$posixes{'word'}}) {
if (/[^\d_]/) { # skip digits and the _
if (uc($_) eq $_) {
$UPPER{$_} = $_;
}
}
}
+
+ # Ordered, where possible, in groups of "this is a subset of the next
+ # one"
+ debug "# :upper: = ", display_characters(@{$posixes{'upper'}}), "\n";
+ debug "# :lower: = ", display_characters(@{$posixes{'lower'}}), "\n";
+ debug "# :cased: = ", display_characters(@{$posixes{'cased'}}), "\n";
+ debug "# :alpha: = ", display_characters(@{$posixes{'alpha'}}), "\n";
+ debug "# :alnum: = ", display_characters(@{$posixes{'alnum'}}), "\n";
+ debug "# w = ", display_characters(@{$posixes{'word'}}), "\n";
+ debug "# :graph: = ", display_characters(@{$posixes{'graph'}}), "\n";
+ debug "# :print: = ", display_characters(@{$posixes{'print'}}), "\n";
+ debug "# d = ", display_characters(@{$posixes{'digit'}}), "\n";
+ debug "# :xdigit: = ", display_characters(@{$posixes{'xdigit'}}), "\n";
+ debug "# :blank: = ", display_characters(@{$posixes{'blank'}}), "\n";
+ debug "# s = ", display_characters(@{$posixes{'space'}}), "\n";
+ debug "# :punct: = ", display_characters(@{$posixes{'punct'}}), "\n";
+ debug "# :cntrl: = ", display_characters(@{$posixes{'cntrl'}}), "\n";
+ debug "# :ascii: = ", display_characters(@{$posixes{'ascii'}}), "\n";
+
foreach (keys %UPPER) {
+
$BoThCaSe{$_}++ if exists $lower{$_};
}
foreach (keys %lower) {
delete $lower{$_};
}
- debug "# UPPER = ", join("", sort keys %UPPER ), "\n";
- debug "# lower = ", join("", sort keys %lower ), "\n";
- debug "# BoThCaSe = ", join("", sort keys %BoThCaSe), "\n";
+ my %Unassigned;
+ foreach my $ord ( 0 .. 255 ) {
+ $Unassigned{chr $ord} = 1;
+ }
+ foreach my $class (keys %posixes) {
+ foreach my $char (@{$posixes{$class}}) {
+ delete $Unassigned{$char};
+ }
+ }
+
+ debug "# UPPER = ", display_characters(keys %UPPER), "\n";
+ debug "# lower = ", display_characters(keys %lower), "\n";
+ debug "# BoThCaSe = ", display_characters(keys %BoThCaSe), "\n";
+ debug "# Unassigned = ", display_characters(sort { ord $a <=> ord $b } keys %Unassigned), "\n";
+
+ my @failures;
+ my @fold_failures;
+ foreach my $x (sort keys %UPPER) {
+ my $ok;
+ my $fold_ok;
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ $ok = $x =~ /[[:upper:]]/;
+ $fold_ok = $x =~ /[[:lower:]]/i;
+ }
+ else {
+ use locale;
+ $ok = $x =~ /[[:upper:]]/;
+ $fold_ok = $x =~ /[[:lower:]]/i;
+ }
+ push @failures, $x unless $ok;
+ push @fold_failures, $x unless $fold_ok;
+ }
+ $locales_test_number++;
+ $first_casing_test_number = $locales_test_number;
+ $test_names{$locales_test_number} = 'Verify that /[[:upper:]]/ matches all alpha X for which uc(X) == X and lc(X) != X';
+ report_multi_result($Locale, $locales_test_number, \@failures);
+
+ $locales_test_number++;
+
+ $test_names{$locales_test_number} = 'Verify that /[[:lower:]]/i matches all alpha X for which uc(X) == X and lc(X) != X';
+ report_multi_result($Locale, $locales_test_number, \@fold_failures);
+
+ undef @failures;
+ undef @fold_failures;
+
+ foreach my $x (sort keys %lower) {
+ my $ok;
+ my $fold_ok;
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ $ok = $x =~ /[[:lower:]]/;
+ $fold_ok = $x =~ /[[:upper:]]/i;
+ }
+ else {
+ use locale;
+ $ok = $x =~ /[[:lower:]]/;
+ $fold_ok = $x =~ /[[:upper:]]/i;
+ }
+ push @failures, $x unless $ok;
+ push @fold_failures, $x unless $fold_ok;
+ }
+
+ $locales_test_number++;
+ $test_names{$locales_test_number} = 'Verify that /[[:lower:]]/ matches all alpha X for which lc(X) == X and uc(X) != X';
+ report_multi_result($Locale, $locales_test_number, \@failures);
+
+ $locales_test_number++;
+ $test_names{$locales_test_number} = 'Verify that /[[:upper:]]/i matches all alpha X for which lc(X) == X and uc(X) != X';
+ report_multi_result($Locale, $locales_test_number, \@fold_failures);
{ # Find the alphabetic characters that are not considered alphabetics
# in the default (C) locale.
no locale;
- @Neoalpha = ();
- for (keys %UPPER, keys %lower) {
- push(@Neoalpha, $_) if (/\W/);
+ @Added_alpha = ();
+ for (keys %UPPER, keys %lower, keys %BoThCaSe) {
+ push(@Added_alpha, $_) if (/\W/);
}
}
- @Neoalpha = sort @Neoalpha;
+ @Added_alpha = sort @Added_alpha;
- debug "# Neoalpha = ", join("",@Neoalpha), "\n";
+ debug "# Added_alpha = ", display_characters(@Added_alpha), "\n";
- my $first_Neoalpha_test_number = $locales_test_number;
- my $final_Neoalpha_test_number = $first_Neoalpha_test_number + 4;
- if (@Neoalpha == 0) {
- # If we have no Neoalphas the remaining tests are no-ops.
