[p5sagit/p5-mst-13.2.git] / t / pragma / locale.t

#!./perl -wT

BEGIN {
    chdir 't' if -d 't';
    unshift @INC, '../lib';
    require Config; import Config;
    if (!$Config{d_setlocale} || $Config{ccflags} =~ /\bD?NO_LOCALE\b/) {
	print "1..0\n";
	exit;
    }
}

use strict;

my $debug = 1;

sub debug {
    print @_ if $debug;
}

sub debugf {
    printf @_ if $debug;
}

my $have_setlocale = 0;
eval {
    require POSIX;
    import POSIX ':locale_h';
    $have_setlocale++;
};

# Visual C's CRT goes silly on strings of the form "en_US.ISO8859-1"
# and mingw32 uses said silly CRT
$have_setlocale = 0 if $^O eq 'MSWin32' && $Config{cc} =~ /^(cl|gcc)/i;

print "1..", ($have_setlocale ? 115 : 98), "\n";

use vars qw(&LC_ALL);

my $a = 'abc %';

sub ok {
    my ($n, $result) = @_;

    print 'not ' unless ($result);
    print "ok $n\n";
}

# First we'll do a lot of taint checking for locales.
# This is the easiest to test, actually, as any locale,
# even the default locale will taint under 'use locale'.

sub is_tainted { # hello, camel two.
    local $^W;	# no warnings 'undef'
    my $dummy;
    not eval { $dummy = join("", @_), kill 0; 1 }
}

sub check_taint ($$) {
    ok $_[0], is_tainted($_[1]);
}

sub check_taint_not ($$) {
    ok $_[0], not is_tainted($_[1]);
}

use locale;	# engage locale and therefore locale taint.

check_taint_not   1, $a;

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";

check_taint      10, sprintf('%e', 123.456);
check_taint      11, sprintf('%f', 123.456);
check_taint      12, sprintf('%g', 123.456);
check_taint_not  13, sprintf('%d', 123.456);
check_taint_not  14, sprintf('%x', 123.456);

$_ = $a;	# untaint $_

$_ = uc($a);	# taint $_

check_taint      15, $_;

/(\w)/;	# taint $&, $`, $', $+, $1.
check_taint      16, $&;
check_taint      17, $`;
check_taint      18, $';
check_taint      19, $+;
check_taint      20, $1;
check_taint_not  21, $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;

/(\W)/;	# taint $&, $`, $', $+, $1.
check_taint      28, $&;
check_taint      29, $`;
check_taint      30, $';
check_taint      31, $+;
check_taint      32, $1;
check_taint_not  33, $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;

/(\S)/;	# taint $&, $`, $', $+, $1.
check_taint      40, $&;
check_taint      41, $`;
check_taint      42, $';
check_taint      43, $+;
check_taint      44, $1;
check_taint_not  45, $2;

$_ = $a;	# untaint $_

check_taint_not  46, $_;

/(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;

$_ = $a;	# untaint $_

check_taint_not  53, $_;

$b = uc($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;

$_ = $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;

$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.

$_ = $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;

$_ = $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;

$_ = $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;

$_ = $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;

# After all this tainting $a should be cool.

check_taint_not  98, $a;

# 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;

# Find locales.

debug "# Scanning for locales...\n";

# Note that it's okay that some languages have their native names
# capitalized here even though that's not "right".  They are lowercased
# anyway later during the scanning process (and besides, some clueless
# vendor might have them capitalized errorneously anyway).

my $locales = <<EOF;
Afrikaans:af:za:1 15
Arabic:ar:dz eg sa:6 arabic8
Brezhoneg Breton:br:fr:1 15
Bulgarski Bulgarian:bg:bg:5
Català Catalan:ca:es:1 15
Chinese:zh:cn tw:cn.EUC eucCN eucTW euc.CN euc.TW GB2312 tw.EUC
Hrvatski Croatian:hr:hr:2
Cymraeg Welsh:cy:cy:1 14 15
Czech:cs:cz:2
Dansk Danish:dk:da:1 15
Nederlands Dutch:nl:be nl:1 15
English American British:en:au ca gb ie nz us uk:1 15 cp850
Esperanto:eo:eo:3
Eesti Estonian:et:ee:4 6 13
Suomi Finnish:fi:fi:1 15
Flamish::fl:1 15
Français French:fr:be ca ch fr lu:1 15
Deutsch German:de:at be ch de lu:1 15
Euskaraz Basque:eu:es fr:1 15
Gáidhlig Gaelic:gd:gb uk:1 14 15
Galego Galician:gl:es:1 15
Ellada Greek:el:gr:7 g8
Føroyskt Faroese:fo:fo:1 15
Frysk:fy:nl:1 15
Greenlandic:kl:gl:4 6
Hebrew:iw:il:8 hebrew8
Hungarian:hu:hu:2
Íslensku Icelandic:is:is:1 15
Indonesian:in: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
Korean:ko:kr:
Sámi Lappish:::4 6 13
Latine Latin:la:va:1 15
Latvian:lv:lv:4 6 13
Lithuanian:lt:lt:4 6 13
Macedonian:mk:mk:1 15
Maltese:mt:mt:3
Norsk Norwegian:no:no:1 15
Occitan:oc:es:1 15
Polski Polish:pl:pl:2
Português Portuguese:po:po br:1 15
Rumanian:ro:ro:2
Russki Russian:ru:ru su ua:5 koi8 koi8r koi8u cp1251
Serbski Serbian:sr:yu:5
Slovak:sk:sk:2
Slovene Slovenian:sl:si:2
Espanõl Spanish:es:ar bo cl co cr do ec es gt hn mx ni pa pe py sv uy ve:1 15
Sqhip Albanian:sq:sq:1 15
Svenska Swedish:sv:fi se:1 15
Thai:th:th:11 tis620
Turkish:tr:tr:9 turkish8
Yiddish:::1 15
EOF

