#!./perl -w
# Tests for sprintf that do not fit the format of sprintf.t.
BEGIN {
chdir 't' if -d 't';
@INC = '../lib';
require './test.pl';
}
# We'll run 12 extra tests (see below) if $Q is false.
eval { my $q = pack "q", 0 };
my $Q = $@ eq '';
my $doubledouble;
# %a and %A depend on the floating point config
# This totally doesn't test non-IEEE-754 float formats.
my @hexfloat;
print "# uvsize = $Config{uvsize}\n";
print "# nvsize = $Config{nvsize}\n";
print "# nv_preserves_uv_bits = $Config{nv_preserves_uv_bits}\n";
print "# d_quad = $Config{d_quad}\n";
if ($Config{nvsize} == 8 &&
(
# IEEE-754 64-bit ("double precision"), the most common out there
($Config{uvsize} == 8 && $Config{nv_preserves_uv_bits} == 53)
||
# If we have a quad we can still get the mantissa bits.
($Config{uvsize} == 4 && $Config{d_quad})
)
) {
@hexfloat = (
[ '%a', '0', '0x0p+0' ],
[ '%a', '1', '0x1p+0' ],
[ '%a', '1.0', '0x1p+0' ],
[ '%a', '0.5', '0x1p-1' ],
[ '%a', '0.25', '0x1p-2' ],
[ '%a', '0.75', '0x1.8p-1' ],
[ '%a', '3.14', '0x1.91eb851eb851fp+1' ],
[ '%a', '-1.0', '-0x1p+0' ],
[ '%a', '-3.14', '-0x1.91eb851eb851fp+1' ],
[ '%a', '0.1', '0x1.999999999999ap-4' ],
[ '%a', '1/7', '0x1.2492492492492p-3' ],
[ '%a', 'sqrt(2)', '0x1.6a09e667f3bcdp+0' ],
[ '%a', 'exp(1)', '0x1.5bf0a8b145769p+1' ],
[ '%a', '2**-10', '0x1p-10' ],
[ '%a', '2**10', '0x1p+10' ],
[ '%a', '1e-9', '0x1.12e0be826d695p-30' ],
[ '%a', '1e9', '0x1.dcd65p+29' ],
[ '%#a', '1', '0x1.p+0' ],
[ '%+a', '1', '+0x1p+0' ],
[ '%+a', '-1', '-0x1p+0' ],
[ '% a', ' 1', ' 0x1p+0' ],
[ '% a', '-1', '-0x1p+0' ],
[ '%8a', '3.14', '0x1.91eb851eb851fp+1' ],
[ '%13a', '3.14', '0x1.91eb851eb851fp+1' ],
[ '%20a', '3.14', '0x1.91eb851eb851fp+1' ],
[ '%.4a', '3.14', '0x1.91ecp+1' ],
[ '%.5a', '3.14', '0x1.91eb8p+1' ],
[ '%.6a', '3.14', '0x1.91eb85p+1' ],
[ '%.20a', '3.14', '0x1.91eb851eb851f0000000p+1' ],
[ '%20.10a', '3.14', ' 0x1.91eb851eb8p+1' ],
[ '%20.15a', '3.14', '0x1.91eb851eb851f00p+1' ],
[ '% 20.10a', '3.14', ' 0x1.91eb851eb8p+1' ],
[ '%020.10a', '3.14', '0x0001.91eb851eb8p+1' ],
[ '%30a', '3.14', ' 0x1.91eb851eb851fp+1' ],
[ '%-30a', '3.14', '0x1.91eb851eb851fp+1 ' ],
[ '%030a', '3.14', '0x00000000001.91eb851eb851fp+1' ],
[ '%-030a', '3.14', '0x1.91eb851eb851fp+1 ' ],
[ '%.40a', '3.14',
'0x1.91eb851eb851f000000000000000000000000000p+1' ],
[ '%A', '3.14', '0X1.91EB851EB851FP+1' ],
);
} elsif (($Config{nvsize} == 16 || $Config{nvsize} == 12) &&
# 80-bit ("extended precision") long double, pack F is the NV
# cd cc cc cc cc cc cc cc fb bf 00 00 00 00 00 00
# cd cc cc cc cc cc cc cc fb bf 00 00
(pack("F", 0.1) =~ /^\xCD/ || # LE
pack("F", 0.1) =~ /\xCD$/)) { # BE (if this ever happens)
@hexfloat = (
