#!./perl -wT
+# This tests plain 'use locale' and adorned 'use locale ":not_characters"'
+# Because these pragmas are compile time, and I (khw) am trying to test
+# 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';
+
BEGIN {
chdir 't' if -d 't';
@INC = '../lib';
}
use strict;
+use feature 'fc';
-my $debug = 1;
+my $debug = $ENV{PERL_DEBUG_FULL_TEST} // 0;
+
+# 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.
+my $acceptable_fold_failure_percentage = 5;
use Dumpvalue;
$have_setlocale = 0 if ((($^O eq 'MSWin32' && !$winxp) || $^O eq 'NetWare') &&
$Config{cc} =~ /^(cl|gcc)/i);
-# UWIN seems to loop after test 98, just skip for now
+# UWIN seems to loop after taint tests, just skip for now
$have_setlocale = 0 if ($^O =~ /^uwin/);
-my $last = $have_setlocale ? &last : &last_without_setlocale;
-
-print "1..$last\n";
-
-sub LC_ALL ();
-
$a = 'abc %';
+my $test_num = 0;
+
sub ok {
- my ($n, $result) = @_;
+ my ($result, $message) = @_;
+ $message = "" unless defined $message;
print 'not ' unless ($result);
- print "ok $n\n";
+ print "ok " . ++$test_num;
+ print " $message";
+ print "\n";
}
# First we'll do a lot of taint checking for locales.
sub is_tainted { # hello, camel two.
no warnings 'uninitialized' ;
my $dummy;
+ local $@;
not eval { $dummy = join("", @_), kill 0; 1 }
}
-sub check_taint ($$) {
- ok $_[0], is_tainted($_[1]);
+sub check_taint ($;$) {
+ my $message_tail = $_[1] // "";
+ $message_tail = ": $message_tail" if $message_tail;
+ ok is_tainted($_[0]), "verify that is tainted$message_tail";
}
-sub check_taint_not ($$) {
- ok $_[0], not is_tainted($_[1]);
+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");
}
-use locale; # engage locale and therefore locale taint.
-
-check_taint_not 1, $a;
+"\tb\t" =~ /^m?(\s)(.*)\1$/;
+check_taint_not $&, "not tainted outside 'use locale'";
+;
-check_taint 2, uc($a);
-check_taint 3, "\U$a";
-check_taint 4, ucfirst($a);
-check_taint 5, "\u$a";
-check_taint 6, lc($a);
-check_taint 7, "\L$a";
-check_taint 8, lcfirst($a);
-check_taint 9, "\l$a";
+use locale; # engage locale and therefore locale taint.
-check_taint_not 10, sprintf('%e', 123.456);
-check_taint_not 11, sprintf('%f', 123.456);
-check_taint_not 12, sprintf('%g', 123.456);
-check_taint_not 13, sprintf('%d', 123.456);
-check_taint_not 14, sprintf('%x', 123.456);
+check_taint_not $a;
+
+check_taint uc($a);
+check_taint "\U$a";
+check_taint ucfirst($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 sprintf('%e', 123.456);
+check_taint_not sprintf('%f', 123.456);
+check_taint_not sprintf('%g', 123.456);
+check_taint_not sprintf('%d', 123.456);
+check_taint_not sprintf('%x', 123.456);
$_ = $a; # untaint $_
$_ = uc($a); # taint $_
-check_taint 15, $_;
+check_taint $_;
/(\w)/; # taint $&, $`, $', $+, $1.
-check_taint 16, $&;
-check_taint 17, $`;
-check_taint 18, $';
-check_taint 19, $+;
-check_taint 20, $1;
-check_taint_not 21, $2;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
/(.)/; # untaint $&, $`, $', $+, $1.
-check_taint_not 22, $&;
-check_taint_not 23, $`;
-check_taint_not 24, $';
-check_taint_not 25, $+;
-check_taint_not 26, $1;
-check_taint_not 27, $2;
+check_taint_not $&;
+check_taint_not $`;
+check_taint_not $';
+check_taint_not $+;
+check_taint_not $1;
+check_taint_not $2;
/(\W)/; # taint $&, $`, $', $+, $1.
-check_taint 28, $&;
-check_taint 29, $`;
-check_taint 30, $';
-check_taint 31, $+;
-check_taint 32, $1;
-check_taint_not 33, $2;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
/(\s)/; # taint $&, $`, $', $+, $1.
-check_taint 34, $&;
-check_taint 35, $`;
-check_taint 36, $';
-check_taint 37, $+;
-check_taint 38, $1;
-check_taint_not 39, $2;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
/(\S)/; # taint $&, $`, $', $+, $1.