my @Locale;
my $Locale;
my @Alnum_;

sub getalnum_ {
    sort grep /\w/, map { chr } 0..255
}

sub trylocale {
    my $locale = shift;
    if (setlocale(LC_ALL, $locale)) {
	push @Locale, $locale;
    }
}

sub decode_encodings {
    my @enc;

    foreach (split(/ /, shift)) {
	if (/^(\d+)$/) {
	    push @enc, "ISO8859-$1";
	    push @enc, "iso8859$1";	# HP
	    if ($1 eq '1') {
		 push @enc, "roman8";	# HP
	    }
	} else {
	    push @enc, $_;
	}
    }

    return @enc;
}

trylocale("C");
trylocale("POSIX");
foreach (0..15) {
    trylocale("ISO8859-$_");
    trylocale("iso8859$_");
    trylocale("iso8859-$_");
    trylocale("iso_8859_$_");
    trylocale("isolatin$_");
    trylocale("isolatin-$_");
    trylocale("iso_latin_$_");
}

foreach my $locale (split(/\n/, $locales)) {
    my ($locale_name, $language_codes, $country_codes, $encodings) =
	split(/:/, $locale);
    my @enc = decode_encodings($encodings);
    foreach my $loc (split(/ /, $locale_name)) {
	trylocale($loc);
	foreach my $enc (@enc) {
	    trylocale("$loc.$enc");
	}
	$loc = lc $loc;
	foreach my $enc (@enc) {
	    trylocale("$loc.$enc");
	}
    }
    foreach my $lang (split(/ /, $language_codes)) {
	trylocale($lang);
	foreach my $country (split(/ /, $country_codes)) {
	    my $lc = "${lang}_${country}";
	    trylocale($lc);
	    foreach my $enc (@enc) {
		trylocale("$lc.$enc");
	    }
	    my $lC = "${lang}_\U${country}";
	    trylocale($lC);
	    foreach my $enc (@enc) {
		trylocale("$lC.$enc");
	    }
	}
    }
}

setlocale(LC_ALL, "C");

@Locale = sort @Locale;

debug "# Locales = @Locale\n";

my %Problem;
my @Neoalpha;

foreach $Locale (@Locale) {
    debug "# Locale = $Locale\n";
    @Alnum_ = getalnum_();
    debug "# \\w = @Alnum_\n";

    unless (setlocale(LC_ALL, $Locale)) {
	foreach (99..103) {
	    $Problem{$_}{$Locale} = -1;
	}
	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 $_) {
		$lower{$_} = $_;
	    }
	}
    }
    foreach (keys %UPPER) {
	$BoThCaSe{$_}++ if exists $lower{$_};
    }
    foreach (keys %lower) {
	$BoThCaSe{$_}++ if exists $UPPER{$_};
    }
    foreach (keys %BoThCaSe) {
	delete $UPPER{$_};
	delete $lower{$_};
    }

    debug "# UPPER    = ", join(" ", sort keys %UPPER   ), "\n";
    debug "# lower    = ", join(" ", sort keys %lower   ), "\n";
    debug "# BoThCaSe = ", join(" ", sort keys %BoThCaSe), "\n";

    # Find the alphabets that are not alphabets in the default locale.

    {
	no locale;
    
	@Neoalpha = ();
	for (keys %UPPER, keys %lower) {
	    push(@Neoalpha, $_) if (/\W/);
	}
    }

    @Neoalpha = sort @Neoalpha;

    debug "# Neoalpha = @Neoalpha\n";

    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";
    } else {

	# Test \w.
    
	debug "# testing 99 with locale '$Locale'\n";
	{
	    my $word = join('', @Neoalpha);

	    $word =~ /^(\w+)$/;

	    if ($1 ne $word) {
		$Problem{99}{$Locale} = 1;
		debug "# failed 99 ($1 vs $word)\n";
	    }
	}

	# Cross-check whole character set.

	debug "# testing 100 with locale '$Locale'\n";
	for (map { chr } 0..255) {
	    if ((/\w/ and /\W/) or (/\d/ and /\D/) or (/\s/ and /\S/)) {
		$Problem{100}{$Locale} = 1;
		debug "# failed 100 for chr(", ord(), ")\n";
	    }
	}