[ '%a', '0', '0x0p+0' ],
[ '%a', '1', '0x8p-3' ],
[ '%a', '1.0', '0x8p-3' ],
[ '%a', '0.5', '0x8p-4' ],
[ '%a', '0.25', '0x8p-5' ],
[ '%a', '0.75', '0xcp-4' ],
[ '%a', '3.14', '0xc.8f5c28f5c28f5c3p-2' ],
[ '%a', '-1.0', '-0x8p-3' ],
[ '%a', '-3.14', '-0xc.8f5c28f5c28f5c3p-2' ],
[ '%a', '0.1', '0xc.ccccccccccccccdp-7' ],
[ '%a', '1/7', '0x9.249249249249249p-6' ],
[ '%a', 'sqrt(2)', '0xb.504f333f9de6484p-3' ],
[ '%a', 'exp(1)', '0xa.df85458a2bb4a9bp-2' ],
[ '%a', '2**-10', '0x8p-13' ],
[ '%a', '2**10', '0x8p+7' ],
[ '%a', '1e-9', '0x8.9705f4136b4a597p-33' ],
[ '%a', '1e9', '0xe.e6b28p+26' ],
[ '%#a', '1', '0x8.p-3' ],
[ '%+a', '1', '+0x8p-3' ],
[ '%+a', '-1', '-0x8p-3' ],
[ '% a', ' 1', ' 0x8p-3' ],
[ '% a', '-1', '-0x8p-3' ],
[ '%8a', '3.14', '0xc.8f5c28f5c28f5c3p-2' ],
[ '%13a', '3.14', '0xc.8f5c28f5c28f5c3p-2' ],
[ '%20a', '3.14', '0xc.8f5c28f5c28f5c3p-2' ],
[ '%.4a', '3.14', '0xc.8f5cp-2' ],
[ '%.5a', '3.14', '0xc.8f5c3p-2' ],
[ '%.6a', '3.14', '0xc.8f5c29p-2' ],
[ '%.20a', '3.14', '0xc.8f5c28f5c28f5c300000p-2' ],
[ '%20.10a', '3.14', ' 0xc.8f5c28f5c3p-2' ],
[ '%20.15a', '3.14', '0xc.8f5c28f5c28f5c3p-2' ],
[ '% 20.10a', '3.14', ' 0xc.8f5c28f5c3p-2' ],
[ '%020.10a', '3.14', '0x000c.8f5c28f5c3p-2' ],
[ '%30a', '3.14', ' 0xc.8f5c28f5c28f5c3p-2' ],
[ '%-30a', '3.14', '0xc.8f5c28f5c28f5c3p-2 ' ],
[ '%030a', '3.14', '0x00000000c.8f5c28f5c28f5c3p-2' ],
[ '%-030a', '3.14', '0xc.8f5c28f5c28f5c3p-2 ' ],
[ '%.40a', '3.14',
'0xc.8f5c28f5c28f5c30000000000000000000000000p-2' ],
[ '%A', '3.14', '0XC.8F5C28F5C28F5C3P-2' ],
);
} elsif (
# IEEE 754 128-bit ("quadruple precision"), e.g. IA-64 (Itanium) in VMS
$Config{nvsize} == 16 &&
# 9a 99 99 99 99 99 99 99 99 99 99 99 99 99 fb 3f (LE), pack F is the NV
(pack("F", 0.1) =~ /^\x9A\x99{6}/ || # LE
pack("F", 0.1) =~ /\x99{6}\x9A$/) # BE
) {
@hexfloat = (
[ '%a', '0', '0x0p+0' ],
[ '%a', '1', '0x1p+0' ],
[ '%a', '1.0', '0x1p+0' ],
[ '%a', '0.5', '0x1p-1' ],
[ '%a', '0.25', '0x1p-2' ],
[ '%a', '0.75', '0x1.8p-1' ],
[ '%a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%a', '-1', '-0x1p+0' ],
[ '%a', '-3.14', '-0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%a', '0.1', '0x1.999999999999999999999999999ap-4' ],
[ '%a', '1/7', '0x1.2492492492492492492492492492p-3' ],
[ '%a', 'sqrt(2)', '0x1.6a09e667f3bcc908b2fb1366ea95p+0' ],
[ '%a', 'exp(1)', '0x1.5bf0a8b1457695355fb8ac404e7ap+1' ],
[ '%a', '2**-10', '0x1p-10' ],
[ '%a', '2**10', '0x1p+10' ],
[ '%a', '1e-09', '0x1.12e0be826d694b2e62d01511f12ap-30' ],
[ '%a', '1e9', '0x1.dcd65p+29' ],
[ '%#a', '1', '0x1.p+0' ],
[ '%+a', '1', '+0x1p+0' ],
[ '%+a', '-1', '-0x1p+0' ],
[ '% a', '1', ' 0x1p+0' ],
[ '% a', '-1', '-0x1p+0' ],