-check_taint 40, $&;
-check_taint 41, $`;
-check_taint 42, $';
-check_taint 43, $+;
-check_taint 44, $1;
-check_taint_not 45, $2;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
$_ = $a; # untaint $_
-check_taint_not 46, $_;
+check_taint_not $_;
/(b)/; # this must not taint
-check_taint_not 47, $&;
-check_taint_not 48, $`;
-check_taint_not 49, $';
-check_taint_not 50, $+;
-check_taint_not 51, $1;
-check_taint_not 52, $2;
+check_taint_not $&;
+check_taint_not $`;
+check_taint_not $';
+check_taint_not $+;
+check_taint_not $1;
+check_taint_not $2;
$_ = $a; # untaint $_
-check_taint_not 53, $_;
+check_taint_not $_;
$b = uc($a); # taint $b
s/(.+)/$b/; # this must taint only the $_
-check_taint 54, $_;
-check_taint_not 55, $&;
-check_taint_not 56, $`;
-check_taint_not 57, $';
-check_taint_not 58, $+;
-check_taint_not 59, $1;
-check_taint_not 60, $2;
+check_taint $_;
+check_taint_not $&;
+check_taint_not $`;
+check_taint_not $';
+check_taint_not $+;
+check_taint_not $1;
+check_taint_not $2;
$_ = $a; # untaint $_
s/(.+)/b/; # this must not taint
-check_taint_not 61, $_;
-check_taint_not 62, $&;
-check_taint_not 63, $`;
-check_taint_not 64, $';
-check_taint_not 65, $+;
-check_taint_not 66, $1;
-check_taint_not 67, $2;
+check_taint_not $_;
+check_taint_not $&;
+check_taint_not $`;
+check_taint_not $';
+check_taint_not $+;
+check_taint_not $1;
+check_taint_not $2;
$b = $a; # untaint $b
($b = $a) =~ s/\w/$&/;
-check_taint 68, $b; # $b should be tainted.
-check_taint_not 69, $a; # $a should be not.
+check_taint $b; # $b should be tainted.
+check_taint_not $a; # $a should be not.
$_ = $a; # untaint $_
s/(\w)/\l$1/; # this must taint
-check_taint 70, $_;
-check_taint 71, $&;
-check_taint 72, $`;
-check_taint 73, $';
-check_taint 74, $+;
-check_taint 75, $1;
-check_taint_not 76, $2;
+check_taint $_;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
$_ = $a; # untaint $_
s/(\w)/\L$1/; # this must taint
-check_taint 77, $_;
-check_taint 78, $&;
-check_taint 79, $`;
-check_taint 80, $';
-check_taint 81, $+;
-check_taint 82, $1;
-check_taint_not 83, $2;
+check_taint $_;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
$_ = $a; # untaint $_
s/(\w)/\u$1/; # this must taint
-check_taint 84, $_;
-check_taint 85, $&;
-check_taint 86, $`;
-check_taint 87, $';
-check_taint 88, $+;
-check_taint 89, $1;
-check_taint_not 90, $2;
+check_taint $_;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
$_ = $a; # untaint $_
s/(\w)/\U$1/; # this must taint
-check_taint 91, $_;
-check_taint 92, $&;
-check_taint 93, $`;
-check_taint 94, $';
-check_taint 95, $+;
-check_taint 96, $1;
-check_taint_not 97, $2;
+check_taint $_;
+check_taint $&;
+check_taint $`;
+check_taint $';
+check_taint $+;
+check_taint $1;
+check_taint_not $2;
# After all this tainting $a should be cool.
-check_taint_not 98, $a;
+check_taint_not $a;
+
+{ # This is just the previous tests copied here with a different
+ # compile-time pragma.
+
+ use locale ':not_characters'; # engage restricted locale with different
+ # tainting rules
+
+ check_taint_not $a;
+
+ check_taint_not uc($a);
+ check_taint_not "\U$a";
+ check_taint_not ucfirst($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";
+
+ check_taint_not sprintf('%e', 123.456);
+ check_taint_not sprintf('%f', 123.456);
+ check_taint_not sprintf('%g', 123.456);
+ check_taint_not sprintf('%d', 123.456);
+ check_taint_not sprintf('%x', 123.456);
+
+ $_ = $a; # untaint $_
+
+ $_ = uc($a); # taint $_
+
+ check_taint_not $_;
+
+ /(\w)/; # taint $&, $`, $', $+, $1.
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ /(.)/; # untaint $&, $`, $', $+, $1.
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ /(\W)/; # taint $&, $`, $', $+, $1.
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ /(\s)/; # taint $&, $`, $', $+, $1.
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ /(\S)/; # taint $&, $`, $', $+, $1.
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ $_ = $a; # untaint $_
+
+ check_taint_not $_;
+
+ /(b)/; # this must not taint
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ $_ = $a; # untaint $_
+
+ check_taint_not $_;
+
+ $b = uc($a); # taint $b
+ s/(.+)/$b/; # this must taint only the $_
+
+ check_taint_not $_;
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ $_ = $a; # untaint $_
+
+ s/(.+)/b/; # this must not taint
+ check_taint_not $_;
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ $b = $a; # untaint $b
+
+ ($b = $a) =~ s/\w/$&/;
+ check_taint_not $b; # $b should be tainted.
