【Big BUG】Why 439!=4.39 * 100, But 438==4.38 * 100?

[gaohuazi]

Description

The following code:

<?php
echo 'PHP版本：' . phpversion() . "\n\n";
/*-------439-----*/
$a = 439;
$b = 4.39 * 100;

var_dump($a);
var_dump($b);
var_dump($a == $b);
echo "\n";

/*-------438-----*/
$a = 438;
$b = 4.38 * 100;

var_dump($a);
var_dump($b);
var_dump($a == $b);
echo "\n\n";
exit;

Resulted in this output:

PHP版本：7.4.26

int(439)
float(439)
bool(false)

int(438)
float(438)
bool(true)

PHP版本：8.1.0

int(439)
float(438.99999999999994)
bool(false)

int(438)
float(438)
bool(true)

PHP Version

PHP 8.1.0 /7.4.26

Operating System

win7 SP1

[iluuu1994]

This is caused by imprecision in floating point arithmetics (IEEE-754).

https://3v4l.org/tMcPL#v8.1.9

var_dump(4.38 * 100);
var_dump(4.39 * 100);

float(438)
float(438.99999999999994)

[thg2k]

That is not a bug, it is the expected behaviour for floating point numbers.

Two solutions:

1. If you want to compare ints: var_dump(439 == (int) round(4.39 * 100.0));
2. If you want to compare floats: var_dump((439 - 4.39 * 100) < 0.00001); (pick a precision that is reasonable with the context)

[gaohuazi]

That is not a bug, it is the expected behaviour for floating point numbers.

Two solutions:

1. If you want to compare ints: var_dump(439 == (int) round(4.39 * 100.0));
2. If you want to compare floats: var_dump((439 - 4.39 * 100) < 0.00001); (pick a precision that is reasonable with the context)

Thanks for your reply,I use function 'bcmul' resloved this problem~

var_dump(bcmul(4.39, 100) == 439);