- debug "# no Neoalpha, skipping tests $locales_test_number..$final_Neoalpha_test_number for locale '$Locale'\n";
- foreach ($locales_test_number+1..$final_Neoalpha_test_number) {
- push @{$Okay{$_}}, $Locale;
- $locales_test_number++;
- }
- } else {
+ # Cross-check the whole 8-bit character set.
- # Test \w.
+ ++$locales_test_number;
+ my @f;
+ $test_names{$locales_test_number} = 'Verify that \w and [:word:] are identical';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:word:]]/ == /\w/;
+ }
+ else {
+ push @f, $_ unless /[[:word:]]/ == /\w/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
- my $word = join('', @Neoalpha);
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that \d and [:digit:] are identical';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:digit:]]/ == /\d/;
+ }
+ else {
+ push @f, $_ unless /[[:digit:]]/ == /\d/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
- ++$locales_test_number;
- $test_names{$locales_test_number} = 'Verify that alnums outside the C locale match \w';
- my $ok;
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that \s and [:space:] are identical';
+ for (map { chr } 0..255) {
if ($is_utf8_locale) {
use locale ':not_characters';
- $ok = $word =~ /^(\w+)$/;
+ push @f, $_ unless /[[:space:]]/ == /\s/;
}
else {
- # Already in 'use locale'; this tests that exiting scopes works
- $ok = $word =~ /^(\w+)$/;
+ push @f, $_ unless /[[:space:]]/ == /\s/;
}
- tryneoalpha($Locale, $locales_test_number, $ok);
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
- # Cross-check the whole 8-bit character set.
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:posix:] and [:^posix:] are mutually exclusive';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless (/[[:alpha:]]/ xor /[[:^alpha:]]/) ||
+ (/[[:alnum:]]/ xor /[[:^alnum:]]/) ||
+ (/[[:ascii:]]/ xor /[[:^ascii:]]/) ||
+ (/[[:blank:]]/ xor /[[:^blank:]]/) ||
+ (/[[:cntrl:]]/ xor /[[:^cntrl:]]/) ||
+ (/[[:digit:]]/ xor /[[:^digit:]]/) ||
+ (/[[:graph:]]/ xor /[[:^graph:]]/) ||
+ (/[[:lower:]]/ xor /[[:^lower:]]/) ||
+ (/[[:print:]]/ xor /[[:^print:]]/) ||
+ (/[[:space:]]/ xor /[[:^space:]]/) ||
+ (/[[:upper:]]/ xor /[[:^upper:]]/) ||
+ (/[[:word:]]/ xor /[[:^word:]]/) ||
+ (/[[:xdigit:]]/ xor /[[:^xdigit:]]/) ||
+
+ # effectively is what [:cased:] would be if it existed.
+ (/[[:upper:]]/i xor /[[:^upper:]]/i);
+ }
+ else {
+ push @f, $_ unless (/[[:alpha:]]/ xor /[[:^alpha:]]/) ||
+ (/[[:alnum:]]/ xor /[[:^alnum:]]/) ||
+ (/[[:ascii:]]/ xor /[[:^ascii:]]/) ||
+ (/[[:blank:]]/ xor /[[:^blank:]]/) ||
+ (/[[:cntrl:]]/ xor /[[:^cntrl:]]/) ||
+ (/[[:digit:]]/ xor /[[:^digit:]]/) ||
+ (/[[:graph:]]/ xor /[[:^graph:]]/) ||
+ (/[[:lower:]]/ xor /[[:^lower:]]/) ||
+ (/[[:print:]]/ xor /[[:^print:]]/) ||
+ (/[[:space:]]/ xor /[[:^space:]]/) ||
+ (/[[:upper:]]/ xor /[[:^upper:]]/) ||
+ (/[[:word:]]/ xor /[[:^word:]]/) ||
+ (/[[:xdigit:]]/ xor /[[:^xdigit:]]/) ||
+ (/[[:upper:]]/i xor /[[:^upper:]]/i);
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
- ++$locales_test_number;
- $test_names{$locales_test_number} = 'Verify that \w and \W are mutually exclusive, as are \d, \D; \s, \S';
- for (map { chr } 0..255) {
+ # The rules for the relationships are given in:
+ # http://www.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap07.html
+
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:lower:] contains at least a-z';
+ for ('a' .. 'z') {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:lower:]]/;
+ }
+ else {
+ push @f, $_ unless /[[:lower:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:lower:] is a subset of [:alpha:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:lower:]]/ and ! /[[:alpha:]]/;
+ }
+ else {
+ push @f, $_ if /[[:lower:]]/ and ! /[[:alpha:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:upper:] contains at least A-Z';
+ for ('A' .. 'Z') {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:upper:]]/;
+ }
+ else {
+ push @f, $_ unless /[[:upper:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:upper:] is a subset of [:alpha:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:upper:]]/ and ! /[[:alpha:]]/;
+ }
+ else {
+ push @f, $_ if /[[:upper:]]/ and ! /[[:alpha:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that /[[:lower:]]/i is a subset of [:alpha:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:lower:]]/i and ! /[[:alpha:]]/;
+ }
+ else {
+ push @f, $_ if /[[:lower:]]/i and ! /[[:alpha:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:alpha:] is a subset of [:alnum:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:alpha:]]/ and ! /[[:alnum:]]/;
+ }
+ else {
+ push @f, $_ if /[[:alpha:]]/ and ! /[[:alnum:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:digit:] contains at least 0-9';
+ for ('0' .. '9') {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:digit:]]/;
+ }
+ else {
+ push @f, $_ unless /[[:digit:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:digit:] is a subset of [:alnum:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:digit:]]/ and ! /[[:alnum:]]/;
+ }
+ else {
+ push @f, $_ if /[[:digit:]]/ and ! /[[:alnum:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:digit:] matches either 10 or 20 code points';
+ report_result($Locale, $locales_test_number, @{$posixes{'digit'}} == 10 || @{$posixes{'digit'}} == 20);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that if there is a second set of digits in [:digit:], they are consecutive';
+ if (@{$posixes{'digit'}} == 20) {
+ my $previous_ord;
+ for (map { chr } 0..255) {
+ next unless /[[:digit:]]/;
+ next if /[0-9]/;
+ if (defined $previous_ord) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if ord $_ != $previous_ord + 1;
+ }
+ else {
