);
sub debug {
return unless $debug;
- my($mess) = join "", @_;
+ my($mess) = join "", '# ', @_;
chop $mess;
print $dumper->stringify($mess,1), "\n";
}
# Find locales.
-debug "# Scanning for locales...\n";
+debug "Scanning for locales...\n";
require POSIX; import POSIX ':locale_h';
-my @Locale = find_locales([ &POSIX::LC_CTYPE, &POSIX::LC_ALL ]);
+my @Locale = find_locales([ &POSIX::LC_CTYPE, &POSIX::LC_NUMERIC, &POSIX::LC_ALL ]);
-debug "# Locales =\n";
+debug "Locales =\n";
for ( @Locale ) {
- debug "# $_\n";
+ debug "$_\n";
}
unless (@Locale) {
return $output;
}
+sub disp_str ($) {
+ my $string = shift;
+
+ # Displays the string unambiguously. ASCII printables are always output
+ # as-is, though perhaps separated by blanks from other characters. If
+ # entirely printable ASCII, just returns the string. Otherwise if valid
+ # UTF-8 it uses the character names for non-printable-ASCII. Otherwise it
+ # outputs hex for each non-ASCII-printable byte.
+
+ return $string if $string =~ / ^ [[:print:]]* $/xa;
+
+ my $result = "";
+ my $prev_was_punct = 1; # Beginning is considered punct
+ if (utf8::valid($string) && utf8::is_utf8($string)) {
+ use charnames ();
+ foreach my $char (split "", $string) {
+
+ # Keep punctuation adjacent to other characters; otherwise
+ # separate them with a blank
+ if ($char =~ /[[:punct:]]/a) {
+ $result .= $char;
+ $prev_was_punct = 1;
+ }
+ elsif ($char =~ /[[:print:]]/a) {
+ $result .= " " unless $prev_was_punct;
+ $result .= $char;
+ $prev_was_punct = 0;
+ }
+ else {
+ $result .= " " unless $prev_was_punct;
+ $result .= charnames::viacode(ord $char);
+ $prev_was_punct = 0;
+ }
+ }
+ }
+ else {
+ use bytes;
+ foreach my $char (split "", $string) {
+ if ($char =~ /[[:punct:]]/a) {
+ $result .= $char;
+ $prev_was_punct = 1;
+ }
+ elsif ($char =~ /[[:print:]]/a) {
+ $result .= " " unless $prev_was_punct;
+ $result .= $char;
+ $prev_was_punct = 0;
+ }
+ else {
+ $result .= " " unless $prev_was_punct;
+ $result .= sprintf("%02X", ord $char);
+ $prev_was_punct = 0;
+ }
+ }
+ }
+
+ return $result;
+}
+
sub report_result {
my ($Locale, $i, $pass_fail, $message) = @_;
$message //= "";
$message = " ($message)" if $message;
unless ($pass_fail) {
$Problem{$i}{$Locale} = 1;
- debug "# failed $i ($test_names{$i}) with locale '$Locale'$message\n";
+ debug "failed $i ($test_names{$i}) with locale '$Locale'$message\n";
} else {
push @{$Okay{$i}}, $Locale;
}
foreach my $Locale (@Locale) {
$locales_test_number = $first_locales_test_number - 1;
- debug "#\n";
- debug "# Locale = $Locale\n";
+ debug "\n";
+ debug "Locale = $Locale\n";
unless (setlocale(&POSIX::LC_ALL, $Locale)) {
$setlocale_failed{$Locale} = $Locale;
my $is_utf8_locale = is_locale_utf8($Locale);
- debug "# is utf8 locale? = $is_utf8_locale\n";
+ debug "is utf8 locale? = $is_utf8_locale\n";
- my $radix = localeconv()->{decimal_point};
- if ($radix !~ / ^ [[:ascii:]] + $/x) {
- use bytes;
- $radix = disp_chars(split "", $radix);
- }
- debug "# radix = $radix\n";
+ debug "radix = " . disp_str(localeconv()->{decimal_point}) . "\n";
if (! $is_utf8_locale) {
use locale;
# Ordered, where possible, in groups of "this is a subset of the next
# one"
- debug "# :upper: = ", disp_chars(@{$posixes{'upper'}}), "\n";