	# Test for read-only scalars' locale vs non-locale comparisons.

	debug "# testing 101 with locale '$Locale'\n";
	{
	    no locale;
	    $a = "qwerty";
	    {
		use locale;
		if ($a cmp "qwerty") {
		    $Problem{101}{$Locale} = 1;
		    debug "# failed 101\n";
		}
	    }
	}

	debug "# testing 102 with locale '$Locale'\n";
	{
	    my ($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)' # 12
		     );
		@test{@test} = 0 x @test;
		$test = 0;
		for my $ti (@test) {
		    $test{$ti} = eval $ti;
		    $test ||= $test{$ti}
		}
		if ($test) {
		    $Problem{102}{$Locale} = 1;
		    debug "# failed 102 at:\n";
		    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 ($x, $y) = (1.23, 1.23);

    my $a = "$x";
    printf ''; # printf used to reset locale to "C"
    my $b = "$y";

    debug "# testing 103 with locale '$Locale'\n";
    unless ($a eq $b) {
	$Problem{103}{$Locale} = 1;
	debug "# failed 103\n";
    }

    my $c = "$x";
    my $z = sprintf ''; # sprintf used to reset locale to "C"
    my $d = "$y";

    debug "# 103..107: a = $a, b = $b, c = $c, d = $d, Locale = $Locale\n";

    debug "# testing 104 with locale '$Locale'\n";
    unless ($c eq $d) {
	$Problem{104}{$Locale} = 1;
	debug "# failed 104\n";
    }

    my $w = 0;
    local $SIG{__WARN__} = sub { $w++ };
    local $^W = 1;

    # the == (among other things) used to warn for locales
    # that had something else than "." as the radix character

    debug "# testing 105 with locale '$Locale'\n";
    unless ($c == 1.23) {
	$Problem{105}{$Locale} = 1;
	debug "# failed 105\n";
    }

    debug "# testing 106 with locale '$Locale'\n";
    unless ($c == $x) {
	$Problem{106}{$Locale} = 1;
	debug "# failed 106\n";
    }

    debug "# testing 107 with locale '$Locale'\n";
    unless ($c == $d) {
	$Problem{107}{$Locale} = 1;
	debug "# failed 107\n";
    }

    {
	no locale;
	
	my $e = "$x";

	debug "# 108..110: e = $e, Locale = $Locale\n";

        debug "# testing 108 with locale '$Locale'\n";
	unless ($e == 1.23) {
	    $Problem{108}{$Locale} = 1;
	    debug "# failed 108\n";
	}

        debug "# testing 109 with locale '$Locale'\n";
	unless ($e == $x) {
	    $Problem{109}{$Locale} = 1;
	    debug "# failed 109\n";
	}

        debug "# testing 110 with locale '$Locale'\n";
	unless ($e == $c) {
	    $Problem{110}{$Locale} = 1;
	    debug "# failed 110\n";
	}
    }

    debug "# testing 111 with locale '$Locale'\n";
    unless ($w == 0) {
	$Problem{110}{$Locale} = 1;
	debug "# failed 111\n";
    }

    my $f = "1.23";

    debug "# 112..114: f = $f, locale = $Locale\n";

    debug "# testing 112 with locale '$Locale'\n";
    unless ($f == 1.23) {
	$Problem{112}{$Locale} = 1;
	debug "# failed 112\n";
    }

    debug "# testing 113 with locale '$Locale'\n";
    unless ($f == $x) {
	$Problem{113}{$Locale} = 1;
	debug "# failed 113\n";
    }

    debug "# testing 114 with locale '$Locale'\n";
    unless ($f == $c) {
	$Problem{114}{$Locale} = 1;
	debug "# failed 114\n";
    }
}

foreach (99..115) {
    if ($Problem{$_}) {
	if ($_ == 102) {
	    print "# The failure of test 102 is not necessarily fatal.\n";
	    print "# It usually indicates a problem in the enviroment,\n";
	    print "# not in Perl itself.\n";
	}
	print "not ";
    }
    print "ok $_\n";
}

my $didwarn = 0;

foreach (99..115) {
    if ($Problem{$_}) {
	my @f = sort keys %{ $Problem{$_} };
	my $f = join(" ", @f);
	$f =~ s/(.{50,60}) /$1\n#\t/g;
	warn
	    "# The locale ", (@f == 1 ? "definition" : "definitions"), "\n#\n",
	    "#\t", $f, "\n#\n",
	    "# on your system may have errors because the locale test $_\n",
            "# failed in ", (@f == 1 ? "that locale" : "those locales"),
            ".\n";
	warn <<EOW;
#
# If your users are not using these locales you are safe for the moment,
# but please report this failure first to perlbug\@perl.com using the
# perlbug script (as described in the INSTALL file) so that the exact
# details of the failures can be sorted out first and then your operating
# system supplier can be alerted about these anomalies.
#
EOW
	$didwarn = 1;
    }
}

if ($didwarn) {
    my @s;
    
    foreach my $l (@Locale) {
	my $p = 0;
	foreach my $t (102..102) {
	    $p++ if $Problem{$t}{$l};
	}
	push @s, $l if $p == 0;
    }
    
    my $s = join(" ", @s);
    $s =~ s/(.{50,60}) /$1\n#\t/g;

    warn
	"# The following locales\n#\n",
        "#\t", $s, "\n#\n",
	"# tested okay.\n#\n",
}

{
    use locale;