+ push @f, $_ if ord $_ != $previous_ord + 1;
+ }
+ }
+ $previous_ord = ord $_;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:digit:] is a subset of [:xdigit:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:digit:]]/ and ! /[[:xdigit:]]/;
+ }
+ else {
+ push @f, $_ if /[[:digit:]]/ and ! /[[:xdigit:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:xdigit:] contains at least A-F, a-f';
+ for ('A' .. 'F', 'a' .. 'f') {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:xdigit:]]/;
+ }
+ else {
+ push @f, $_ unless /[[:xdigit:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that any additional members of [:xdigit:], are in groups of 6 consecutive code points';
+ my $previous_ord;
+ my $count = 0;
+ for (map { chr } 0..255) {
+ next unless /[[:xdigit:]]/;
+ next if /[[:digit:]]/;
+ next if /[A-Fa-f]/;
+ if (defined $previous_ord) {
if ($is_utf8_locale) {
use locale ':not_characters';
- $ok = (/\w/ xor /\W/) ||
- (/\d/ xor /\D/) ||
- (/\s/ xor /\S/);
+ push @f, $_ if ord $_ != $previous_ord + 1;
}
else {
- $ok = (/\w/ xor /\W/) ||
- (/\d/ xor /\D/) ||
- (/\s/ xor /\S/);
+ push @f, $_ if ord $_ != $previous_ord + 1;
}
- tryneoalpha($Locale, $locales_test_number, $ok);
- }
+ }
+ $count++;
+ if ($count == 6) {
+ undef $previous_ord;
+ }
+ else {
+ $previous_ord = ord $_;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:xdigit:] is a subset of [:graph:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:xdigit:]]/ and ! /[[:graph:]]/;
+ }
+ else {
+ push @f, $_ if /[[:xdigit:]]/ and ! /[[:graph:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ # Note that xdigit doesn't have to be a subset of alnum
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:punct:] is a subset of [:graph:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:punct:]]/ and ! /[[:graph:]]/;
+ }
+ else {
+ push @f, $_ if /[[:punct:]]/ and ! /[[:graph:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that the space character is not in [:graph:]';
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, " " if " " =~ /[[:graph:]]/;
+ }
+ else {
+ push @f, " " if " " =~ /[[:graph:]]/;
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:space:] contains at least [\f\n\r\t\cK ]';
+ for (' ', "\f", "\n", "\r", "\t", "\cK") {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:space:]]/;
+ }
+ else {
+ push @f, $_ unless /[[:space:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:blank:] contains at least [\t ]';
+ for (' ', "\t") {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ unless /[[:blank:]]/;
+ }
+ else {
+ push @f, $_ unless /[[:blank:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:blank:] is a subset of [:space:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:blank:]]/ and ! /[[:space:]]/;
+ }
+ else {
+ push @f, $_ if /[[:blank:]]/ and ! /[[:space:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:graph:] is a subset of [:print:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:graph:]]/ and ! /[[:print:]]/;
+ }
+ else {
+ push @f, $_ if /[[:graph:]]/ and ! /[[:print:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that the space character is in [:print:]';
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, " " if " " !~ /[[:print:]]/;
+ }
+ else {
+ push @f, " " if " " !~ /[[:print:]]/;
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that isn\'t both [:cntrl:] and [:print:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if (/[[:print:]]/ and /[[:cntrl:]]/);
+ }
+ else {
+ push @f, $_ if (/[[:print:]]/ and /[[:cntrl:]]/);
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that isn\'t both [:alpha:] and [:digit:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:alpha:]]/ and /[[:digit:]]/;
+ }
+ else {
+ push @f, $_ if /[[:alpha:]]/ and /[[:digit:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that isn\'t both [:alnum:] and [:punct:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:alnum:]]/ and /[[:punct:]]/;
+ }
+ else {
+ push @f, $_ if /[[:alnum:]]/ and /[[:punct:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that isn\'t both [:xdigit:] and [:punct:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if (/[[:punct:]]/ and /[[:xdigit:]]/);
+ }
+ else {
+ push @f, $_ if (/[[:punct:]]/ and /[[:xdigit:]]/);
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
- # Test for read-only scalars' locale vs non-locale comparisons.
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that isn\'t both [:graph:] and [:space:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if (/[[:graph:]]/ and /[[:space:]]/);
+ }
+ else {
+ push @f, $_ if (/[[:graph:]]/ and /[[:space:]]/);
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ $final_casing_test_number = $locales_test_number;
+
+ # Test for read-only scalars' locale vs non-locale comparisons.
+
+ {
+ no locale;
+ my $ok;
+ $a = "qwerty";
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ $ok = ($a cmp "qwerty") == 0;
+ }
+ else {
+ use locale;
+ $ok = ($a cmp "qwerty") == 0;
+ }
+ report_result($Locale, ++$locales_test_number, $ok);
+ $test_names{$locales_test_number} = 'Verify that cmp works with a read-only scalar; no- vs locale';
+ }
+
+ {
+ my ($from, $to, $lesser, $greater,
+ @test, %test, $test, $yes, $no, $sign);
- {
- no locale;
- $a = "qwerty";
+ ++$locales_test_number;
+ $test_names{$locales_test_number} = 'Verify that "le", "ne", etc work';
+ $not_necessarily_a_problem_test_number = $locales_test_number;
+ for (0..9) {
+ # Select a slice.
+ $from = int(($_*@{$posixes{'word'}})/10);
+ $to = $from + int(@{$posixes{'word'}}/10);
+ $to = $#{$posixes{'word'}} if ($to > $#{$posixes{'word'}});
+ $lesser = join('', @{$posixes{'word'}}[$from..$to]);
+ # Select a slice one character on.