[ '%8a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%13a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%20a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%.4a', '3.14', '0x1.91ecp+1' ],
[ '%.5a', '3.14', '0x1.91eb8p+1' ],
[ '%.6a', '3.14', '0x1.91eb85p+1' ],
[ '%.20a', '3.14', '0x1.91eb851eb851eb851eb8p+1' ],
[ '%20.10a', '3.14', ' 0x1.91eb851eb8p+1' ],
[ '%20.15a', '3.14', '0x1.91eb851eb851eb8p+1' ],
[ '% 20.10a', '3.14', ' 0x1.91eb851eb8p+1' ],
[ '%020.10a', '3.14', '0x0001.91eb851eb8p+1' ],
[ '%30a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%-30a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%030a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%-030a', '3.14', '0x1.91eb851eb851eb851eb851eb851fp+1' ],
[ '%.40a', '3.14',
'0x1.91eb851eb851eb851eb851eb851f000000000000p+1' ],
[ '%A', '3.14', '0X1.91EB851EB851EB851EB851EB851FP+1' ],
);
} elsif (
# "double-double", two 64-bit doubles end to end
$Config{nvsize} == 16 &&
# bf b9 99 99 99 99 99 9a bc 59 99 99 99 99 99 9a (BE), pack F is the NV
(pack("F", 0.1) =~ /^\x9A\x99{5}\x59\xBC/ || # LE
pack("F", 0.1) =~ /\xBC\x59\x99{5}\x9A$/) # BE
) {
$doubledouble = 1;
@hexfloat = (
[ '%a', '0', '0x0p+0' ],
[ '%a', '1', '0x1p+0' ],
[ '%a', '1.0', '0x1p+0' ],
[ '%a', '0.5', '0x1p-1' ],
[ '%a', '0.25', '0x1p-2' ],
[ '%a', '0.75', '0x1.8p-1' ],
[ '%a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%a', '-1', '-0x0p+0' ],
[ '%a', '-3.14', '-0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%a', '0.1', '0x1.999999999999999999999999998p-4' ],
[ '%a', '1/7', '0x1.249249249249249249249249248p-3' ],
[ '%a', 'sqrt(2)', '0x1.6a09e667f3bcc908b2fb1366ea8p+0' ],
[ '%a', 'exp(1)', '0x1.5bf0a8b1457695355fb8ac404e8p+1' ],
[ '%a', '2**-10', '0x1p-10' ],
[ '%a', '2**10', '0x1p+10' ],
[ '%a', '1e-09', '0x1.12e0be826d694b2e62d01511f14p-30' ],
[ '%a', '1e9', '0x1.dcd65p+29' ],
[ '%#a', '1', '0x1.p+0' ],
[ '%+a', '1', '+0x1p+0' ],
[ '%+a', '-1', '-0x1p+0' ],
[ '% a', '1', ' 0x1p+0' ],
[ '% a', '-1', '-0x1p+0' ],
[ '%8a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%13a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%20a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%.4a', '3.14', '0x1.91ecp+1' ],
[ '%.5a', '3.14', '0x1.91eb8p+1' ],
[ '%.6a', '3.14', '0x1.91eb85p+1' ],
[ '%.20a', '3.14', '0x1.91eb851eb851eb851eb8p+1' ],
[ '%20.10a', '3.14', ' 0x1.91eb851eb8p+1' ],
[ '%20.15a', '3.14', '0x1.91eb851eb851eb8p+1' ],
[ '% 20.10a', '3.14', ' 0x1.91eb851eb8p+1' ],
[ '%020.10a', '3.14', '0x0001.91eb851eb8p+1' ],
[ '%30a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%-30a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%030a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%-030a', '3.14', '0x1.91eb851eb851eb851eb851eb85p+1' ],