- debug "# :lower: = ", disp_chars(@{$posixes{'lower'}}), "\n";
- debug "# :cased: = ", disp_chars(@{$posixes{'cased'}}), "\n";
- debug "# :alpha: = ", disp_chars(@{$posixes{'alpha'}}), "\n";
- debug "# :alnum: = ", disp_chars(@{$posixes{'alnum'}}), "\n";
- debug "# w = ", disp_chars(@{$posixes{'word'}}), "\n";
- debug "# :graph: = ", disp_chars(@{$posixes{'graph'}}), "\n";
- debug "# :print: = ", disp_chars(@{$posixes{'print'}}), "\n";
- debug "# d = ", disp_chars(@{$posixes{'digit'}}), "\n";
- debug "# :xdigit: = ", disp_chars(@{$posixes{'xdigit'}}), "\n";
- debug "# :blank: = ", disp_chars(@{$posixes{'blank'}}), "\n";
- debug "# s = ", disp_chars(@{$posixes{'space'}}), "\n";
- debug "# :punct: = ", disp_chars(@{$posixes{'punct'}}), "\n";
- debug "# :cntrl: = ", disp_chars(@{$posixes{'cntrl'}}), "\n";
- debug "# :ascii: = ", disp_chars(@{$posixes{'ascii'}}), "\n";
+ debug ":upper: = ", disp_chars(@{$posixes{'upper'}}), "\n";
+ debug ":lower: = ", disp_chars(@{$posixes{'lower'}}), "\n";
+ debug ":cased: = ", disp_chars(@{$posixes{'cased'}}), "\n";
+ debug ":alpha: = ", disp_chars(@{$posixes{'alpha'}}), "\n";
+ debug ":alnum: = ", disp_chars(@{$posixes{'alnum'}}), "\n";
+ debug " w = ", disp_chars(@{$posixes{'word'}}), "\n";
+ debug ":graph: = ", disp_chars(@{$posixes{'graph'}}), "\n";
+ debug ":print: = ", disp_chars(@{$posixes{'print'}}), "\n";
+ debug " d = ", disp_chars(@{$posixes{'digit'}}), "\n";
+ debug ":xdigit: = ", disp_chars(@{$posixes{'xdigit'}}), "\n";
+ debug ":blank: = ", disp_chars(@{$posixes{'blank'}}), "\n";
+ debug " s = ", disp_chars(@{$posixes{'space'}}), "\n";
+ debug ":punct: = ", disp_chars(@{$posixes{'punct'}}), "\n";
+ debug ":cntrl: = ", disp_chars(@{$posixes{'cntrl'}}), "\n";
+ debug ":ascii: = ", disp_chars(@{$posixes{'ascii'}}), "\n";
foreach (keys %UPPER) {
}
}
- debug "# UPPER = ", disp_chars(sort { ord $a <=> ord $b } keys %UPPER), "\n";
- debug "# lower = ", disp_chars(sort { ord $a <=> ord $b } keys %lower), "\n";
- debug "# BoThCaSe = ", disp_chars(sort { ord $a <=> ord $b } keys %BoThCaSe), "\n";
- debug "# Unassigned = ", disp_chars(sort { ord $a <=> ord $b } keys %Unassigned), "\n";
+ debug "UPPER = ", disp_chars(sort { ord $a <=> ord $b } keys %UPPER), "\n";
+ debug "lower = ", disp_chars(sort { ord $a <=> ord $b } keys %lower), "\n";
+ debug "BoThCaSe = ", disp_chars(sort { ord $a <=> ord $b } keys %BoThCaSe), "\n";
+ debug "Unassigned = ", disp_chars(sort { ord $a <=> ord $b } keys %Unassigned), "\n";
my @failures;
my @fold_failures;
@Added_alpha = sort { ord $a <=> ord $b } @Added_alpha;
- debug "# Added_alpha = ", disp_chars(@Added_alpha), "\n";
+ debug "Added_alpha = ", disp_chars(@Added_alpha), "\n";
# Cross-check the whole 8-bit character set.
}
report_result($Locale, $locales_test_number, $test == 0);
if ($test) {
- debug "# lesser = '$lesser'\n";
- debug "# greater = '$greater'\n";
- debug "# lesser cmp greater = ",
+ debug "lesser = '$lesser'\n";
+ debug "greater = '$greater'\n";
+ debug "lesser cmp greater = ",
$lesser cmp $greater, "\n";
- debug "# greater cmp lesser = ",
+ debug "greater cmp lesser = ",
$greater cmp $lesser, "\n";
- debug "# (greater) from = $from, to = $to\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#";
+ debugf("\n#");
}
last;
