16-bit MD5 compression for 8-bit string algorithm doubts

Generates a 8-bit random string

function make_coupon_card() {        $code = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ';        $rand = $code[rand(0,25)]            .strtoupper(dechex(date('m')))            .date('d').substr(time(),-5)            .substr(microtime(),2,5)            .sprintf('%02d',rand(0,99));        for(            $a = md5( $rand, true ),            $s = '0123456789ABCDEFGHIJKLMNOPQRSTUV',            $d = '',            $f = 0;            $f < 8;                $g = ord( $a[ $f ] ),            $d .= $s[ ( $g ^ ord( $a[ $f + 8 ] ) ) - $g & 0x1F ],            $f++        );        return $d;    }  

For ($g ^ ord ($a [$f + 8])-$g & 0x1F Bitwise XOR or subtract itself from the operation, the final range is between 0-31, how is this determined?

Reply content:

Generates a 8-bit random string

function make_coupon_card() {        $code = 'ABCDEFGHIJKLMNOPQRSTUVWXYZ';        $rand = $code[rand(0,25)]            .strtoupper(dechex(date('m')))            .date('d').substr(time(),-5)            .substr(microtime(),2,5)            .sprintf('%02d',rand(0,99));        for(            $a = md5( $rand, true ),            $s = '0123456789ABCDEFGHIJKLMNOPQRSTUV',            $d = '',            $f = 0;            $f < 8;                $g = ord( $a[ $f ] ),            $d .= $s[ ( $g ^ ord( $a[ $f + 8 ] ) ) - $g & 0x1F ],            $f++        );        return $d;    }  

For ($g ^ ord ($a [$f + 8])-$g & 0x1F Bitwise XOR or subtract itself from the operation, the final range is between 0-31, how is this determined?

( $g ^ ord( $a[ $f + 8 ] ) ) - $g & 0x1F
In short there is a key point in which operators have - & higher precedence.
So the overall view should be ( ( $g ^ ord( $a[ $f + 8 ] ) ) - $g ) and 0x1F proceed with the operation,
And that 0x1F is the decimal 31 , and the result range is limited to 0 - 31 between.