+ $from++; $to++;
+ $to = $#{$posixes{'word'}} if ($to > $#{$posixes{'word'}});
+ $greater = join('', @{$posixes{'word'}}[$from..$to]);
if ($is_utf8_locale) {
use locale ':not_characters';
- $ok = ($a cmp "qwerty") == 0;
+ ($yes, $no, $sign) = ($lesser lt $greater
+ ? (" ", "not ", 1)
+ : ("not ", " ", -1));
}
else {
use locale;
- $ok = ($a cmp "qwerty") == 0;
+ ($yes, $no, $sign) = ($lesser lt $greater
+ ? (" ", "not ", 1)
+ : ("not ", " ", -1));
}
- tryneoalpha($Locale, ++$locales_test_number, $ok);
- $test_names{$locales_test_number} = 'Verify that cmp works with a read-only scalar; no- vs locale';
- }
-
- {
- my ($from, $to, $lesser, $greater,
- @test, %test, $test, $yes, $no, $sign);
-
- ++$locales_test_number;
- $test_names{$locales_test_number} = 'Verify that "le", "ne", etc work';
- $not_necessarily_a_problem_test_number = $locales_test_number;
- for (0..9) {
- # Select a slice.
- $from = int(($_*@Alnum_)/10);
- $to = $from + int(@Alnum_/10);
- $to = $#Alnum_ if ($to > $#Alnum_);
- $lesser = join('', @Alnum_[$from..$to]);
- # Select a slice one character on.
- $from++; $to++;
- $to = $#Alnum_ if ($to > $#Alnum_);
- $greater = join('', @Alnum_[$from..$to]);
+ # all these tests should FAIL (return 0). Exact lt or gt cannot
+ # be tested because in some locales, say, eacute and E may test
+ # equal.
+ @test =
+ (
+ $no.' ($lesser le $greater)', # 1
+ 'not ($lesser ne $greater)', # 2
+ ' ($lesser eq $greater)', # 3
+ $yes.' ($lesser ge $greater)', # 4
+ $yes.' ($lesser ge $greater)', # 5
+ $yes.' ($greater le $lesser )', # 7
+ 'not ($greater ne $lesser )', # 8
+ ' ($greater eq $lesser )', # 9
+ $no.' ($greater ge $lesser )', # 10
+ 'not (($lesser cmp $greater) == -($sign))' # 11
+ );
+ @test{@test} = 0 x @test;
+ $test = 0;
+ for my $ti (@test) {
if ($is_utf8_locale) {
use locale ':not_characters';
- ($yes, $no, $sign) = ($lesser lt $greater
- ? (" ", "not ", 1)
- : ("not ", " ", -1));
+ $test{$ti} = eval $ti;
}
else {
- use locale;
- ($yes, $no, $sign) = ($lesser lt $greater
- ? (" ", "not ", 1)
- : ("not ", " ", -1));
+ # Already in 'use locale';
+ $test{$ti} = eval $ti;
}
- # all these tests should FAIL (return 0).
- # Exact lt or gt cannot be tested because
- # in some locales, say, eacute and E may test equal.
- @test =
- (
- $no.' ($lesser le $greater)', # 1
- 'not ($lesser ne $greater)', # 2
- ' ($lesser eq $greater)', # 3
- $yes.' ($lesser ge $greater)', # 4
- $yes.' ($lesser ge $greater)', # 5
- $yes.' ($greater le $lesser )', # 7
- 'not ($greater ne $lesser )', # 8
- ' ($greater eq $lesser )', # 9
- $no.' ($greater ge $lesser )', # 10
- 'not (($lesser cmp $greater) == -($sign))' # 11
- );
- @test{@test} = 0 x @test;
- $test = 0;
- for my $ti (@test) {
- if ($is_utf8_locale) {
- use locale ':not_characters';
- $test{$ti} = eval $ti;
- }
- else {
- # Already in 'use locale';
- $test{$ti} = eval $ti;
+ $test ||= $test{$ti}
+ }
+ report_result($Locale, $locales_test_number, $test == 0);
+ if ($test) {
+ debug "# lesser = '$lesser'\n";
+ debug "# greater = '$greater'\n";
+ debug "# lesser cmp greater = ",
+ $lesser cmp $greater, "\n";
+ debug "# greater cmp lesser = ",
+ $greater cmp $lesser, "\n";
+ debug "# (greater) from = $from, to = $to\n";
+ for my $ti (@test) {
+ debugf("# %-40s %-4s", $ti,
+ $test{$ti} ? 'FAIL' : 'ok');
+ if ($ti =~ /\(\.*(\$.+ +cmp +\$[^\)]+)\.*\)/) {
+ debugf("(%s == %4d)", $1, eval $1);
}
- $test ||= $test{$ti}
- }
- tryneoalpha($Locale, $locales_test_number, $test == 0);
- if ($test) {
- debug "# lesser = '$lesser'\n";
- debug "# greater = '$greater'\n";
- debug "# lesser cmp greater = ",
- $lesser cmp $greater, "\n";
- debug "# greater cmp lesser = ",
- $greater cmp $lesser, "\n";
- debug "# (greater) from = $from, to = $to\n";
- for my $ti (@test) {
- debugf("# %-40s %-4s", $ti,
- $test{$ti} ? 'FAIL' : 'ok');
- if ($ti =~ /\(\.*(\$.+ +cmp +\$[^\)]+)\.*\)/) {
- debugf("(%s == %4d)", $1, eval $1);
- }
- debug "\n#";
- }
-
- last;
- }
- }
- }
- }
+ debug "\n#";
+ }
- if ($locales_test_number != $final_Neoalpha_test_number) {
- die("The delta for \$final_Neoalpha needs to be updated from "
- . ($final_Neoalpha_test_number - $first_Neoalpha_test_number)
- . " to "
- . ($locales_test_number - $first_Neoalpha_test_number)
- );
+ last;
+ }
+ }
}
my $ok1;
my $ok11;
my $ok12;
my $ok13;
+ my $ok14;
+ my $ok15;
+ my $ok16;
+ my $ok17;
+ my $ok18;
my $c;
my $d;
my $e;
my $f;
my $g;
+ my $h;
+ my $i;
+ my $j;
if (! $is_utf8_locale) {
use locale;
{
no locale;
- # The earlier test was $e = "$x". But this fails [perl
- # #108378], and the "no locale" was commented out. But doing
- # that made all the tests in the block after this one
- # meaningless, as originally it was testing the nesting of a
- # "no locale" scope, and how it recovers after that scope is
- # done. So I (khw) filed a bug report and changed this so it
- # wouldn't fail. It seemed too much work to add TODOs
- # instead. Should this be fixed, the following test names
- # would need to be revised; they mostly don't really test
- # anything currently.
