Bit shifting in C is different from PHP-CodePudding

I have a question about a small piece of code in C to make the same piece of code work in PHP, it has to do with a bit shift and I can't figure out what's wrong.

unsigned u = 3910796769;
u  = u << 8;
printf("%u\n",u); 
//Result : 52422369

PHP:

$u = 3910796769;
$u  = $u << 8;
printf("%u\n",$u);
//Result : 1005074769633

CodePudding user response：

Well, unsigned in C is 32bit, you cannot even shift the number you provided once without triggering an overflow, but you have shifted it 8 times and added one more time, like multiplying the number by 257, you should get the result mod 2^32 == 4294967296:

unsigned u = 3910796769;
u  = u << 8;

this should be 256*u u == 257 * u == 1005074769633 ~= 52422369 (mod 4294967296) You can test it.

[...]
//Result : 52422369  /* correct (mod 2^32) */

PHP probably uses 64bit integers for the operations, and the result properly fits in 64bit.

$u = 3910796769;
$u  = $u << 8;
printf("%u\n",$u);
//Result : 1005074769633

But if you try:

#include <stdio.h>
#include <stdint.h>

int main()
{
    uint64_t u = 3910796769;
    u  = u << 8;
    printf("%Lu\n", u); 
    //Result : 1005074769633
}

you will get the correct result.

CodePudding user response：

In my case, I needed to select elements from an array filled with 32-bit values using a specific formula. The answer from @Eugene Sh helped me do this in PHP.

$u = 3910796769;
$u  = $u << 8;
$u = $u & 0xFFFFFFFF
printf("%u\n",$u);
//Result : 52422369