[ '%.40a', '3.14',
'0x1.91eb851eb851eb851eb851eb8500000000000000p+1' ],
[ '%A', '3.14', '0X1.91EB851EB851EB851EB851EB85P+1' ],
);
} else {
print "# no hexfloat tests\n";
}
plan tests => 1408 + ($Q ? 0 : 12) + @hexfloat;
use strict;
use Config;
is(
sprintf("%.40g ",0.01),
sprintf("%.40g", 0.01)." ",
q(the sprintf "%.<number>g" optimization)
);
is(
sprintf("%.40f ",0.01),
sprintf("%.40f", 0.01)." ",
q(the sprintf "%.<number>f" optimization)
);
# cases of $i > 1 are against [perl #39126]
for my $i (1, 5, 10, 20, 50, 100) {
chop(my $utf8_format = "%-*s\x{100}");
my $string = "\xB4"x$i; # latin1 ACUTE or ebcdic COPYRIGHT
my $expect = $string." "x$i; # followed by 2*$i spaces
is(sprintf($utf8_format, 3*$i, $string), $expect,
"width calculation under utf8 upgrade, length=$i");
}
# check simultaneous width & precision with wide characters
for my $i (1, 3, 5, 10) {
my $string = "\x{0410}"x($i+10); # cyrillic capital A
my $expect = "\x{0410}"x$i; # cut down to exactly $i characters
my $format = "%$i.${i}s";
is(sprintf($format, $string), $expect,
"width & precision interplay with utf8 strings, length=$i");
}
# Used to mangle PL_sv_undef
fresh_perl_like(
'print sprintf "xxx%n\n"; print undef',
qr/Modification of a read-only value attempted at - line 1\./,
{ switches => [ '-w' ] },
q(%n should not be able to modify read-only constants),
);
# check overflows
for (int(~0/2+1), ~0, "9999999999999999999") {
is(eval {sprintf "%${_}d", 0}, undef, "no sprintf result expected %${_}d");
like($@, qr/^Integer overflow in format string for sprintf /, "overflow in sprintf");
is(eval {printf "%${_}d\n", 0}, undef, "no printf result expected %${_}d");
like($@, qr/^Integer overflow in format string for printf /, "overflow in printf");
}
# check %NNN$ for range bounds
{
my ($warn, $bad) = (0,0);
local $SIG{__WARN__} = sub {
if ($_[0] =~ /missing argument/i) {
$warn++
}
else {
$bad++
}
};
my $fmt = join('', map("%$_\$s%" . ((1 << 31)-$_) . '$s', 1..20));
my $result = sprintf $fmt, qw(a b c d);
is($result, "abcd", "only four valid values in $fmt");
is($warn, 36, "expected warnings");
is($bad, 0, "unexpected warnings");
}
# Tests for "missing argument" and "redundant argument" warnings
{
my ($warn_missing, $warn_redundant, $warn_bad) = (0,0,0);
local $SIG{__WARN__} = sub {
if ($_[0] =~ /missing argument/i) {
$warn_missing++
}
elsif ($_[0] =~ /redundant argument/i) {
$warn_redundant++
}
else {
$warn_bad++
}
};
my @tests = (
# The "", "%s", and "%-p" formats have special-case handling
# in sv.c
{
fmt => "",
args => [ qw( x ) ],
res => "",
m => 0,
r => 1,
},
{
fmt => "%s",