+ check_taint_not $a; # $a should be not.
+
+ $_ = $a; # untaint $_
+
+ s/(\w)/\l$1/; # this must taint
+ check_taint_not $_;
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ $_ = $a; # untaint $_
+
+ s/(\w)/\L$1/; # this must taint
+ check_taint_not $_;
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ $_ = $a; # untaint $_
+
+ s/(\w)/\u$1/; # this must taint
+ check_taint_not $_;
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ $_ = $a; # untaint $_
+
+ s/(\w)/\U$1/; # this must taint
+ check_taint_not $_;
+ check_taint_not $&;
+ check_taint_not $`;
+ check_taint_not $';
+ check_taint_not $+;
+ check_taint_not $1;
+ check_taint_not $2;
+
+ # After all this tainting $a should be cool.
+
+ check_taint_not $a;
+}
-sub last_without_setlocale { 98 }
+# Here are in scope of 'use locale'
# I think we've seen quite enough of taint.
# Let us do some *real* locale work now,
# unless setlocale() is missing (i.e. minitest).
-exit unless $have_setlocale;
+unless ($have_setlocale) {
+ print "1..$test_num\n";
+ exit;
+}
+
+# The test number before our first setlocale()
+my $final_without_setlocale = $test_num;
# Find locales.
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 @Word_;
+my @Digit_;
+my @Space_;
+my @Alpha_;
my @Alnum_;
-
-my @utf8locale;
-my %utf8skip;
-
-sub getalnum_ {
- sort grep /\w/, map { chr } 0..255
-}
+my @Ascii_;
+my @Blank_;
+my @Cntrl_;
+my @Graph_;
+my @Lower_;
+my @Print_;
+my @Upper_;
+my @Xdigit_;
+my @Cased_;
sub trylocale {
my $locale = shift;
- if (setlocale(LC_ALL, $locale)) {
- push @Locale, $locale;
+ return if grep { $locale eq $_ } @Locale;
+ return unless setlocale(&POSIX::LC_ALL, $locale);
+ my $badutf8;
+ {
+ local $SIG{__WARN__} = sub {
+ $badutf8 = $_[0] =~ /Malformed UTF-8/;
+ };
+ $Locale =~ /UTF-?8/i;
}
+
+ if ($badutf8) {
+ ok(0, "Locale name contains malformed utf8");
+ return;
+ }
+ push @Locale, $locale;
}
sub decode_encodings {
}
close(LOCALES);
} elsif ($^O eq 'VMS' && defined($ENV{'SYS$I18N_LOCALE'}) && -d 'SYS$I18N_LOCALE') {
-# The SYS$I18N_LOCALE logical name search list was not present on
+# The SYS$I18N_LOCALE logical name search list was not present on
# VAX VMS V5.5-12, but was on AXP && VAX VMS V6.2 as well as later versions.
opendir(LOCALES, "SYS\$I18N_LOCALE:");
while ($_ = readdir(LOCALES)) {
}
}
-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;
-my %Neoalpha;
+my @Added_alpha; # Alphas that aren't in the C locale.
+my %test_names;
+
+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 $hex = "";
+ 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]/) {
+ $class = 4;
+ }
+ elsif ($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 {
+ $hex .= sprintf(" %02X", $range_start);
+ $hex .= 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/^ //;
+ $hex =~ s/^ // if ! length $output;
+ return "$output$hex";
+}
-sub tryneoalpha {
- my ($Locale, $i, $test) = @_;
- unless ($test) {
+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 "# Locale = $Locale\n";
- @Alnum_ = getalnum_();
- debug "# w = ", join("",@Alnum_), "\n";
- unless (setlocale(LC_ALL, $Locale)) {
- foreach (99..103) {
- $Problem{$_}{$Locale} = -1;
- }
+ unless (setlocale(&POSIX::LC_ALL, $Locale)) {
+ $setlocale_failed{$Locale} = $Locale;
next;
}
- # Sieve the uppercase and the lowercase.