}

# eof
Commit	Line	Data
8ebc5c01	1	#!./perl -wT
	2
	3	BEGIN {
	4	chdir 't' if -d 't';
284102e8	5	unshift @INC, '../lib';
b002077a	6	require Config; import Config;
97a0514d	7	if (!$Config{d_setlocale} \|\| $Config{ccflags} =~ /\bD?NO_LOCALE\b/) {
b002077a	8	print "1..0\n";
	9	exit;
	10	}
8ebc5c01	11	}
	12
	13	use strict;
	14
284102e8	15	my $debug = 1;
284102e8	16
6be75cd7	17	sub debug {
	18	print @_ if $debug;
	19	}
	20
	21	sub debugf {
	22	printf @_ if $debug;
	23	}
	24
8ebc5c01	25	my $have_setlocale = 0;
	26	eval {
	27	require POSIX;
	28	import POSIX ':locale_h';
	29	$have_setlocale++;
	30	};
	31
6dead956	32	# Visual C's CRT goes silly on strings of the form "en_US.ISO8859-1"
f6c6487a	33	# and mingw32 uses said silly CRT
f6c6487a	34	$have_setlocale = 0 if $^O eq 'MSWin32' && $Config{cc} =~ /^(cl\|gcc)/i;
6dead956	35
d43ce814	36	print "1..", ($have_setlocale ? 115 : 98), "\n";
8ebc5c01	37
097ee67d	38	use vars qw(&LC_ALL);
8ebc5c01	39
097ee67d	40	my $a = 'abc %';
8ebc5c01	41
	42	sub ok {
	43	my ($n, $result) = @_;
	44
	45	print 'not ' unless ($result);
	46	print "ok $n\n";
	47	}
	48
	49	# First we'll do a lot of taint checking for locales.
	50	# This is the easiest to test, actually, as any locale,
	51	# even the default locale will taint under 'use locale'.
	52
	53	sub is_tainted { # hello, camel two.
3fe9a6f1	54	local $^W; # no warnings 'undef'
8ebc5c01	55	my $dummy;
	56	not eval { $dummy = join("", @_), kill 0; 1 }
	57	}
	58
	59	sub check_taint ($$) {
	60	ok $_[0], is_tainted($_[1]);
	61	}
	62
	63	sub check_taint_not ($$) {
	64	ok $_[0], not is_tainted($_[1]);
	65	}
	66
	67	use locale; # engage locale and therefore locale taint.
	68
	69	check_taint_not 1, $a;
	70
	71	check_taint 2, uc($a);
	72	check_taint 3, "\U$a";
	73	check_taint 4, ucfirst($a);
	74	check_taint 5, "\u$a";
	75	check_taint 6, lc($a);
	76	check_taint 7, "\L$a";
	77	check_taint 8, lcfirst($a);
	78	check_taint 9, "\l$a";
	79
	80	check_taint 10, sprintf('%e', 123.456);
	81	check_taint 11, sprintf('%f', 123.456);
	82	check_taint 12, sprintf('%g', 123.456);
	83	check_taint_not 13, sprintf('%d', 123.456);
	84	check_taint_not 14, sprintf('%x', 123.456);
	85
	86	$_ = $a; # untaint $_
	87
	88	$_ = uc($a); # taint $_
	89
	90	check_taint 15, $_;
	91
	92	/(\w)/; # taint $&, $`, $', $+, $1.
	93	check_taint 16, $&;
	94	check_taint 17, $`;
	95	check_taint 18, $';
	96	check_taint 19, $+;
	97	check_taint 20, $1;
	98	check_taint_not 21, $2;
	99
	100	/(.)/; # untaint $&, $`, $', $+, $1.
	101	check_taint_not 22, $&;
	102	check_taint_not 23, $`;
	103	check_taint_not 24, $';
	104	check_taint_not 25, $+;
	105	check_taint_not 26, $1;
	106	check_taint_not 27, $2;
	107
	108	/(\W)/; # taint $&, $`, $', $+, $1.
	109	check_taint 28, $&;
	110	check_taint 29, $`;
	111	check_taint 30, $';
	112	check_taint 31, $+;
	113	check_taint 32, $1;
	114	check_taint_not 33, $2;
	115
	116	/(\s)/; # taint $&, $`, $', $+, $1.
	117	check_taint 34, $&;
	118	check_taint 35, $`;
119	check_taint 36, $';
120	check_taint 37, $+;
121	check_taint 38, $1;
122	check_taint_not 39, $2;
123
124	/(\S)/; # taint $&, $`, $', $+, $1.
125	check_taint 40, $&;
126	check_taint 41, $`;
127	check_taint 42, $';
128	check_taint 43, $+;
129	check_taint 44, $1;
130	check_taint_not 45, $2;
131
132	$_ = $a; # untaint $_
133
134	check_taint_not 46, $_;
135
136	/(b)/; # this must not taint
137	check_taint_not 47, $&;
138	check_taint_not 48, $`;
139	check_taint_not 49, $';
140	check_taint_not 50, $+;
141	check_taint_not 51, $1;
142	check_taint_not 52, $2;
143
144	$_ = $a; # untaint $_
145
146	check_taint_not 53, $_;
147
148	$b = uc($a); # taint $b
149	s/(.+)/$b/; # this must taint only the $_
150
151	check_taint 54, $_;
152	check_taint_not 55, $&;
153	check_taint_not 56, $`;
154	check_taint_not 57, $';
155	check_taint_not 58, $+;
156	check_taint_not 59, $1;
157	check_taint_not 60, $2;
158