- $e = $x;
+ $e = "$x";
$ok6 = $e == 1.23;
$ok7 = $e == $x;
$f = "1.23";
$g = 2.34;
+ $h = 1.5;
+ $i = 1.25;
+ $j = "$h:$i";
$ok9 = $f == 1.23;
$ok10 = $f == $x;
$ok11 = $f == $c;
$ok12 = abs(($f + $g) - 3.57) < 0.01;
$ok13 = $w == 0;
+ $ok14 = $ok15 = $ok16 = 1; # Skip for non-utf8 locales
+ }
+ {
+ no locale;
+ $ok17 = "1.5:1.25" eq sprintf("%g:%g", $h, $i);
}
+ $ok18 = $j eq sprintf("%g:%g", $h, $i);
}
else {
use locale ':not_characters';
$ok5 = $c == $d;
{
no locale;
- $e = $x;
+ $e = "$x";
$ok6 = $e == 1.23;
$ok7 = $e == $x;
$f = "1.23";
$g = 2.34;
+ $h = 1.5;
+ $i = 1.25;
+ $j = "$h:$i";
$ok9 = $f == 1.23;
$ok10 = $f == $x;
$ok11 = $f == $c;
$ok12 = abs(($f + $g) - 3.57) < 0.01;
$ok13 = $w == 0;
+
+ # Look for non-ASCII error messages, and verify that the first
+ # such is in UTF-8 (the others almost certainly will be like the
+ # first).
+ $ok14 = 1;
+ foreach my $err (keys %!) {
+ use Errno;
+ $! = eval "&Errno::$err"; # Convert to strerror() output
+ my $strerror = "$!";
+ if ("$strerror" =~ /\P{ASCII}/) {
+ $ok14 = utf8::is_utf8($strerror);
+ last;
+ }
+ }
+
+ # Similarly, we verify that a non-ASCII radix is in UTF-8. This
+ # also catches if there is a disparity between sprintf and
+ # stringification.
+
+ my $string_g = "$g";
+ my $sprintf_g = sprintf("%g", $g);
+
+ $ok15 = $string_g =~ / ^ \p{ASCII}+ $ /x || utf8::is_utf8($string_g);
+ $ok16 = $sprintf_g eq $string_g;
+ }
+ {
+ no locale;
+ $ok17 = "1.5:1.25" eq sprintf("%g:%g", $h, $i);
}
+ $ok18 = $j eq sprintf("%g:%g", $h, $i);
}
- tryneoalpha($Locale, ++$locales_test_number, $ok1);
+ report_result($Locale, ++$locales_test_number, $ok1);
$test_names{$locales_test_number} = 'Verify that an intervening printf doesn\'t change assignment results';
my $first_a_test = $locales_test_number;
debug "# $first_a_test..$locales_test_number: \$a = $a, \$b = $b, Locale = $Locale\n";
- tryneoalpha($Locale, ++$locales_test_number, $ok2);
+ report_result($Locale, ++$locales_test_number, $ok2);
$test_names{$locales_test_number} = 'Verify that an intervening sprintf doesn\'t change assignment results';
my $first_c_test = $locales_test_number;
- tryneoalpha($Locale, ++$locales_test_number, $ok3);
+ report_result($Locale, ++$locales_test_number, $ok3);
$test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a constant';
- tryneoalpha($Locale, ++$locales_test_number, $ok4);
+ report_result($Locale, ++$locales_test_number, $ok4);
$test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a scalar';
- tryneoalpha($Locale, ++$locales_test_number, $ok5);
+ report_result($Locale, ++$locales_test_number, $ok5);
$test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a scalar and an intervening sprintf';
debug "# $first_c_test..$locales_test_number: \$c = $c, \$d = $d, Locale = $Locale\n";
- tryneoalpha($Locale, ++$locales_test_number, $ok6);
- $test_names{$locales_test_number} = 'Verify that can assign numerically under inner no-locale block';
+ report_result($Locale, ++$locales_test_number, $ok6);
+ $test_names{$locales_test_number} = 'Verify that can assign stringified under inner no-locale block';
my $first_e_test = $locales_test_number;
- tryneoalpha($Locale, ++$locales_test_number, $ok7);
+ report_result($Locale, ++$locales_test_number, $ok7);
$test_names{$locales_test_number} = 'Verify that "==" with a scalar still works in inner no locale';
- tryneoalpha($Locale, ++$locales_test_number, $ok8);
+ report_result($Locale, ++$locales_test_number, $ok8);
$test_names{$locales_test_number} = 'Verify that "==" with a scalar and an intervening sprintf still works in inner no locale';
debug "# $first_e_test..$locales_test_number: \$e = $e, no locale\n";
- tryneoalpha($Locale, ++$locales_test_number, $ok9);
+ report_result($Locale, ++$locales_test_number, $ok9);
$test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix still works when doing "==" with a constant';
my $first_f_test = $locales_test_number;
- tryneoalpha($Locale, ++$locales_test_number, $ok10);
+ report_result($Locale, ++$locales_test_number, $ok10);