-
- my %UPPER = ();
- my %lower = ();
- my %BoThCaSe = ();
- for (@Alnum_) {
- if (/[^\d_]/) { # skip digits and the _
- if (uc($_) eq $_) {
- $UPPER{$_} = $_;
- }
- if (lc($_) eq $_) {
- $lower{$_} = $_;
- }
- }
+ # We test UTF-8 locales only under ':not_characters'; otherwise they have
+ # 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 = (); # 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 (! $is_utf8_locale) {
+ use locale;
+ @Word_ = grep /\w/, map { chr } 0..255;
+ @Digit_ = grep /\d/, map { chr } 0..255;
+ @Space_ = grep /\s/, map { chr } 0..255;
+ @Alpha_ = grep /[[:alpha:]]/, map {chr } 0..255;
+ @Alnum_ = grep /[[:alnum:]]/, map {chr } 0..255;
+ @Ascii_ = grep /[[:ascii:]]/, map {chr } 0..255;
+ @Blank_ = grep /[[:blank:]]/, map {chr } 0..255;
+ @Cntrl_ = grep /[[:cntrl:]]/, map {chr } 0..255;
+ @Graph_ = grep /[[:graph:]]/, map {chr } 0..255;
+ @Lower_ = grep /[[:lower:]]/, map {chr } 0..255;
+ @Print_ = grep /[[:print:]]/, map {chr } 0..255;
+ @Upper_ = grep /[[:upper:]]/, map {chr } 0..255;
+ @Xdigit_ = grep /[[:xdigit:]]/, map {chr } 0..255;
+ @Cased_ = grep /[[:upper:]]/i, map {chr } 0..255;
+
+ # Sieve the uppercase and the lowercase.
+
+ for (@Word_) {
+ if (/[^\d_]/) { # skip digits and the _
+ if (uc($_) eq $_) {
+ $UPPER{$_} = $_;
+ }
+ if (lc($_) eq $_) {
+ $lower{$_} = $_;
+ }
+ }
+ }
+ }
+ else {
+ use locale ':not_characters';
+ @Word_ = grep /\w/, map { chr } 0..255;
+ @Digit_ = grep /\d/, map { chr } 0..255;
+ @Space_ = grep /\s/, map { chr } 0..255;
+ @Alpha_ = grep /[[:alpha:]]/, map {chr } 0..255;
+ @Alnum_ = grep /[[:alnum:]]/, map {chr } 0..255;
+ @Ascii_ = grep /[[:ascii:]]/, map {chr } 0..255;
+ @Blank_ = grep /[[:blank:]]/, map {chr } 0..255;
+ @Cntrl_ = grep /[[:cntrl:]]/, map {chr } 0..255;
+ @Graph_ = grep /[[:graph:]]/, map {chr } 0..255;
+ @Lower_ = grep /[[:lower:]]/, map {chr } 0..255;
+ @Print_ = grep /[[:print:]]/, map {chr } 0..255;
+ @Upper_ = grep /[[:upper:]]/, map {chr } 0..255;
+ @Xdigit_ = grep /[[:xdigit:]]/, map {chr } 0..255;
+ @Cased_ = grep /[[:upper:]]/i, map {chr } 0..255;
+ for (@Word_) {
+ if (/[^\d_]/) { # skip digits and the _
+ if (uc($_) eq $_) {
+ $UPPER{$_} = $_;
+ }
+ if (lc($_) eq $_) {
+ $lower{$_} = $_;
+ }
+ }
+ }
}
+
+ debug "# :upper: = ", display_characters(@Upper_), "\n";
+ debug "# :lower: = ", display_characters(@Lower_), "\n";
+ debug "# :cased: = ", display_characters(@Cased_), "\n";
+ debug "# :alpha: = ", display_characters(@Alpha_), "\n";
+ debug "# :alnum: = ", display_characters(@Alnum_), "\n";
+ debug "# w = ", display_characters(@Word_), "\n";
+ debug "# :graph: = ", display_characters(@Graph_), "\n";
+ debug "# :print: = ", display_characters(@Print_), "\n";
+ debug "# d = ", display_characters(@Digit_), "\n";
+ debug "# :xdigit: = ", display_characters(@Xdigit_), "\n";
+ debug "# :blank: = ", display_characters(@Blank_), "\n";
+ debug "# s = ", display_characters(@Space_), "\n";
+ debug "# :cntrl: = ", display_characters(@Cntrl_), "\n";
+ debug "# :ascii: = ", display_characters(@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";
+ debug "# UPPER = ", display_characters(keys %UPPER), "\n";
+ debug "# lower = ", display_characters(keys %lower), "\n";
+ debug "# BoThCaSe = ", display_characters(keys %BoThCaSe), "\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);
- # Find the alphabets that are not alphabets in the default locale.
+ $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/);
- $Neoalpha{$_} = $_;
+
+ @Added_alpha = ();
+ for (keys %UPPER, keys %lower, keys %BoThCaSe) {
+ push(@Added_alpha, $_) if (/\W/);
}
}
- @Neoalpha = sort @Neoalpha;
-
- debug "# Neoalpha = ", join("",@Neoalpha), "\n";
+ @Added_alpha = sort @Added_alpha;
- if (@Neoalpha == 0) {
- # If we have no Neoalphas the remaining tests are no-ops.
- debug "# no Neoalpha, skipping tests 99..102 for locale '$Locale'\n";
- foreach (99..102) {
- push @{$Okay{$_}}, $Locale;
- }
- } else {
+ debug "# Added_alpha = ", display_characters(@Added_alpha), "\n";
- # Test \w.