159	$_ = $a; # untaint $_
160
161	s/(.+)/b/; # this must not taint
162	check_taint_not 61, $_;
163	check_taint_not 62, $&;
164	check_taint_not 63, $`;
165	check_taint_not 64, $';
166	check_taint_not 65, $+;
167	check_taint_not 66, $1;
168	check_taint_not 67, $2;
169
170	$b = $a; # untaint $b
171
172	($b = $a) =~ s/\w/$&/;
173	check_taint 68, $b; # $b should be tainted.
174	check_taint_not 69, $a; # $a should be not.
175
176	$_ = $a; # untaint $_
177
178	s/(\w)/\l$1/; # this must taint
179	check_taint 70, $_;
180	check_taint 71, $&;
181	check_taint 72, $`;
182	check_taint 73, $';
183	check_taint 74, $+;
184	check_taint 75, $1;
185	check_taint_not 76, $2;
186
187	$_ = $a; # untaint $_
188
189	s/(\w)/\L$1/; # this must taint
190	check_taint 77, $_;
191	check_taint 78, $&;
192	check_taint 79, $`;
193	check_taint 80, $';
194	check_taint 81, $+;
195	check_taint 82, $1;
196	check_taint_not 83, $2;
197
198	$_ = $a; # untaint $_
199
200	s/(\w)/\u$1/; # this must taint
201	check_taint 84, $_;
202	check_taint 85, $&;
203	check_taint 86, $`;
204	check_taint 87, $';
205	check_taint 88, $+;
206	check_taint 89, $1;
207	check_taint_not 90, $2;
208
209	$_ = $a; # untaint $_
210
211	s/(\w)/\U$1/; # this must taint
212	check_taint 91, $_;
213	check_taint 92, $&;
214	check_taint 93, $`;
215	check_taint 94, $';
216	check_taint 95, $+;
217	check_taint 96, $1;
218	check_taint_not 97, $2;
219
220	# After all this tainting $a should be cool.
221
222	check_taint_not 98, $a;
223
224	# I think we've seen quite enough of taint.
225	# Let us do some real locale work now,
284102e8	226	# unless setlocale() is missing (i.e. minitest).
8ebc5c01	227
	228	exit unless $have_setlocale;
	229
284102e8	230	# Find locales.
284102e8	231
6be75cd7	232	debug "# Scanning for locales...\n";
	233
	234	# Note that it's okay that some languages have their native names
	235	# capitalized here even though that's not "right". They are lowercased
	236	# anyway later during the scanning process (and besides, some clueless
	237	# vendor might have them capitalized errorneously anyway).
	238
284102e8	239	my $locales = <<EOF;
6be75cd7	240	Afrikaans:af:za:1 15
284102e8	241	Arabic:ar:dz eg sa:6 arabic8
6be75cd7	242	Brezhoneg Breton:br:fr:1 15
	243	Bulgarski Bulgarian:bg:bg:5
	244	Català Catalan:ca:es:1 15
	245	Chinese:zh:cn tw:cn.EUC eucCN eucTW euc.CN euc.TW GB2312 tw.EUC
	246	Hrvatski Croatian:hr:hr:2
	247	Cymraeg Welsh:cy:cy:1 14 15
284102e8	248	Czech:cs:cz:2
6be75cd7	249	Dansk Danish:dk:da:1 15
	250	Nederlands Dutch:nl:be nl:1 15
	251	English American British:en:au ca gb ie nz us uk:1 15 cp850
	252	Esperanto:eo:eo:3
	253	Eesti Estonian:et:ee:4 6 13
	254	Suomi Finnish:fi:fi:1 15
	255	Flamish::fl:1 15
d43ce814	256	Français French:fr:be ca ch fr lu:1 15
6be75cd7	257	Deutsch German:de:at be ch de lu:1 15
	258	Euskaraz Basque:eu:es fr:1 15
	259	Gáidhlig Gaelic:gd:gb uk:1 14 15
	260	Galego Galician:gl:es:1 15
	261	Ellada Greek:el:gr:7 g8
	262	Føroyskt Faroese:fo:fo:1 15
	263	Frysk:fy:nl:1 15
	264	Greenlandic:kl:gl:4 6
284102e8	265	Hebrew:iw:il:8 hebrew8
284102e8	266	Hungarian:hu:hu:2
6be75cd7	267	Íslensku Icelandic:is:is:1 15
	268	Indonesian:in:id:1 15
	269	Gaeilge Irish:ga:IE:1 14 15
	270	Italiano Italian:it:ch it:1 15
	271	Nihongo Japanese:ja:jp:euc eucJP jp.EUC sjis
284102e8	272	Korean:ko:kr:
6be75cd7	273	Sámi Lappish:::4 6 13
	274	Latine Latin:la:va:1 15
	275	Latvian:lv:lv:4 6 13
	276	Lithuanian:lt:lt:4 6 13
	277	Macedonian:mk:mk:1 15
	278	Maltese:mt:mt:3
d43ce814	279	Norsk Norwegian:no:no:1 15
6be75cd7	280	Occitan:oc:es:1 15
	281	Polski Polish:pl:pl:2
	282	Português Portuguese:po:po br:1 15
284102e8	283	Rumanian:ro:ro:2
6be75cd7	284	Russki Russian:ru:ru su ua:5 koi8 koi8r koi8u cp1251
6be75cd7	285	Serbski Serbian:sr:yu:5
284102e8	286	Slovak:sk:sk:2
6be75cd7	287	Slovene Slovenian:sl:si:2
6be75cd7	288	Espanõl Spanish:es:ar bo cl co cr do ec es gt hn mx ni pa pe py sv uy ve:1 15
