Add URL-to-path and clean path conversion

This commit is contained in:
Kijin Sung 2016-03-15 13:48:07 +09:00
parent 86e91d116e
commit 9d2fe0270b
4 changed files with 140 additions and 8 deletions

View file

@ -46,4 +46,24 @@ class FilenameFilter
return $filename;
}
/**
* Clean a path to remove ./, ../, trailing slashes, etc.
*
* @param string $path
* @return string
*/
public static function cleanPath($path)
{
$path = str_replace('\\', '/', $path);
$path = preg_replace('@[\?#].+$@', '', $path);
$path = preg_replace('@/{2,}@', '/', $path);
$path = preg_replace('@/\.{3,}/@', '/', $path);
$path = preg_replace('@/(\./)+@', '/', $path);
while (preg_match('@/[^/]+/\.\.(?:/|$)@', $path, $matches))
{
$path = str_replace($matches[0], '/', $path);
}
return rtrim($path, '/');
}
}

View file

@ -18,10 +18,7 @@ class URL
*/
public static function getCurrentURL(array $changes = array())
{
$proto = \RX_SSL ? 'https://' : 'http://';
$host = isset($_SERVER['HTTP_HOST']) ? $_SERVER['HTTP_HOST'] : 'localhost';
$local = isset($_SERVER['REQUEST_URI']) ? $_SERVER['REQUEST_URI'] : '/';
$url = $proto . $host . $local;
$url = self::getCurrentDomainURL(isset($_SERVER['REQUEST_URI']) ? $_SERVER['REQUEST_URI'] : '/');
if (count($changes))
{
return self::modifyURL($url, $changes);
@ -32,6 +29,19 @@ class URL
}
}
/**
* Get the current domain.
*
* @param string $path
* @return string
*/
public static function getCurrentDomainURL($path = '/')
{
$proto = \RX_SSL ? 'https://' : 'http://';
$host = isset($_SERVER['HTTP_HOST']) ? $_SERVER['HTTP_HOST'] : 'localhost';
return $proto . $host . '/' . ltrim($path, '/');
}
/**
* Convert a URL to its canonical format.
*
@ -42,9 +52,7 @@ class URL
{
if (preg_match('#^\.?/([^/]|$)#', $url) || !preg_match('#^(https?:|/)#', $url))
{
$proto = \RX_SSL ? 'https://' : 'http://';
$host = isset($_SERVER['HTTP_HOST']) ? $_SERVER['HTTP_HOST'] : 'localhost';
$url = $proto . $host . \RX_BASEURL . ltrim($url, './');
$url = self::getCurrentDomainURL(\RX_BASEURL . ltrim($url, './'));
}
return preg_replace_callback('#^(https?:|)//([^/]+)#i', function($matches) {
if ($matches[1] === '') $matches[1] = \RX_SSL ? 'https:' : 'http:';
@ -125,6 +133,53 @@ class URL
}
}
/**
* Convert a server-side path to a URL.
*
* This method returns false if the path cannot be converted to a URL,
* e.g. if the path is outside of the document root.
*
* @param string $path
* @return string|false
*/
public static function fromServerPath($path)
{
$cleanpath = Filters\FilenameFilter::cleanPath($path);
if (substr($path, -1) === '/')
{
$cleanpath .= '/';
}
$root = Filters\FilenameFilter::cleanPath($_SERVER['DOCUMENT_ROOT']);
if ($cleanpath === $root)
{
return self::getCurrentDomainURL('/');
}
if (starts_with($root . '/', $cleanpath))
{
return self::getCurrentDomainURL(substr($cleanpath, strlen($root)));
}
return false;
}
/**
* Convert a URL to a server-side path.
*
* This method returns false if the URL cannot be converted to a server-side path,
* e.g. if the URL belongs to an external domain.
*
* @param string $url
* @return string
*/
public static function toServerPath($url)
{
$url = self::getCanonicalURL($url);
if (!self::isInternalURL($url))
{
return false;
}
return Filters\FilenameFilter::cleanPath($_SERVER['DOCUMENT_ROOT'] . parse_url($url, \PHP_URL_PATH));
}
/**
* Encode UTF-8 domain into IDNA (punycode)
*