-
- my $word = join('', @Neoalpha);
+ # Cross-check the whole 8-bit character set.
- my $badutf8;
- {
- local $SIG{__WARN__} = sub {
- $badutf8 = $_[0] =~ /Malformed UTF-8/;
- };
- $Locale =~ /utf-?8/i;
- }
+ ++$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, $locales_test_number, \@f);
+
+ ++$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;
+ 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';
+ push @f, $_ unless /[[:space:]]/ == /\s/;
+ }
+ else {
+ push @f, $_ unless /[[:space:]]/ == /\s/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
+
+ ++$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, $locales_test_number, \@f);
+
+ # The rules for the relationships are given in:
+ # http://www.opengroup.org/onlinepubs/009695399/basedefs/xbd_chap07.html
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:lower:] 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:] 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:] 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, $locales_test_number, @Digit_ == 10 || @Digit_ ==20);
+
+ ++$locales_test_number;
+ undef @f;
+ $test_names{$locales_test_number} = 'Verify that [:digit:] (if is 10 code points) is a subset of [:xdigit:]';
+ if (@Digit_ == 10) {
+ 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 [:alnum:] is a subset of [:graph:]';
+ for (map { chr } 0..255) {
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ push @f, $_ if /[[:alnum:]]/ and ! /[[:graph:]]/;
+ }
+ else {
+ push @f, $_ if /[[:alnum:]]/ and ! /[[:graph:]]/;
+ }
+ }
+ report_multi_result($Locale, $locales_test_number, \@f);
- if ($badutf8) {
- debug "# Locale name contains bad UTF-8, skipping test 99 for locale '$Locale'\n";
- } elsif ($Locale =~ /utf-?8/i) {
- debug "# unknown whether locale and Unicode have the same \\w, skipping test 99 for locale '$Locale'\n";
- push @{$Okay{99}}, $Locale;
- } else {
- if ($word =~ /^(\w+)$/) {
- tryneoalpha($Locale, 99, 1);
- } else {
- tryneoalpha($Locale, 99, 0);
- }
- }
+ # Note that xdigit doesn't have to be a subset of alnum
- # Cross-check the whole 8-bit character set.
+ ++$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, $locales_test_number, \@f);
+
+ ++$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 [: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 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 [: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, $locales_test_number, \@f);
+
+ ++$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, $locales_test_number, \@f);
- for (map { chr } 0..255) {
- tryneoalpha($Locale, 100,
- (/\w/ xor /\W/) ||
- (/\d/ xor /\D/) ||
- (/\s/ xor /\S/));
- }
+ $final_casing_test_number = $locales_test_number;
- # Test for read-only scalars' locale vs non-locale comparisons.
+ # Test for read-only scalars' locale vs non-locale comparisons.
- {
- no locale;
- $a = "qwerty";
- {
- use locale;
- tryneoalpha($Locale, 101, ($a cmp "qwerty") == 0);
- }
- }
+ {
+ 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);
-
- 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..$to]);
- ($yes, $no, $sign) = ($lesser lt $greater
- ? (" ", "not ", 1)
- : ("not ", " ", -1));
- # 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) {
- $test{$ti} = eval $ti;
- $test ||= $test{$ti}
- }
- tryneoalpha($Locale, 102, $test == 0);
- if ($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 $1);
- }
- debug "\n#";
- }
-
- last;
- }
- }
- }
+ {
+ my ($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.
+ $from = int(($_*@Word_)/10);
+ $to = $from + int(@Word_/10);
+ $to = $#Word_ if ($to > $#Word_);
+ $lesser = join('', @Word_[$from..$to]);
+ # Select a slice one character on.
+ $from++; $to++;
+ $to = $#Word_ if ($to > $#Word_);
+ $greater = join('', @Word_[$from..$to]);
+ if ($is_utf8_locale) {
+ use locale ':not_characters';
+ ($yes, $no, $sign) = ($lesser lt $greater
+ ? (" ", "not ", 1)
+ : ("not ", " ", -1));
+ }
+ else {
+ use locale;
+ ($yes, $no, $sign) = ($lesser lt $greater
+ ? (" ", "not ", 1)
+ : ("not ", " ", -1));
+ }
+ # 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);
+ }
+ debug "\n#";
+ }
+
+ last;
+ }
+ }
}
- use locale;
+ my $ok1;
+ my $ok2;
+ my $ok3;
+ my $ok4;
+ my $ok5;
+ my $ok6;
+ my $ok7;
+ my $ok8;
+ my $ok9;
+ my $ok10;
+ my $ok11;
+ my $ok12;
+ my $ok13;
+ my $ok14;
+ my $ok15;
+ my $ok16;
+
+ my $c;
+ my $d;
+ my $e;
+ my $f;
+ my $g;
+
+ if (! $is_utf8_locale) {
+ use locale;
- my ($x, $y) = (1.23, 1.23);
+ my ($x, $y) = (1.23, 1.23);
+
+ $a = "$x";
+ printf ''; # printf used to reset locale to "C"
+ $b = "$y";
+ $ok1 = $a eq $b;
+
+ $c = "$x";
+ my $z = sprintf ''; # sprintf used to reset locale to "C"
+ $d = "$y";
+ $ok2 = $c eq $d;
+ {
+
+ use warnings;
+ my $w = 0;
+ local $SIG{__WARN__} =
+ sub {
+ print "# @_\n";
+ $w++;
+ };
+
+ # The == (among other ops) used to warn for locales
+ # that had something else than "." as the radix character.