d43ce814	289	Sqhip Albanian:sq:sq:1 15
d43ce814	290	Svenska Swedish:sv:fi se:1 15
6be75cd7	291	Thai:th:th:11 tis620
284102e8	292	Turkish:tr:tr:9 turkish8
6be75cd7	293	Yiddish:::1 15
284102e8	294	EOF
	295
	296	my @Locale;
	297	my $Locale;
	298	my @Alnum_;
	299
	300	sub getalnum_ {
8ebc5c01	301	sort grep /\w/, map { chr } 0..255
	302	}
	303
284102e8	304	sub trylocale {
	305	my $locale = shift;
	306	if (setlocale(LC_ALL, $locale)) {
	307	push @Locale, $locale;
	308	}
	309	}
8ebc5c01	310
284102e8	311	sub decode_encodings {
284102e8	312	my @enc;
8ebc5c01	313
284102e8	314	foreach (split(/ /, shift)) {
	315	if (/^(\d+)$/) {
	316	push @enc, "ISO8859-$1";
	317	push @enc, "iso8859$1"; # HP
	318	if ($1 eq '1') {
	319	push @enc, "roman8"; # HP
	320	}
	321	} else {
	322	push @enc, $_;
8ebc5c01	323	}
	324	}
	325
284102e8	326	return @enc;
8ebc5c01	327	}
8ebc5c01	328
284102e8	329	trylocale("C");
	330	trylocale("POSIX");
	331	foreach (0..15) {
	332	trylocale("ISO8859-$_");
284102e8	333	trylocale("iso8859$_");
097ee67d	334	trylocale("iso8859-$_");
	335	trylocale("iso_8859_$_");
	336	trylocale("isolatin$_");
	337	trylocale("isolatin-$_");
	338	trylocale("iso_latin_$_");
8ebc5c01	339	}
8ebc5c01	340
284102e8	341	foreach my $locale (split(/\n/, $locales)) {
	342	my ($locale_name, $language_codes, $country_codes, $encodings) =
	343	split(/:/, $locale);
	344	my @enc = decode_encodings($encodings);
	345	foreach my $loc (split(/ /, $locale_name)) {
	346	trylocale($loc);
	347	foreach my $enc (@enc) {
	348	trylocale("$loc.$enc");
	349	}
	350	$loc = lc $loc;
	351	foreach my $enc (@enc) {
	352	trylocale("$loc.$enc");
	353	}
	354	}
	355	foreach my $lang (split(/ /, $language_codes)) {
	356	trylocale($lang);
	357	foreach my $country (split(/ /, $country_codes)) {
	358	my $lc = "${lang}_${country}";
	359	trylocale($lc);
	360	foreach my $enc (@enc) {
	361	trylocale("$lc.$enc");
	362	}
	363	my $lC = "${lang}_\U${country}";
	364	trylocale($lC);
	365	foreach my $enc (@enc) {
	366	trylocale("$lC.$enc");
	367	}
	368	}
	369	}
	370	}
4599a1de	371
d43ce814	372	setlocale(LC_ALL, "C");
d43ce814	373
4599a1de	374	@Locale = sort @Locale;
4599a1de	375
284102e8	376	debug "# Locales = @Locale\n";
8ebc5c01	377
284102e8	378	my %Problem;
097ee67d	379	my @Neoalpha;
284102e8	380
	381	foreach $Locale (@Locale) {
	382	debug "# Locale = $Locale\n";
	383	@Alnum_ = getalnum_();
	384	debug "# \\w = @Alnum_\n";
	385
	386	unless (setlocale(LC_ALL, $Locale)) {
	387	foreach (99..103) {
	388	$Problem{$_}{$Locale} = -1;
8ebc5c01	389	}
284102e8	390	next;
8ebc5c01	391	}
8ebc5c01	392
284102e8	393	# Sieve the uppercase and the lowercase.
284102e8	394
097ee67d	395	my %UPPER = ();
	396	my %lower = ();
	397	my %BoThCaSe = ();
284102e8	398	for (@Alnum_) {
	399	if (/[^\d_]/) { # skip digits and the _
	400	if (uc($_) eq $_) {
	401	$UPPER{$_} = $_;
	402	}
	403	if (lc($_) eq $_) {
	404	$lower{$_} = $_;
	405	}
	406	}
	407	}
	408	foreach (keys %UPPER) {
097ee67d	409	$BoThCaSe{$_}++ if exists $lower{$_};
284102e8	410	}
284102e8	411	foreach (keys %lower) {
097ee67d	412	$BoThCaSe{$_}++ if exists $UPPER{$_};
284102e8	413	}
097ee67d	414	foreach (keys %BoThCaSe) {
284102e8	415	delete $UPPER{$_};
	416	delete $lower{$_};
	417	}
	418
	419	debug "# UPPER = ", join(" ", sort keys %UPPER ), "\n";
	420	debug "# lower = ", join(" ", sort keys %lower ), "\n";
097ee67d	421	debug "# BoThCaSe = ", join(" ", sort keys %BoThCaSe), "\n";
284102e8	422
284102e8	423	# Find the alphabets that are not alphabets in the default locale.
8ebc5c01	424
284102e8	425	{
284102e8	426	no locale;
8ebc5c01	427
284102e8	428	@Neoalpha = ();
	429	for (keys %UPPER, keys %lower) {
	430	push(@Neoalpha, $_) if (/\W/);
	431	}
8ebc5c01	432	}
8ebc5c01	433
284102e8	434	@Neoalpha = sort @Neoalpha;
8ebc5c01	435
284102e8	436	debug "# Neoalpha = @Neoalpha\n";
8ebc5c01	437