$test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix still works when doing "==" with a scalar';
- tryneoalpha($Locale, ++$locales_test_number, $ok11);
+ report_result($Locale, ++$locales_test_number, $ok11);
$test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix still works when doing "==" with a scalar and an intervening sprintf';
- tryneoalpha($Locale, ++$locales_test_number, $ok12);
+ report_result($Locale, ++$locales_test_number, $ok12);
$test_names{$locales_test_number} = 'Verify that after a no-locale block, a different locale radix can participate in an addition and function call as numeric';
- tryneoalpha($Locale, ++$locales_test_number, $ok13);
+ report_result($Locale, ++$locales_test_number, $ok13);
$test_names{$locales_test_number} = 'Verify that don\'t get warning under "==" even if radix is not a dot';
+ report_result($Locale, ++$locales_test_number, $ok14);
+ $test_names{$locales_test_number} = 'Verify that non-ASCII UTF-8 error messages are in UTF-8';
+
+ report_result($Locale, ++$locales_test_number, $ok15);
+ $test_names{$locales_test_number} = 'Verify that a number with a UTF-8 radix has a UTF-8 stringification';
+
+ report_result($Locale, ++$locales_test_number, $ok16);
+ $test_names{$locales_test_number} = 'Verify that a sprintf of a number with a UTF-8 radix yields UTF-8';
+
+ report_result($Locale, ++$locales_test_number, $ok17);
+ $test_names{$locales_test_number} = 'Verify that a sprintf of a number outside locale scope uses a dot radix';
+
+ report_result($Locale, ++$locales_test_number, $ok18);
+ $test_names{$locales_test_number} = 'Verify that a sprintf of a number back within locale scope uses locale radix';
+
debug "# $first_f_test..$locales_test_number: \$f = $f, \$g = $g, back to locale = $Locale\n";
# Does taking lc separately differ from taking
my $y = "aa";
my $z = "AB";
- tryneoalpha($Locale, ++$locales_test_number,
+ report_result($Locale, ++$locales_test_number,
lcA($x, $y) == 1 && lcB($x, $y) == 1 ||
lcA($x, $z) == 0 && lcB($x, $z) == 0);
}
my $y = "aa";
my $z = "AB";
- tryneoalpha($Locale, ++$locales_test_number,
+ report_result($Locale, ++$locales_test_number,
lcC($x, $y) == 1 && lcD($x, $y) == 1 ||
lcC($x, $z) == 0 && lcD($x, $z) == 0);
}
my @f = ();
++$locales_test_number;
$test_names{$locales_test_number} = 'Verify case insensitive matching works';
- foreach my $x (keys %UPPER) {
+ foreach my $x (sort keys %UPPER) {
if (! $is_utf8_locale) {
my $y = lc $x;
next unless uc $y eq $x;
print "# UPPER $x lc $y ",
- $x =~ /$y/i ? 1 : 0, " ",
- $y =~ /$x/i ? 1 : 0, "\n" if 0;
+ $x =~ /$y/i ? 1 : 0, " ",
+ $y =~ /$x/i ? 1 : 0, "\n" if 0;
#
# If $x and $y contain regular expression characters
# AND THEY lowercase (/i) to regular expression characters,
# With utf8 both will fail since the locale concept
# of upper/lower does not work well in Unicode.
push @f, $x unless $x =~ /$y/i == $y =~ /$x/i;
+
+ # fc is not a locale concept, so Perl uses lc for it.
+ push @f, $x unless lc $x eq fc $x;
}
else {
use locale ':not_characters';
my $y = lc $x;
next unless uc $y eq $x;
print "# UPPER $x lc $y ",
- $x =~ /$y/i ? 1 : 0, " ",
- $y =~ /$x/i ? 1 : 0, "\n" if 0;
+ $x =~ /$y/i ? 1 : 0, " ",
+ $y =~ /$x/i ? 1 : 0, "\n" if 0;
# Here, we can fully test things, unlike plain 'use locale',
# because this form does work well with Unicode
push @f, $x unless $x =~ /$y/i && $y =~ /$x/i;
+
+ # The places where Unicode's lc is different from fc are
+ # skipped here by virtue of the 'next unless uc...' line above
+ push @f, $x unless lc $x eq fc $x;
}
}
- foreach my $x (keys %lower) {
+ foreach my $x (sort keys %lower) {
if (! $is_utf8_locale) {
my $y = uc $x;
next unless lc $y eq $x;
print "# lower $x uc $y ",
- $x =~ /$y/i ? 1 : 0, " ",
- $y =~ /$x/i ? 1 : 0, "\n" if 0;
+ $x =~ /$y/i ? 1 : 0, " ",
+ $y =~ /$x/i ? 1 : 0, "\n" if 0;
if ($x =~ $re || $y =~ $re) { # See above.
print "# Regex characters in '$x' or '$y', skipping test $locales_test_number for locale '$Locale'\n";
next;