+
+ $ok3 = $c == 1.23;
+ $ok4 = $c == $x;
+ $ok5 = $c == $d;
+ {
+ no locale;
+
+ $e = "$x";
+
+ $ok6 = $e == 1.23;
+ $ok7 = $e == $x;
+ $ok8 = $e == $c;
+ }
+
+ $f = "1.23";
+ $g = 2.34;
+
+ $ok9 = $f == 1.23;
+ $ok10 = $f == $x;
+ $ok11 = $f == $c;
+ $ok12 = abs(($f + $g) - 3.57) < 0.01;
+ $ok13 = $w == 0;
+ $ok14 = $ok15 = $ok16 = 1; # Skip for non-utf8 locales
+ }
+ }
+ else {
+ use locale ':not_characters';
+
+ my ($x, $y) = (1.23, 1.23);
+ $a = "$x";
+ printf ''; # printf used to reset locale to "C"
+ $b = "$y";
+ $ok1 = $a eq $b;
+
+ $c = "$x";
+ my $z = sprintf ''; # sprintf used to reset locale to "C"
+ $d = "$y";
+ $ok2 = $c eq $d;
+ {
+ use warnings;
+ my $w = 0;
+ local $SIG{__WARN__} =
+ sub {
+ print "# @_\n";
+ $w++;
+ };
+ $ok3 = $c == 1.23;
+ $ok4 = $c == $x;
+ $ok5 = $c == $d;
+ {
+ no locale;
+ $e = "$x";
+
+ $ok6 = $e == 1.23;
+ $ok7 = $e == $x;
+ $ok8 = $e == $c;
+ }
+
+ $f = "1.23";
+ $g = 2.34;
+
+ $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}/) {
+ my $utf8_strerror = $strerror;
+ utf8::upgrade($utf8_strerror);
+
+ # If $! was already in UTF-8, the upgrade was a no-op;
+ # otherwise they will be different byte strings.
+ use bytes;
+ $ok14 = $utf8_strerror eq $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 $utf8_string_g = "$g";
+ utf8::upgrade($utf8_string_g);
+
+ my $utf8_sprintf_g = sprintf("%g", $g);
+ utf8::upgrade($utf8_sprintf_g);
+ use bytes;
+ $ok15 = $utf8_string_g eq $string_g;
+ $ok16 = $utf8_sprintf_g eq $string_g;
+ }
+ }
- $a = "$x";
- printf ''; # printf used to reset locale to "C"
- $b = "$y";
+ 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 "# 103..107: a = $a, b = $b, Locale = $Locale\n";
+ debug "# $first_a_test..$locales_test_number: \$a = $a, \$b = $b, Locale = $Locale\n";
- tryneoalpha($Locale, 103, $a eq $b);
+ report_result($Locale, ++$locales_test_number, $ok2);
+ $test_names{$locales_test_number} = 'Verify that an intervening sprintf doesn\'t change assignment results';
- my $c = "$x";
- my $z = sprintf ''; # sprintf used to reset locale to "C"
- my $d = "$y";
+ my $first_c_test = $locales_test_number;
- debug "# 104..107: c = $c, d = $d, Locale = $Locale\n";
+ report_result($Locale, ++$locales_test_number, $ok3);
+ $test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a constant';
- tryneoalpha($Locale, 104, $c eq $d);
+ report_result($Locale, ++$locales_test_number, $ok4);
+ $test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a scalar';
- {
- use warnings;
- my $w = 0;
- local $SIG{__WARN__} =
- sub {
- print "# @_\n";
- $w++;
- };
+ report_result($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';
- # The == (among other ops) used to warn for locales
- # that had something else than "." as the radix character.
+ debug "# $first_c_test..$locales_test_number: \$c = $c, \$d = $d, Locale = $Locale\n";
- tryneoalpha($Locale, 105, $c == 1.23);
+ report_result($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, 106, $c == $x);
+ report_result($Locale, ++$locales_test_number, $ok7);
+ $test_names{$locales_test_number} = 'Verify that "==" with a scalar still works in inner no locale';
- tryneoalpha($Locale, 107, $c == $d);
+ report_result($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';
- {
-# no locale; # XXX did this ever work correctly?