284102e8	438	if (@Neoalpha == 0) {
284102e8	439	# If we have no Neoalphas the remaining tests are no-ops.
6be75cd7	440	debug "# no Neoalpha, skipping tests 99..102 for locale '$Locale'\n";
6be75cd7	441	} else {
8ebc5c01	442
6be75cd7	443	# Test \w.
284102e8	444
6be75cd7	445	debug "# testing 99 with locale '$Locale'\n";
	446	{
	447	my $word = join('', @Neoalpha);
8ebc5c01	448
6be75cd7	449	$word =~ /^(\w+)$/;
8ebc5c01	450
6be75cd7	451	if ($1 ne $word) {
	452	$Problem{99}{$Locale} = 1;
	453	debug "# failed 99 ($1 vs $word)\n";
	454	}
284102e8	455	}
8ebc5c01	456
6be75cd7	457	# Cross-check whole character set.
8ebc5c01	458
6be75cd7	459	debug "# testing 100 with locale '$Locale'\n";
	460	for (map { chr } 0..255) {
	461	if ((/\w/ and /\W/) or (/\d/ and /\D/) or (/\s/ and /\S/)) {
	462	$Problem{100}{$Locale} = 1;
	463	debug "# failed 100 for chr(", ord(), ")\n";
	464	}
284102e8	465	}
8ebc5c01	466
6be75cd7	467	# Test for read-only scalars' locale vs non-locale comparisons.
284102e8	468
6be75cd7	469	debug "# testing 101 with locale '$Locale'\n";
284102e8	470	{
6be75cd7	471	no locale;
	472	$a = "qwerty";
	473	{
	474	use locale;
	475	if ($a cmp "qwerty") {
	476	$Problem{101}{$Locale} = 1;
	477	debug "# failed 101\n";
	478	}
8ebc5c01	479	}
8ebc5c01	480	}
8ebc5c01	481
6be75cd7	482	debug "# testing 102 with locale '$Locale'\n";
	483	{
	484	my ($from, $to, $lesser, $greater,
	485	@test, %test, $test, $yes, $no, $sign);
	486
	487	for (0..9) {
	488	# Select a slice.
	489	$from = int(($_*@Alnum_)/10);
	490	$to = $from + int(@Alnum_/10);
	491	$to = $#Alnum_ if ($to > $#Alnum_);
	492	$lesser = join('', @Alnum_[$from..$to]);
	493	# Select a slice one character on.
	494	$from++; $to++;
	495	$to = $#Alnum_ if ($to > $#Alnum_);
	496	$greater = join('', @Alnum_[$from..$to]);
	497	($yes, $no, $sign) = ($lesser lt $greater
	498	? (" ", "not ", 1)
	499	: ("not ", " ", -1));
	500	# all these tests should FAIL (return 0).
	501	# Exact lt or gt cannot be tested because
	502	# in some locales, say, eacute and E may test equal.
	503	@test =
	504	(
	505	$no.' ($lesser le $greater)', # 1
	506	'not ($lesser ne $greater)', # 2
	507	' ($lesser eq $greater)', # 3
	508	$yes.' ($lesser ge $greater)', # 4
	509	$yes.' ($lesser ge $greater)', # 5
	510	$yes.' ($greater le $lesser )', # 7
	511	'not ($greater ne $lesser )', # 8
	512	' ($greater eq $lesser )', # 9
	513	$no.' ($greater ge $lesser )', # 10
	514	'not (($lesser cmp $greater) == -$sign)' # 12
	515	);
	516	@test{@test} = 0 x @test;
	517	$test = 0;
284102e8	518	for my $ti (@test) {
6be75cd7	519	$test{$ti} = eval $ti;
6be75cd7	520	$test \|\|= $test{$ti}
284102e8	521	}
6be75cd7	522	if ($test) {
	523	$Problem{102}{$Locale} = 1;
	524	debug "# failed 102 at:\n";
	525	debug "# lesser = '$lesser'\n";
	526	debug "# greater = '$greater'\n";
	527	debug "# lesser cmp greater = ",
	528	$lesser cmp $greater, "\n";
	529	debug "# greater cmp lesser = ",
	530	$greater cmp $lesser, "\n";
	531	debug "# (greater) from = $from, to = $to\n";
	532	for my $ti (@test) {
	533	debugf("# %-40s %-4s", $ti,
	534	$test{$ti} ? 'FAIL' : 'ok');
	535	if ($ti =~ /\(\.(\$.+ +cmp +\$[^\)]+)\.\)/) {
	536	debugf("(%s == %4d)", $1, eval $1);
	537	}
	538	debug "\n#";
	539	}
284102e8	540
6be75cd7	541	last;
6be75cd7	542	}
284102e8	543	}
8ebc5c01	544	}
8ebc5c01	545	}
6be75cd7	546
	547	use locale;
	548
	549	my ($x, $y) = (1.23, 1.23);
	550
	551	my $a = "$x";
	552	printf ''; # printf used to reset locale to "C"
	553	my $b = "$y";
	554
	555	debug "# testing 103 with locale '$Locale'\n";
	556	unless ($a eq $b) {
	557	$Problem{103}{$Locale} = 1;
	558	debug "# failed 103\n";
	559	}
	560
	561	my $c = "$x";
	562	my $z = sprintf ''; # sprintf used to reset locale to "C"
	563	my $d = "$y";
	564
	565	debug "# 103..107: a = $a, b = $b, c = $c, d = $d, Locale = $Locale\n";
	566
	567	debug "# testing 104 with locale '$Locale'\n";