# With utf8 both will fail since the locale concept
# of upper/lower does not work well in Unicode.
push @f, $x unless $x =~ /$y/i == $y =~ /$x/i;
+
+ push @f, $x unless lc $x eq fc $x;
}
else {
use locale ':not_characters';
my $y = uc $x;
next unless lc $y eq $x;
print "# lower $x uc $y ",
- $x =~ /$y/i ? 1 : 0, " ",
- $y =~ /$x/i ? 1 : 0, "\n" if 0;
+ $x =~ /$y/i ? 1 : 0, " ",
+ $y =~ /$x/i ? 1 : 0, "\n" if 0;
push @f, $x unless $x =~ /$y/i && $y =~ /$x/i;
+
+ push @f, $x unless lc $x eq fc $x;
}
}
- tryneoalpha($Locale, $locales_test_number, @f == 0);
+ report_multi_result($Locale, $locales_test_number, \@f);
+ }
+
+ # [perl #109318]
+ {
+ my @f = ();
+ ++$locales_test_number;
+ $test_names{$locales_test_number} = 'Verify atof with locale radix and negative exponent';
+
+ my $radix = POSIX::localeconv()->{decimal_point};
+ my @nums = (
+ "3.14e+9", "3${radix}14e+9", "3.14e-9", "3${radix}14e-9",
+ "-3.14e+9", "-3${radix}14e+9", "-3.14e-9", "-3${radix}14e-9",
+ );
+
+ if (! $is_utf8_locale) {
+ use locale;
+ for my $num (@nums) {
+ push @f, $num
+ unless sprintf("%g", $num) =~ /3.+14/;
+ }
+ }
+ else {
+ use locale ':not_characters';
+ for my $num (@nums) {
+ push @f, $num
+ unless sprintf("%g", $num) =~ /3.+14/;
+ }
+ }
+
+ report_result($Locale, $locales_test_number, @f == 0);
if (@f) {
- print "# failed $locales_test_number locale '$Locale' characters @f\n"
+ print "# failed $locales_test_number locale '$Locale' numbers @f\n"
}
}
}
print "# It usually indicates a problem in the environment,\n";
print "# not in Perl itself.\n";
}
+ if ($Okay{$_} && ($_ >= $first_casing_test_number
+ && $_ <= $final_casing_test_number))
+ {
+ # Round to nearest .1%
+ my $percent_fail = (int(.5 + (1000 * scalar(keys $Problem{$_})
+ / scalar(@Locale))))
+ / 10;
+ if (! $debug && $percent_fail < $acceptable_fold_failure_percentage)
+ {
+ $test_names{$_} .= 'TODO';
+ print "# ", 100 - $percent_fail, "% of locales pass the following test, so it is likely that the failures\n";
+ print "# are errors in the locale definitions. The test is marked TODO, as the\n";
+ print "# problem is not likely to be Perl's\n";
+ }
+ }
+ print "#\n";
+ if ($debug) {
+ print "# The code points that had this failure are given above. Look for lines\n";
+ print "# that match 'failed $_'\n";
+ }
+ else {
+ print "# For more details, rerun, with environment variable PERL_DEBUG_FULL_TEST=1.\n";
+ print "# Then look at that output for lines that match 'failed $_'\n";
+ }
print "not ";
}
print "ok $_";
- print " $test_names{$_}" if defined $test_names{$_};
+ if (defined $test_names{$_}) {
+ # If TODO is in the test name, make it thus
+ my $todo = $test_names{$_} =~ s/TODO\s*//;
+ print " $test_names{$_}";
+ print " # TODO" if $todo;
+ }
print "\n";
}
+$test_num = $final_locales_test_number;
+
+unless ( $^O eq 'dragonfly' ) {
+ # perl #115808
+ use warnings;
+ my $warned = 0;
+ local $SIG{__WARN__} = sub {
+ $warned = $_[0] =~ /uninitialized/;
+ };
+ my $z = "y" . setlocale(&POSIX::LC_ALL, "xyzzy");
+ ok($warned, "variable set to setlocale(BAD LOCALE) is considered uninitialized");
+}
+
+# Test that tainting and case changing works on utf8 strings. These tests are
+# placed last to avoid disturbing the hard-coded test numbers that existed at
+# the time these were added above this in this file.
+# This also tests that locale overrides unicode_strings in the same scope for
+# non-utf8 strings.
+setlocale(&POSIX::LC_ALL, "C");
+{
+ use locale;
+ use feature 'unicode_strings';
+
+ foreach my $function ("uc", "ucfirst", "lc", "lcfirst", "fc") {
+ my @list; # List of code points to test for $function
+
+ # Used to calculate the changed case for ASCII characters by using the
+ # ord, instead of using one of the functions under test.
+ my $ascii_case_change_delta;
+ my $above_latin1_case_change_delta; # Same for the specific ords > 255
+ # that we use
+
+ # We test an ASCII character, which should change case and be tainted;
+ # a Latin1 character, which shouldn't change case under this C locale,
+ # and is tainted.
+ # an above-Latin1 character that when the case is changed would cross
+ # the 255/256 boundary, so doesn't change case and isn't tainted
+ # (the \x{149} is one of these, but changes into 2 characters, the
+ # first one of which doesn't cross the boundary.
+ # the final one in each list is an above-Latin1 character whose case
+ # does change, and shouldn't be tainted. The code below uses its
+ # position in its list as a marker to indicate that it, unlike the
+ # other code points above ASCII, has a successful case change
+ if ($function =~ /^u/) {
+ @list = ("", "a", "\xe0", "\xff", "\x{fb00}", "\x{149}", "\x{101}");
+ $ascii_case_change_delta = -32;
+ $above_latin1_case_change_delta = -1;
+ }
+ else {
+ @list = ("", "A", "\xC0", "\x{17F}", "\x{100}");
+ $ascii_case_change_delta = +32;
+ $above_latin1_case_change_delta = +1;
+ }
+ foreach my $is_utf8_locale (0 .. 1) {
+ foreach my $j (0 .. $#list) {
+ my $char = $list[$j];
+
+ for my $encoded_in_utf8 (0 .. 1) {
+ my $should_be;
+ my $changed;
+ if (! $is_utf8_locale) {
+ $should_be = ($j == $#list)
+ ? chr(ord($char) + $above_latin1_case_change_delta)
+ : (length $char == 0 || ord($char) > 127)
+ ? $char
+ : chr(ord($char) + $ascii_case_change_delta);
+
+ # This monstrosity is in order to avoid using an eval,
+ # which might perturb the results
+ $changed = ($function eq "uc")
+ ? uc($char)
+ : ($function eq "ucfirst")
+ ? ucfirst($char)
+ : ($function eq "lc")
+ ? lc($char)
+ : ($function eq "lcfirst")
+ ? lcfirst($char)
+ : ($function eq "fc")
+ ? fc($char)
+ : die("Unexpected function \"$function\"");
+ }
+ else {
+ {
+ no locale;
+
+ # For utf8-locales the case changing functions
+ # should work just like they do outside of locale.
+ # Can use eval here because not testing it when
+ # not in locale.