# 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).
+ # first). This is only done if the current locale has LC_MESSAGES
$ok14 = 1;
$ok14_5 = 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);
- no locale;
- $ok14_5 = "$!" !~ /\P{ASCII}/;
- last;
+ if (setlocale(&POSIX::LC_MESSAGES, $Locale)) {
+ foreach my $err (keys %!) {
+ use Errno;
+ $! = eval "&Errno::$err"; # Convert to strerror() output
+ my $strerror = "$!";
+ if ("$strerror" =~ /\P{ASCII}/) {
+ $ok14 = utf8::is_utf8($strerror);
+ no locale;
+ $ok14_5 = "$!" !~ /\P{ASCII}/;
+ last;
+ }
}
}
$ok18 = $j eq sprintf("%g:%g", $h, $i);
}
- { # These tests aren't affected by :not_characters
-
+ $ok19 = $ok20 = 1;
+ if (setlocale(&POSIX::LC_TIME, $Locale)) { # These tests aren't affected by
+ # :not_characters
my @times = CORE::localtime();
use locale;
$ok19 = POSIX::strftime("%p", @times) ne "%p"; # [perl #119425]
- my $date = POSIX::strftime("%A %B %Z", @times);
+ my $date = POSIX::strftime("'%A' '%B' '%Z' '%p'", @times);
+ debug("'Day' 'Month' 'TZ' 'am/pm' = ", disp_str($date));
# If there is any non-ascii, it better be UTF-8 in a UTF-8 locale, and
# not UTF-8 if the locale isn't UTF-8.
$test_names{$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 = $Locale\n";
+ debug "$first_a_test..$locales_test_number: \$a = $a, \$b = $b, Locale = $Locale\n";
report_result($Locale, ++$locales_test_number, $ok2);
$test_names{$locales_test_number} = 'Verify that an intervening sprintf doesn\'t change assignment results';
$test_names{$locales_test_number} = 'Verify that a different locale radix works when doing "==" with a scalar and an intervening sprintf';
$problematical_tests{$locales_test_number} = 1;
- debug "# $first_c_test..$locales_test_number: \$c = $c, \$d = $d, Locale = $Locale\n";
+ debug "$first_c_test..$locales_test_number: \$c = $c, \$d = $d, Locale = $Locale\n";
report_result($Locale, ++$locales_test_number, $ok6);
$test_names{$locales_test_number} = 'Verify that can assign stringified under inner no-locale block';
$test_names{$locales_test_number} = 'Verify that "==" with a scalar and an intervening sprintf still works in inner no locale';
$problematical_tests{$locales_test_number} = 1;
- debug "# $first_e_test..$locales_test_number: \$e = $e, no locale\n";
+ debug "$first_e_test..$locales_test_number: \$e = $e, no locale\n";
report_result($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';
report_result($Locale, ++$locales_test_number, $ok20);
$test_names{$locales_test_number} = 'Verify that strftime returns date with UTF-8 flag appropriately set';
+ $problematical_tests{$locales_test_number} = 1; # This is broken in
+ # OS X 10.9.3
report_result($Locale, ++$locales_test_number, $ok21);
$test_names{$locales_test_number} = '"$!" is ASCII only outside of locale scope';
- debug "# $first_f_test..$locales_test_number: \$f = $f, \$g = $g, back to locale = $Locale\n";
+ debug "$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.)
if (! $is_utf8_locale) {
my $y = lc $x;
next unless uc $y eq $x;
- debug_more( "# UPPER=", disp_chars(($x)),
+ debug_more( "UPPER=", disp_chars(($x)),
"; lc=", disp_chars(($y)), "; ",
"; fc=", disp_chars((fc $x)), "; ",
disp_chars(($x)), "=~/", disp_chars(($y)), "/i=",
use locale ':not_characters';
my $y = lc $x;
next unless uc $y eq $x;
- debug_more( "# UPPER=", disp_chars(($x)),
+ debug_more( "UPPER=", disp_chars(($x)),
"; lc=", disp_chars(($y)), "; ",
"; fc=", disp_chars((fc $x)), "; ",
disp_chars(($x)), "=~/", disp_chars(($y)), "/i=",
if (! $is_utf8_locale) {
my $y = uc $x;
next unless lc $y eq $x;
- debug_more( "# lower=", disp_chars(($x)),
+ debug_more( "lower=", disp_chars(($x)),
"; uc=", disp_chars(($y)), "; ",
"; fc=", disp_chars((fc $x)), "; ",
disp_chars(($x)), "=~/", disp_chars(($y)), "/i=",
use locale ':not_characters';
my $y = uc $x;
next unless lc $y eq $x;
- debug_more( "# lower=", disp_chars(($x)),
+ debug_more( "lower=", disp_chars(($x)),
"; uc=", disp_chars(($y)), "; ",
"; fc=", disp_chars((fc $x)), "; ",
disp_chars(($x)), "=~/", disp_chars(($y)), "/i=",
print "# problem is not likely to be Perl's\n";
}
}
- elsif ($debug) {
+ if ($debug) {
print "# $percent_fail% of locales (",
scalar(keys $Problem{$test_num}),
" of ",
scalar(@Locale),
- ") fail the following test\n";
+ ") fail the above test (TODO cut-off is ",
+ $acceptable_failure_percentage,
+ "%)\n";
}
}
print "#\n";