Implicit conversion from float (number) to int loses precision

Question

I have used this formula before php 8.1

<?php
$number = 0;
echo log10(abs($number)) / 3 | 0;

echo PHP_EOL;

$number = 100;
echo log10(abs($number)) / 3 | 0;

echo PHP_EOL;

$number = 1100;
echo log10(abs($number)) / 3 | 0;

echo PHP_EOL;

$number = 10000000;
echo log10(abs($number)) / 3 | 0;
?>

It worked fine but now after upgrading I keep getting these errors

Deprecated: Implicit conversion from float -INF to int loses precision

Deprecated: Implicit conversion from float 0.6666666666666666 to int loses precision

Deprecated: Implicit conversion from float 1.0137975617194084 to int loses precision

Deprecated: Implicit conversion from float 2.3333333333333335 to int loses precision

I can't find or understand from the 8.1 documentation why this is happening now

Answer 1

// Implicit variant
$number= "2";
$calc = 2 + $number;

// Cast variant
$number = "2";
$calc= 2 + (int) $number;

// Explicit variant
$number = "2";
$calc = 2 + intval($number);

// Everything is good variant ^^
$number = 2;
$calc = 2 + $number;

Answer 2

When you perform a bitwise OR operation via the | operator, you get an implicit conversion to an integer. This is a...weird...way to convert to an integer. To avoid the warning, just cast explicitly.

Implicit:

echo log10(abs($number)) / 3 | 0;

Explicitly pass the function:

echo intval(log10(abs($number)) / 3);

Or via cast:

echo (int) (log10(abs($number)) / 3);