+ $should_be = eval "$function('$char')";
+ die "Unexpected eval error $@ from 'eval \"$function('$char')\"'" if $@;
+
+ }
+ use locale ':not_characters';
+ $changed = ($function eq "uc")
+ ? uc($char)
+ : ($function eq "ucfirst")
+ ? ucfirst($char)
+ : ($function eq "lc")
+ ? lc($char)
+ : ($function eq "lcfirst")
+ ? lcfirst($char)
+ : ($function eq "fc")
+ ? fc($char)
+ : die("Unexpected function \"$function\"");
+ }
+ ok($changed eq $should_be,
+ "$function(\"$char\") in C locale "
+ . (($is_utf8_locale)
+ ? "(use locale ':not_characters'"
+ : "(use locale")
+ . (($encoded_in_utf8)
+ ? "; encoded in utf8)"
+ : "; not encoded in utf8)")
+ . " should be \"$should_be\", got \"$changed\"");
+
+ # Tainting shouldn't happen for utf8 locales, empty
+ # strings, or those characters above 255.
+ (! $is_utf8_locale && length($char) > 0 && ord($char) < 256)
+ ? check_taint($changed)
+ : check_taint_not($changed);
+
+ # Use UTF-8 next time through the loop
+ utf8::upgrade($char);
+ }
+ }
+ }
+ }
+}
+
# Give final advice.
my $didwarn = 0;
"# The locale ", (@f == 1 ? "definition" : "definitions"), "\n#\n",
"#\t", $f, "\n#\n",
"# on your system may have errors because the locale test $_\n",
+ "# \"$test_names{$_}\"\n",
"# failed in ", (@f == 1 ? "that locale" : "those locales"),
".\n";
print <<EOW;
$s =~ s/(.{50,60}) /$1\n#\t/g;
warn
- "# The following locales\n#\n",
+ "# The following locales\n#\n",
"#\t", $s, "\n#\n",
"# tested okay.\n#\n",
} else {
"# The following locales\n#\n",
"#\t", $F, "\n#\n",
"# had problems.\n#\n",
+ "# For more details, rerun, with environment variable PERL_DEBUG_FULL_TEST=1.\n";
} else {
warn "# None of your locales were broken.\n";
}
}
-$test_num = $final_locales_test_number;
-
-# Test that tainting and case changing works on utf8 strings. These tests are
-# placed last to avoid disturbing the hard-coded test numbers that existed at
-# the time these were added above this in this file.
-setlocale(LC_ALL, "C");
-{
- use locale;
-
- foreach my $function ("uc", "ucfirst", "lc", "lcfirst") {
- my @list; # List of code points to test for $function
-
- # Used to calculate the changed case for ASCII characters by using the
- # ord, instead of using one of the functions under test.
- my $ascii_case_change_delta;
- my $above_latin1_case_change_delta; # Same for the specific ords > 255
- # that we use
-
- # We test an ASCII character, which should change case and be tainted;
- # a Latin1 character, which shouldn't change case under this C locale,
- # and is tainted.
- # an above-Latin1 character that when the case is changed would cross
- # the 255/256 boundary, so doesn't change case and isn't tainted
- # (the \x{149} is one of these, but changes into 2 characters, the
- # first one of which doesn't cross the boundary.
- # the final one in each list is an above-Latin1 character whose case
- # does change, and shouldn't be tainted. The code below uses its
- # position in its list as a marker to indicate that it, unlike the
- # other code points above ASCII, has a successful case change
- if ($function =~ /^u/) {
- @list = ("", "a", "\xe0", "\xff", "\x{fb00}", "\x{149}", "\x{101}");
- $ascii_case_change_delta = -32;
- $above_latin1_case_change_delta = -1;
- }
- else {
- @list = ("", "A", "\xC0", "\x{1E9E}", "\x{100}");
- $ascii_case_change_delta = +32;
- $above_latin1_case_change_delta = +1;
- }
- foreach my $is_utf8_locale (0 .. 1) {
- foreach my $j (0 .. $#list) {
- my $char = $list[$j];
- utf8::upgrade($char);
- my $should_be;
- my $changed;
- if (! $is_utf8_locale) {
- $should_be = ($j == $#list)
- ? chr(ord($char) + $above_latin1_case_change_delta)
- : (length $char == 0 || ord($char) > 127)
- ? $char
- : chr(ord($char) + $ascii_case_change_delta);
-
- # This monstrosity is in order to avoid using an eval, which
- # might perturb the results
- $changed = ($function eq "uc")
- ? uc($char)
- : ($function eq "ucfirst")
- ? ucfirst($char)
- : ($function eq "lc")
- ? lc($char)
- : ($function eq "lcfirst")
- ? lcfirst($char)
- : die("Unexpected function \"$function\"");
- }
- else {
- {
- no locale;
-
- # For utf8-locales the case changing functions should
- # work just like they do outside of locale. Can use
- # eval here because not testing it when not in locale.
- $should_be = eval "$function('$char')";
- die "Unexpected eval error $@ from 'eval \"$function('$char')\"'" if $@;
-
- }
- use locale ':not_characters';
- $changed = ($function eq "uc")
- ? uc($char)
- : ($function eq "ucfirst")
- ? ucfirst($char)
- : ($function eq "lc")
- ? lc($char)
- : ($function eq "lcfirst")
- ? lcfirst($char)
- : die("Unexpected function \"$function\"");
- }
- ok($changed eq $should_be, "$function(\"$char\") in C locale "
- . (($is_utf8_locale)
- ? "(use locale ':not_characters')"
- : "(use locale)")
- . " should be \"$should_be\", got \"$changed\"");
-
- # Tainting shouldn't happen for utf8 locales, empty strings,
- # or those characters above 255.
- (! $is_utf8_locale && length($char) > 0 && ord($char) < 256)
- ? check_taint($changed)
- : check_taint_not($changed);
- }
- }
- }
-}
-
print "1..$test_num\n";
# eof
https://perl5.git.perl.org / perl5.git / blobdiff