-
- my $e = "$x";
+ debug "# $first_e_test..$locales_test_number: \$e = $e, no locale\n";
- debug "# 108..110: e = $e, Locale = $Locale\n";
+ report_result($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, 108, $e == 1.23);
+ report_result($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, 109, $e == $x);
-
- tryneoalpha($Locale, 110, $e == $c);
- }
-
- my $f = "1.23";
- my $g = 2.34;
+ report_result($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';
- debug "# 111..115: f = $f, g = $g, locale = $Locale\n";
+ report_result($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, 111, $f == 1.23);
+ report_result($Locale, ++$locales_test_number, $ok13);
+ $test_names{$locales_test_number} = 'Verify that don\'t get warning under "==" even if radix is not a dot';
- tryneoalpha($Locale, 112, $f == $x);
-
- tryneoalpha($Locale, 113, $f == $c);
+ report_result($Locale, ++$locales_test_number, $ok14);
+ $test_names{$locales_test_number} = 'Verify that non-ASCII UTF-8 error messages are in UTF-8';
- tryneoalpha($Locale, 114, abs(($f + $g) - 3.57) < 0.01);
+ report_result($Locale, ++$locales_test_number, $ok15);
+ $test_names{$locales_test_number} = 'Verify that a number with a UTF-8 radix has a UTF-8 stringification';
- tryneoalpha($Locale, 115, $w == 0);
- }
+ report_result($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';
+
+ debug "# $first_f_test..$locales_test_number: \$f = $f, \$g = $g, back to locale = $Locale\n";
# Does taking lc separately differ from taking
# the lc "in-line"? (This was the bug 19990704.002, change #3568.)
# The bug was in the caching of the 'o'-magic.
- {
+ if (! $is_utf8_locale) {
use locale;
sub lcA {
my $y = "aa";
my $z = "AB";
- tryneoalpha($Locale, 116,
+ report_result($Locale, ++$locales_test_number,
lcA($x, $y) == 1 && lcB($x, $y) == 1 ||
lcA($x, $z) == 0 && lcB($x, $z) == 0);
}
+ else {
+ use locale ':not_characters';
+
+ sub lcC {
+ my $lc0 = lc $_[0];
+ my $lc1 = lc $_[1];
+ return $lc0 cmp $lc1;
+ }
+
+ sub lcD {
+ return lc($_[0]) cmp lc($_[1]);
+ }
+
+ my $x = "ab";
+ my $y = "aa";
+ my $z = "AB";
+
+ report_result($Locale, ++$locales_test_number,
+ lcC($x, $y) == 1 && lcD($x, $y) == 1 ||
+ lcC($x, $z) == 0 && lcD($x, $z) == 0);
+ }
+ $test_names{$locales_test_number} = 'Verify "lc(foo) cmp lc(bar)" is the same as using intermediaries for the cmp';
# Does lc of an UPPER (if different from the UPPER) match
# case-insensitively the UPPER, and does the UPPER match
my $re = qr/[\[\(\{\*\+\?\|\^\$\\]/;
my @f = ();
- foreach my $x (keys %UPPER) {
- 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;
- #
- # If $x and $y contain regular expression characters
- # AND THEY lowercase (/i) to regular expression characters,
- # regcomp() will be mightily confused. No, the \Q doesn't
- # help here (maybe regex engine internal lowercasing
- # is done after the \Q?) An example of this happening is
- # the bg_BG (Bulgarian) locale under EBCDIC (OS/390 USS):
- # the chr(173) (the "[") is the lowercase of the chr(235).
- #
- # Similarly losing EBCDIC locales include cs_cz, cs_CZ,
- # el_gr, el_GR, en_us.IBM-037 (!), en_US.IBM-037 (!),
- # et_ee, et_EE, hr_hr, hr_HR, hu_hu, hu_HU, lt_LT,
- # mk_mk, mk_MK, nl_nl.IBM-037, nl_NL.IBM-037,
- # pl_pl, pl_PL, ro_ro, ro_RO, ru_ru, ru_RU,
- # sk_sk, sk_SK, sl_si, sl_SI, tr_tr, tr_TR.
- #
- # Similar things can happen even under (bastardised)
- # non-EBCDIC locales: in many European countries before the
- # advent of ISO 8859-x nationally customised versions of
- # ISO 646 were devised, reusing certain punctuation
- # characters for modified characters needed by the
- # country/language. For example, the "|" might have
- # stood for U+00F6 or LATIN SMALL LETTER O WITH DIAERESIS.
- #
- if ($x =~ $re || $y =~ $re) {
- print "# Regex characters in '$x' or '$y', skipping test 117 for locale '$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;
- }
-
- foreach my $x (keys %lower) {
- 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;
- if ($x =~ $re || $y =~ $re) { # See above.