View file

@ -25,6 +25,17 @@ class URLTest extends \Codeception\TestCase\Test
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . '/index.php?xe=sucks&l=ko', Rhymix\Framework\URL::getCurrentURL(array('l' => 'ko', 'foo' => null)));
}
public function testGetCurrentDomainURL()
{
$protocol = \RX_SSL ? 'https://' : 'http://';
$_SERVER['HTTP_HOST'] = 'www.rhymix.org';
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . '/', Rhymix\Framework\URL::getCurrentDomainURL());
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . '/', Rhymix\Framework\URL::getCurrentDomainURL('/'));
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . '/foo/bar', Rhymix\Framework\URL::getCurrentDomainURL('/foo/bar'));
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . '/index.php?foo=bar', Rhymix\Framework\URL::getCurrentDomainURL('index.php?foo=bar'));
}
public function testGetCanonicalURL()
{
$protocol = \RX_SSL ? 'https://' : 'http://';
@ -82,6 +93,30 @@ class URLTest extends \Codeception\TestCase\Test
// This function is checked in Security::checkCSRF()
}
public function testURLFromServerPath()
{
$protocol = \RX_SSL ? 'https://' : 'http://';
$_SERVER['HTTP_HOST'] = 'www.rhymix.org';
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . \RX_BASEURL, Rhymix\Framework\URL::fromServerPath(\RX_BASEDIR));
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . \RX_BASEURL . 'index.php', Rhymix\Framework\URL::fromServerPath(\RX_BASEDIR . 'index.php'));
$this->assertEquals($protocol . $_SERVER['HTTP_HOST'] . \RX_BASEURL . 'foo/bar', Rhymix\Framework\URL::fromServerPath(\RX_BASEDIR . '/foo/bar'));
$this->assertEquals(false, Rhymix\Framework\URL::fromServerPath('C:/Windows'));
}
public function testURLToServerPath()
{
$protocol = \RX_SSL ? 'https://' : 'http://';
$_SERVER['HTTP_HOST'] = 'www.rhymix.org';
$this->assertEquals(\RX_BASEDIR . 'index.php', Rhymix\Framework\URL::toServerPath($protocol . $_SERVER['HTTP_HOST'] . \RX_BASEURL . 'index.php'));
$this->assertEquals(\RX_BASEDIR . 'foo/bar', Rhymix\Framework\URL::toServerPath($protocol . $_SERVER['HTTP_HOST'] . \RX_BASEURL . '/foo/bar?arg=baz'));
$this->assertEquals(\RX_BASEDIR . 'foo/bar', Rhymix\Framework\URL::toServerPath('./foo/bar'));
$this->assertEquals(\RX_BASEDIR . 'foo/bar', Rhymix\Framework\URL::toServerPath('foo/bar/../bar'));
$this->assertEquals(false, Rhymix\Framework\URL::toServerPath('http://other.domain.com/'));
$this->assertEquals(false, Rhymix\Framework\URL::toServerPath('//other.domain.com/'));
}
public function testEncodeIdna()
{
$this->assertEquals('xn--9i1bl3b186bf9e.xn--3e0b707e', Rhymix\Framework\URL::encodeIdna('퓨니코드.한국'));

View file

@ -1,5 +1,7 @@
<?php
use Rhymix\Framework\Filters\FilenameFilter;
class FilenameFilterTest extends \Codeception\TestCase\Test
{
public function testFilenameFilterClean()
@ -35,8 +37,28 @@ class FilenameFilterTest extends \Codeception\TestCase\Test
foreach ($tests as $from => $to)
{
$result = Rhymix\Framework\Filters\FilenameFilter::clean($from);
$result = FilenameFilter::clean($from);
$this->assertEquals($to, $result);
}
}
public function testFilenameFilterCleanPath()
{
// Remove extra dots and slashes.
$this->assertEquals('/usr/share/foo/bar.jpg', FilenameFilter::cleanPath('/usr/share/foo//./baz/../bar.jpg'));
$this->assertEquals('/usr/share/foo/bar.jpg', FilenameFilter::cleanPath('/usr/share/foo/././baz/../../foo/bar.jpg'));
$this->assertEquals('/usr/share', FilenameFilter::cleanPath('/usr/share/foo/..'));
$this->assertEquals('/usr/share', FilenameFilter::cleanPath('/usr/share/foo/bar/../baz/../../'));
// Test Windows paths.
$this->assertEquals('C:/Windows/Notepad.exe', FilenameFilter::cleanPath('C:\\Windows\\System32\\..\\Notepad.exe'));
// Do not remove .. if there is no parent directory.
$this->assertEquals('C:/../foobar', FilenameFilter::cleanPath('C:\\..\foobar\\'));
$this->assertEquals('/../foobar', FilenameFilter::cleanPath('/../foobar/'));
// Remove query strings and URL fragments.
$this->assertEquals('index.php', FilenameFilter::cleanPath('index.php?foo=bar'));
$this->assertEquals('index.php', FilenameFilter::cleanPath('index.php#baz'));
}
}