args => [ qw( x y ) ],
res => "x",
m => 0,
r => 1,
},
{
fmt => "%-p",
args => [ qw( x y ) ],
res => qr/^[0-9a-f]+$/as,
m => 0,
r => 1,
},
# Other non-specialcased patterns
{
fmt => "%s : %s",
args => [ qw( a b c ) ],
res => "a : b",
m => 0,
r => 1,
},
{
fmt => "%s : %s : %s",
args => [ qw( a b c d e ) ],
res => "a : b : c",
m => 0,
# Note how we'll only warn about redundant arguments once,
# even though both "d" and "e" are redundant...
r => 1,
},
{
fmt => "%s : %s : %s",
args => [ ],
res => " : : ",
# ...But when arguments are missing we'll warn about every
# missing argument. This difference between the two
# warnings is a feature.
m => 3,
r => 0,
},
# Tests for format parameter indexes.
#
# Deciding what to do about these is a bit tricky, and so is
# "correctly" warning about missing arguments on them.
#
# Should we warn if you supply 4 arguments but only use
# argument 1,3 & 4? Or only if you supply 5 arguments and your
# highest used argument is 4?
#
# For some uses of this printf feature (e.g. i18n systems)
# it's a always a logic error to not print out every provided
# argument, but for some other uses skipping some might be a
# feature (although you could argue that then printf should be
# called as e.g:
#
# printf q[%1$s %3$s], x(), undef, z();
#
# Instead of:
#
# printf q[%1$s %3$s], x(), y(), z();
#
# Since calling the (possibly expensive) y() function is
# completely redundant there.
#
# We deal with all these potential problems by not even
# trying. If the pattern contains any format parameter indexes
# whatsoever we'll never warn about redundant arguments.
{
fmt => '%1$s : %2$s',
args => [ qw( x y z ) ],
res => "x : y",
m => 0,
r => 0,
},
{
fmt => '%2$s : %4$s : %5$s',
args => [ qw( a b c d )],
res => "b : d : ",
m => 1,
r => 0,
},
{
fmt => '%s : %1$s : %s',
args => [ qw( x y z ) ],
res => "x : x : y",
m => 0,
r => 0,
},
);
for my $i (0..$#tests) {
my $test = $tests[$i];
my $result = sprintf $test->{fmt}, @{$test->{args}};
my $prefix = "For format '$test->{fmt}' and arguments/result '@{$test->{args}}'/'$result'";
if (ref $test->{res} eq 'Regexp') {
like($result, $test->{res}, "$prefix got the right result");
} else {
is($result, $test->{res}, "$prefix got the right result");
}
is($warn_missing, $test->{m}, "$prefix got '$test->{m}' 'missing argument' warnings");
is($warn_redundant, $test->{r}, "$prefix got '$test->{r}' 'redundant argument' warnings");
is($warn_bad, 0, "$prefix No unknown warnings");
($warn_missing, $warn_redundant, $warn_bad) = (0,0,0);
}
}
{
foreach my $ord (0 .. 255) {
my $bad = 0;
local $SIG{__WARN__} = sub {
if ($_[0] !~ /^Invalid conversion in sprintf/) {
warn $_[0];
$bad++;
}
};