- print "# Regex characters in '$x' or '$y', skipping test 117 for locale '$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;
+ ++$locales_test_number;
+ $test_names{$locales_test_number} = 'Verify case insensitive matching works';
+ 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;
+ #
+ # If $x and $y contain regular expression characters
+ # AND THEY lowercase (/i) to regular expression characters,
+ # regcomp() will be mightily confused. No, the \Q doesn't
+ # help here (maybe regex engine internal lowercasing
+ # is done after the \Q?) An example of this happening is
+ # the bg_BG (Bulgarian) locale under EBCDIC (OS/390 USS):
+ # the chr(173) (the "[") is the lowercase of the chr(235).
+ #
+ # Similarly losing EBCDIC locales include cs_cz, cs_CZ,
+ # el_gr, el_GR, en_us.IBM-037 (!), en_US.IBM-037 (!),
+ # et_ee, et_EE, hr_hr, hr_HR, hu_hu, hu_HU, lt_LT,
+ # mk_mk, mk_MK, nl_nl.IBM-037, nl_NL.IBM-037,
+ # pl_pl, pl_PL, ro_ro, ro_RO, ru_ru, ru_RU,
+ # sk_sk, sk_SK, sl_si, sl_SI, tr_tr, tr_TR.
+ #
+ # Similar things can happen even under (bastardised)
+ # non-EBCDIC locales: in many European countries before the
+ # advent of ISO 8859-x nationally customised versions of
+ # ISO 646 were devised, reusing certain punctuation
+ # characters for modified characters needed by the
+ # country/language. For example, the "|" might have
+ # stood for U+00F6 or LATIN SMALL LETTER O WITH DIAERESIS.
+ #
+ if ($x =~ $re || $y =~ $re) {
+ print "# Regex characters in '$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;
+
+ # 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;
+
+ # 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 (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;
+ 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;
+ push @f, $x unless $x =~ /$y/i && $y =~ /$x/i;
+
+ push @f, $x unless lc $x eq fc $x;
+ }
}
- tryneoalpha($Locale, 117, @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, $locales_test_number, @f == 0);
if (@f) {
- print "# failed 117 locale '$Locale' characters @f\n"
+ print "# failed $locales_test_number locale '$Locale' numbers @f\n"
}
}
}
+my $final_locales_test_number = $locales_test_number;
+
# Recount the errors.
-foreach (&last_without_setlocale()+1..$last) {
- if ($Problem{$_} || !defined $Okay{$_} || !@{$Okay{$_}}) {
- if ($_ == 102) {
- print "# The failure of test 102 is not necessarily fatal.\n";
+foreach ($first_locales_test_number..$final_locales_test_number) {
+ if (%setlocale_failed) {
+ print "not ";
+ }
+ elsif ($Problem{$_} || !defined $Okay{$_} || !@{$Okay{$_}}) {
+ if (defined $not_necessarily_a_problem_test_number
+ && $_ == $not_necessarily_a_problem_test_number)
+ {
+ print "# The failure of test $not_necessarily_a_problem_test_number is not necessarily fatal.\n";
print "# It usually indicates a problem in the environment,\n";
print "# not in Perl itself.\n";
}
+ if ($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 $_\n";
+ print "ok $_";
+ 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;
+
+{ # 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 $didwarn = 0;
-foreach (99..$last) {
+foreach ($first_locales_test_number..$final_locales_test_number) {
if ($Problem{$_}) {
my @f = sort keys %{ $Problem{$_} };
my $f = join(" ", @f);
"# 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{$_}\"\n",
"# failed in ", (@f == 1 ? "that locale" : "those locales"),
".\n";
print <<EOW;
if ($didwarn) {
my (@s, @F);
-
+
foreach my $l (@Locale) {
my $p = 0;
- foreach my $t (102..$last) {
- $p++ if $Problem{$t}{$l};
+ if ($setlocale_failed{$l}) {
+ $p++;
+ }
+ else {
+ foreach my $t
+ ($first_locales_test_number..$final_locales_test_number)
+ {
+ $p++ if $Problem{$t}{$l};
+ }
}
push @s, $l if $p == 0;
- push @F, $l unless $p == 0;
+ push @F, $l unless $p == 0;
}
-
+
if (@s) {
my $s = join(" ", @s);
$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";
}
-
- if (@utf8locale) {
- my $S = join(" ", @utf8locale);
- $S =~ s/(.{50,60}) /$1\n#\t/g;
-
- warn "#\n# The following locales\n#\n",
- "#\t", $S, "\n#\n",
- "# were skipped for the tests ",
- join(" ", sort {$a<=>$b} keys %utf8skip), "\n",
- "# because UTF-8 and locales do not work together in Perl.\n#\n";
- }
}
-sub last { 117 }
+print "1..$test_num\n";
# eof
https://perl5.git.perl.org / perl5.git / blobdiff