hash collisions

collisions.php

<?php
declare(strict_types = 1);

ini_set("memory_limit", "-1");
$algo = "xxh64";
$dictionary = "crackstation.txt";

// MODE_dict_in_ram_hashes_in_ram($algo, $file);
MODE_dict_on_disk_hash_on_sqlite($algo, $dictionary);

function MODE_dict_on_disk_hash_on_sqlite(string $algo, string $dictionary): void
{
    $sqlite_file = __FILE__ . ".sqlite3";
    if (file_exists($sqlite_file)) {
        if (! unlink($sqlite_file)) {
            throw new \RuntimeException("unable to unlink sqlite3 file!");
        }
    }
    $db = new \PDO('sqlite:' . $sqlite_file, '', '', array(
        \PDO::ATTR_EMULATE_PREPARES => false,
        \PDO::ATTR_ERRMODE => \PDO::ERRMODE_EXCEPTION,
        \PDO::ATTR_DEFAULT_FETCH_MODE => \PDO::FETCH_ASSOC
    ));
    $db->exec('PRAGMA synchronous = OFF');
    $db->exec('PRAGMA journal_mode = OFF');
    $sql = 'CREATE TABLE t (
	h BLOB PRIMARY KEY,
   	w BLOB NOT NULL) WITHOUT ROWID;';
    $db->exec($sql);
    $fp = fopen($dictionary, "rb");
    $stm = $db->prepare("INSERT INTO t VALUES(?,?)");
    $hash = "";
    $word = "";
    $stm->bindParam(1, $hash);
    $stm->bindParam(2, $word);
    while (($word = fgets($fp))) {
        $word = rtrim($word, "\n");
        $hash = hash($algo, $word, true);
        try {
            $stm->execute();
        } catch (\Throwable $ex) {
            $collision = $db->query("SELECT w FROM t WHERE h = " . $db->quote($hash))
                ->fetch(\PDO::FETCH_NUM)[0];
            print_collision($collision, $word);
        }
        //echo ".";
    }
    var_dump($db->query("SELECT COUNT(*) FROM t")->fetch());
}

function MODE_dict_on_disk_hashes_in_ram(string $algo, string $dictionary): void
{
    $fp = fopen($dictionary, "rb");
    $hashes = [];
    $counter = 0;
    register_shutdown_function(function () use (&$counter) {
        var_dump([
            "counter" => $counter
        ]);
    });
    while (($word = fgets($fp))) {
        ++ $counter;
        $word = rtrim($word, "\n");
        $hash = hash($algo, $word, true);
        if (isset($hashes[$hash])) {
            print_collision($hashes[$hash], $word);
        }
        $hashes[$hash] = $word;
    }
}

function MODE_dict_in_ram_hashes_in_ram(string $algo, string $dictionary): void
{
    echo "loading..";
    $words = file($dictionary, FILE_IGNORE_NEW_LINES);
    echo ". done!\n";
    $hashes = array();
    foreach ($words as $word) {
        $hash = hash($algo, $word, true);
        if (isset($hashes[$hash])) {
            print_collision($hashes[$hash], $word);
        }
        $hashes[$hash] = $word;
    }
}

function print_collision(string $str1, string $str2): void
{
    echo "\"{$str1}\" (hex " . bin2hex($str1) . " ) collides with \"{$str2}\" (hex " . bin2hex($str2) . " )\n";
}