my $r = eval {sprintf '%v' . chr $ord};
is ($bad, 0, "pattern '%v' . chr $ord");
}
}
sub mysprintf_int_flags {
my ($fmt, $num) = @_;
die "wrong format $fmt" if $fmt !~ /^%([-+ 0]+)([1-9][0-9]*)d\z/;
my $flag = $1;
my $width = $2;
my $sign = $num < 0 ? '-' :
$flag =~ /\+/ ? '+' :
$flag =~ /\ / ? ' ' :
'';
my $abs = abs($num);
my $padlen = $width - length($sign.$abs);
return
$flag =~ /0/ && $flag !~ /-/ # do zero padding
? $sign . '0' x $padlen . $abs
: $flag =~ /-/ # left or right
? $sign . $abs . ' ' x $padlen
: ' ' x $padlen . $sign . $abs;
}
# Whole tests for "%4d" with 2 to 4 flags;
# total counts: 3 * (4**2 + 4**3 + 4**4) == 1008
my @flags = ("-", "+", " ", "0");
for my $num (0, -1, 1) {
for my $f1 (@flags) {
for my $f2 (@flags) {
for my $f3 ('', @flags) { # '' for doubled flags
my $flag = $f1.$f2.$f3;
my $width = 4;
my $fmt = '%'."${flag}${width}d";
my $result = sprintf($fmt, $num);
my $expect = mysprintf_int_flags($fmt, $num);
is($result, $expect, qq/sprintf("$fmt",$num)/);
next if $f3 eq '';
for my $f4 (@flags) { # quadrupled flags
my $flag = $f1.$f2.$f3.$f4;
my $fmt = '%'."${flag}${width}d";
my $result = sprintf($fmt, $num);
my $expect = mysprintf_int_flags($fmt, $num);
is($result, $expect, qq/sprintf("$fmt",$num)/);
}
}
}
}
}
# test that %f doesn't panic with +Inf, -Inf, NaN [perl #45383]
foreach my $n (2**1e100, -2**1e100, 2**1e100/2**1e100) { # +Inf, -Inf, NaN
eval { my $f = sprintf("%f", $n); };
is $@, "", "sprintf(\"%f\", $n)";
}
# test %ll formats with and without HAS_QUAD
my @tests = (
[ '%lld' => [qw( 4294967296 -100000000000000 )] ],
[ '%lli' => [qw( 4294967296 -100000000000000 )] ],
[ '%llu' => [qw( 4294967296 100000000000000 )] ],
[ '%Ld' => [qw( 4294967296 -100000000000000 )] ],
[ '%Li' => [qw( 4294967296 -100000000000000 )] ],
[ '%Lu' => [qw( 4294967296 100000000000000 )] ],
);
for my $t (@tests) {
my($fmt, $nums) = @$t;
for my $num (@$nums) {
my $w = '';
local $SIG{__WARN__} = sub { $w .= shift };
my $sprintf_got = sprintf($fmt, $num);
if ($Q) {
is($sprintf_got, $num, "quad: $fmt -> $num");
is($w, '', "no warnings for: $fmt -> $num");
} else {
is($sprintf_got, $fmt, "quad unsupported: $fmt -> $fmt");
like($w, qr/Invalid conversion in sprintf: "$fmt"/, "got warning about invalid conversion from fmt : $fmt");
like($w, qr/Redundant argument in sprintf/, "got warning about redundant argument in sprintf from fmt : $fmt");
}
}
}
# Check unicode vs byte length
for my $width (1,2,3,4,5,6,7) {
for my $precis (1,2,3,4,5,6,7) {
my $v = "\x{20ac}\x{20ac}";
my $format = "%" . $width . "." . $precis . "s";
my $chars = ($precis > 2 ? 2 : $precis);
my $space = ($width < 2 ? 0 : $width - $chars);
fresh_perl_is(
