사용자 입력값 정제

정제와 유효성 검사의 차이

HTML 출력을 위한 htmlspecialchars()

<?php
$name = '<script>alert(1)</script>';

// Dangerous — executes script:
// echo $name;

// Safe:
echo htmlspecialchars($name, ENT_QUOTES | ENT_HTML5, 'UTF-8');
// &lt;script&gt;alert(1)&lt;/script&gt;

htmlentities()와 htmlspecialchars() 비교

<?php
$str = '<b>caf&eacute;</b> & "bar"';

// htmlspecialchars: only & < > ' "
echo htmlspecialchars($str, ENT_QUOTES, 'UTF-8');

// htmlentities: also converts é → &eacute;
echo htmlentities($str, ENT_QUOTES, 'UTF-8');
// Prefer htmlspecialchars — htmlentities can break UTF-8

filter_var FILTER_SANITIZE_*

<?php
$rawEmail = 'user(bad)@exam<ple.com';
$clean = filter_var($rawEmail, FILTER_SANITIZE_EMAIL);
echo $clean; // user@example.com

$rawUrl = 'http://exa mple.com/pa th?q=foo bar';
$cleanUrl = filter_var($rawUrl, FILTER_SANITIZE_URL);
echo $cleanUrl;

strip_tags()

<?php
$input = '<p>Hello <b>World</b> <script>alert(1)</script></p>';

// Remove all tags:
echo strip_tags($input);
// Hello World alert(1)

// Keep only <b> and <i>:
echo strip_tags($input, '<b><i>');
// Hello <b>World</b>

SQL을 위한 정제

<?php
$pdo = new PDO('mysql:host=localhost;dbname=test', 'user', 'pass');

// WRONG — even with addslashes this is dangerous:
// $sql = "SELECT * FROM users WHERE email = '" . addslashes($email) . "'";

// CORRECT — prepared statement:
$stmt = $pdo->prepare('SELECT * FROM users WHERE email = :email');
$stmt->execute([':email' => $email]);
$user = $stmt->fetch();

addslashes()와 준비된 문 비교

<?php
// For embedding PHP string into JavaScript (still escape via json_encode):
$jsData = json_encode(['name' => $userInput], JSON_HEX_QUOT);
echo "<script>var data = $jsData;</script>";

파일 이름 정제

<?php
function sanitizeFilename(string $name): string {
    // Remove path traversal chars and other dangerous ones
    $name = basename($name);                          // strip any path
    $name = preg_replace('/[^a-zA-Z0-9._-]/', '_', $name); // only safe chars
    $name = preg_replace('/_{2,}/', '_', $name);      // collapse underscores
    return strtolower($name);
}

echo sanitizeFilename('../../../etc/passwd');   // etc_passwd
echo sanitizeFilename('My Photo (1).JPG');      // my_photo_1_.jpg

식별자를 위한 정수 캐스팅

<?php
// Never do this:
$id = $_GET['id'];
$sql = "SELECT * FROM posts WHERE id = $id";

// Safe approach 1: cast to int
$id = (int) ($_GET['id'] ?? 0);
if ($id <= 0) die('Invalid ID');

// Safe approach 2: validate
$id = filter_input(INPUT_GET, 'id', FILTER_VALIDATE_INT);
if (!$id) die('Invalid ID');

출력 문맥이 중요합니다

urlencode 및 rawurlencode

<?php
$search = 'hello world & PHP';

echo urlencode($search);
// hello+world+%26+PHP  (+ for spaces)

echo rawurlencode($search);
// hello%20world%20%26%20PHP  (RFC 3986 — use in paths)

$url = 'https://example.com/search?q=' . urlencode($search);
echo $url;

빠른 확인

복습: 입력 정제