	568	unless ($c eq $d) {
	569	$Problem{104}{$Locale} = 1;
	570	debug "# failed 104\n";
	571	}
	572
	573	my $w = 0;
	574	local $SIG{__WARN__} = sub { $w++ };
	575	local $^W = 1;
	576
	577	# the == (among other things) used to warn for locales
	578	# that had something else than "." as the radix character
	579
	580	debug "# testing 105 with locale '$Locale'\n";
	581	unless ($c == 1.23) {
	582	$Problem{105}{$Locale} = 1;
	583	debug "# failed 105\n";
	584	}
	585
	586	debug "# testing 106 with locale '$Locale'\n";
	587	unless ($c == $x) {
	588	$Problem{106}{$Locale} = 1;
	589	debug "# failed 106\n";
	590	}
	591
	592	debug "# testing 107 with locale '$Locale'\n";
	593	unless ($c == $d) {
	594	$Problem{107}{$Locale} = 1;
	595	debug "# failed 107\n";
	596	}
	597
	598	{
	599	no locale;
	600
	601	my $e = "$x";
	602
	603	debug "# 108..110: e = $e, Locale = $Locale\n";
	604
	605	debug "# testing 108 with locale '$Locale'\n";
	606	unless ($e == 1.23) {
	607	$Problem{108}{$Locale} = 1;
	608	debug "# failed 108\n";
	609	}
610
611	debug "# testing 109 with locale '$Locale'\n";
612	unless ($e == $x) {
613	$Problem{109}{$Locale} = 1;
614	debug "# failed 109\n";
615	}
616
617	debug "# testing 110 with locale '$Locale'\n";
618	unless ($e == $c) {
619	$Problem{110}{$Locale} = 1;
620	debug "# failed 110\n";
621	}
622	}
623
624	debug "# testing 111 with locale '$Locale'\n";
625	unless ($w == 0) {
626	$Problem{110}{$Locale} = 1;
627	debug "# failed 111\n";
628	}
629
630	my $f = "1.23";
631
632	debug "# 112..114: f = $f, locale = $Locale\n";
633
634	debug "# testing 112 with locale '$Locale'\n";
635	unless ($f == 1.23) {
636	$Problem{112}{$Locale} = 1;
637	debug "# failed 112\n";
638	}
639
640	debug "# testing 113 with locale '$Locale'\n";
641	unless ($f == $x) {
642	$Problem{113}{$Locale} = 1;
643	debug "# failed 113\n";
644	}
645
646	debug "# testing 114 with locale '$Locale'\n";
647	unless ($f == $c) {
648	$Problem{114}{$Locale} = 1;
649	debug "# failed 114\n";
650	}
8ebc5c01	651	}
284102e8	652
d43ce814	653	foreach (99..115) {
284102e8	654	if ($Problem{$_}) {
097ee67d	655	if ($_ == 102) {
097ee67d	656	print "# The failure of test 102 is not necessarily fatal.\n";
284102e8	657	print "# It usually indicates a problem in the enviroment,\n";
	658	print "# not in Perl itself.\n";
	659	}
	660	print "not ";
8ebc5c01	661	}
284102e8	662	print "ok $_\n";
8ebc5c01	663	}
fb73857a	664
284102e8	665	my $didwarn = 0;
284102e8	666
d43ce814	667	foreach (99..115) {
284102e8	668	if ($Problem{$_}) {
	669	my @f = sort keys %{ $Problem{$_} };
	670	my $f = join(" ", @f);
	671	$f =~ s/(.{50,60}) /$1\n#\t/g;
	672	warn
	673	"# The locale ", (@f == 1 ? "definition" : "definitions"), "\n#\n",
	674	"#\t", $f, "\n#\n",
	675	"# on your system may have errors because the locale test $_\n",
	676	"# failed in ", (@f == 1 ? "that locale" : "those locales"),
	677	".\n";
	678	warn <<EOW;
	679	#
	680	# If your users are not using these locales you are safe for the moment,
	681	# but please report this failure first to perlbug\@perl.com using the
	682	# perlbug script (as described in the INSTALL file) so that the exact
	683	# details of the failures can be sorted out first and then your operating
	684	# system supplier can be alerted about these anomalies.
	685	#
	686	EOW
	687	$didwarn = 1;
fb73857a	688	}
fb73857a	689	}
774d564b	690
284102e8	691	if ($didwarn) {
	692	my @s;
	693
	694	foreach my $l (@Locale) {
	695	my $p = 0;
097ee67d	696	foreach my $t (102..102) {
284102e8	697	$p++ if $Problem{$t}{$l};
8ebc5c01	698	}
284102e8	699	push @s, $l if $p == 0;
8ebc5c01	700	}
284102e8	701
	702	my $s = join(" ", @s);
	703	$s =~ s/(.{50,60}) /$1\n#\t/g;
	704
	705	warn
	706	"# The following locales\n#\n",
	707	"#\t", $s, "\n#\n",
	708	"# tested okay.\n#\n",
8ebc5c01	709	}
90248788	710
097ee67d	711	{
	712	use locale;
	713
097ee67d	714	}
097ee67d	715
90248788	716	# eof