'my $v = "\x{20ac}\x{20ac}"; my $x = sprintf "'.$format.'", $v; $x =~ /^(\s*)(\S*)$/; print "$_" for map {length} $1, $2',
"$space$chars",
{},
q(sprintf ").$format.q(", "\x{20ac}\x{20ac}"),
);
}
}
# Overload count
package o {
use overload
'""', sub { ++our $count; $_[0][0]; },
'0+', sub { ++our $numcount; $_[0][1]; }
}
my $o = bless ["\x{100}",42], o::;
() = sprintf "%1s", $o;
is $o::count, '1', 'sprinf %1s overload count';
$o::count = 0;
() = sprintf "%.1s", $o;
is $o::count, '1', 'sprinf %.1s overload count';
$o::count = 0;
() = sprintf "%d", $o;
is $o::count, 0, 'sprintf %d string overload count is 0';
is $o::numcount, 1, 'sprintf %d number overload count is 1';
my $ppc64_linux = $Config{archname} =~ /^ppc64-linux/;
for my $t (@hexfloat) {
my ($format, $arg, $expected) = @$t;
$arg = eval $arg;
my $result = sprintf($format, $arg);
my $ok = $result eq $expected;
if ($doubledouble && $ppc64_linux && $arg =~ /^2.71828/) {
# ppc64-linux has buggy exp(1).
local $::TODO = "$Config{archname} exp(1)";
ok($ok, "'$format' '$arg' -> '$result' cf '$expected'");
next;
}
unless ($ok) {
# It seems that there can be difference in the last bits:
# [perl #122578]
# got "0x1.5bf0a8b14576ap+1"
# expected "0x1.5bf0a8b145769p+1"
# (Android on ARM)
#
# Exact cause unknown but suspecting different fp rounding modes,
# (towards zero? towards +inf? towards -inf?) about which Perl
# is blissfully unaware.
#
# Try extracting one (or sometimes two) last mantissa
# hexdigits, and see if they differ in value by one.
my ($rh, $eh) = ($result, $expected);
sub extract_prefix {
($_[0] =~ s/(-?0x[0-9a-fA-F]+\.)//) && return $1;
}
my $rp = extract_prefix($rh);
my $ep = extract_prefix($eh);
print "# rp = $rp, ep = $ep (rh $rh, eh $eh)\n";
if ($rp eq $ep) { # If prefixes match.
sub extract_exponent {
($_[0] =~ s/([pP][+-]?\d+)//) && return $1;
}
my $re = extract_exponent($rh);
my $ee = extract_exponent($eh);
print "# re = $re, ee = $ee (rh $rh, eh $eh)\n";
if ($re eq $ee) { # If exponents match.
# Remove the common prefix of the mantissa bits.
my $la = length($rh);
my $lb = length($eh);
my $i;
for ($i = 0; $i < $la && $i < $lb; $i++) {
last if substr($rh, $i, 1) ne substr($eh, $i, 1);
}
$rh = substr($rh, $i);
$eh = substr($eh, $i);
print "# (rh $rh, eh $eh)\n";
if ($rh ne $eh) {
# If necessary, pad the shorter one on the right
# with one zero (for example "...1f" vs "...2",
# we want to compare "1f" to "20").
if (length $rh < length $eh) {
$rh .= '0';
} elsif (length $eh < length $rh) {
$eh .= '0';
}
print "# (rh $rh, eh $eh)\n";
if (length $eh == length $rh) {
if (abs(hex($eh) - hex($rh)) == 1) {
$ok = 1;
}
}
}
}
}
}
ok($ok, "'$format' '$arg' -> '$result' cf '$expected'");
}