Question

How do I determine whether a given string is a valid Windows filename? I’m thinking of some function that I can give a string and that returns a boolean. It should check for disallowed characters (<>:”/\|?*) and for reserved words (CON, PRN, et cetera).

isValidWindowsFilename('readme.txt'); // true
isValidWindowsFilename('foo/bar'); // false
isValidWindowsFilename('CON'); // false

I’ve found a MSDN reference describing exactly what is valid: http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247(v=vs.85).aspx

I’ve also found the same question for Java, with an answer that does what I need, except that it’s Java and not PHP: Validate a file name on Windows

Answer 1

Here’s the answer on the Java question, ported to PHP.

/**
 * @param string $filename
 * @return boolean Whether the string is a valid Windows filename.
 */
function isValidWindowsFilename($filename) {
    $regex = <<<'EOREGEX'
~                               # start of regular expression
^                               # Anchor to start of string.
(?!                             # Assert filename is not: CON, PRN, AUX, NUL, COM1, COM2, COM3, COM4, COM5, COM6, COM7, COM8, COM9, LPT1, LPT2, LPT3, LPT4, LPT5, LPT6, LPT7, LPT8, and LPT9.
    (CON|PRN|AUX|NUL|COM[1-9]|LPT[1-9])
    (\.[^.]*)?                  # followed by optional extension
    $                           # and end of string
)                               # End negative lookahead assertion.
[^<>:"/\\|?*\x00-\x1F]*         # Zero or more valid filename chars.
[^<>:"/\\|?*\x00-\x1F\ .]       # Last char is not a space or dot.
$                               # Anchor to end of string.
                                #
                                # tilde = end of regular expression.
                                # i = pattern modifier PCRE_CASELESS. Make the match case insensitive.
                                # x = pattern modifier PCRE_EXTENDED. Allows these comments inside the regex.
                                # D = pattern modifier PCRE_DOLLAR_ENDONLY. A dollar should not match a newline if it is the final character.
~ixD
EOREGEX;

    return preg_match($regex, $filename) === 1;
}

Note that this function does not impose any limit on the length of the filename, but a real filename may be limited to 260 or 32767 chars depending on the platform.

Here’s some test code to verify its correctness.

$tests = array(
    // Valid filenames
    'readme.txt' => true,
    'foo.AUX' => true,
    'foo.AUX.txt' => true,
    '.gitignore' => true, // starting with a period is ok.

    // Invalid filenames
    'x<y' => false, // less than not allowed.
    'x>y' => false, // greater than not allowed.
    'q: why not.txt' => false, // colon not allowed.
    'He said "hi".doc' => false, // double quote not allowed.
    'foo/bar' => false, // Forward slash not allowed.
    'foo\\bar' => false, // Backslash not allowed.
    'cat readme | backup' => false, // vertical bar (pipe) not allowed.
    'ls foo?.rtf' => false, // question mark not allowed
    'ls foo*.rtf' => false, // asterisk not allowed
    "null\0char" => false, // null character not allowed
    '.' => false, // must not end in period
    '..' => false, // must not end in period
    'period.' => false, // must not end in period
    'space ' => false, // must not end with space

    // Do not use the following reserved device names for the name of a file:
    // CON, PRN, AUX, NUL, COM1, COM2, COM3, COM4, COM5, COM6, COM7, COM8, COM9, LPT1, LPT2, LPT3, LPT4, LPT5, LPT6, LPT7, LPT8, and LPT9.
    // Also avoid these names followed immediately by an extension; for example, NUL.txt is not recommended.
    'CON' => false,
    'prn.txt' => false,
    'LPT9.php' => false,
);

// Disallow characters whose integer representations are in the range from 1 through 31
for ($i = 1; $i <= 31; $i++) {
    $tests[chr($i)] = false;
}
?>

<style>
.pass {
    background-color: #efe;
}
.fail {
    background-color: #fee;
}
</style>

<table>
<thead><tr><th>filename</th><th>expected</th><th>actual</th></tr></thead>
<tbody>
<? foreach ($tests as $filename => $expected) {
    $actual = isValidWindowsFilename($filename);
    ?>
    <tr><td><?=htmlentities($filename)?></td><td><?= $expected ? 'valid' : 'invalid' ?></td><td class="<?= $expected === $actual ? 'pass' : 'fail' ?>"><?= $actual ? 'valid' : 'invalid' ?></td></tr>
    <?
} ?>
</tbody>
</table>