HTTP API: Add a `$component` parameter to `wp_parse_url()` to give it parity with PHP's `parse_url()` function.

Fixes #36356
Props jrf


git-svn-id: https://develop.svn.wordpress.org/trunk@38694 602fd350-edb4-49c9-b593-d223f7449a82
This commit is contained in:
John Blackbourn 2016-09-30 21:46:43 +00:00
parent 41fc5ba881
commit e0f4a218fb
2 changed files with 128 additions and 20 deletions

View File

@ -633,35 +633,68 @@ function ms_allowed_http_request_hosts( $is_external, $host ) {
* when URL parsing failed.
*
* @since 4.4.0
* @since 4.7.0 The $component parameter was added for parity with PHP's parse_url().
*
* @param string $url The URL to parse.
* @return bool|array False on failure; Array of URL components on success;
* See parse_url()'s return values.
* @param string $url The URL to parse.
* @param int $component The specific component to retrieve. Use one of the PHP
* predefined constants to specify which one.
* Defaults to -1 (= return all parts as an array).
* @see http://php.net/manual/en/function.parse-url.php
* @return mixed False on failure; Array of URL components on success;
* When a specific component has been requested: null if the component doesn't
* exist in the given URL; a sting or - in the case of PHP_URL_PORT - integer
* when it does; See parse_url()'s return values.
*/
function wp_parse_url( $url ) {
$parts = @parse_url( $url );
if ( ! $parts ) {
// < PHP 5.4.7 compat, trouble with relative paths including a scheme break in the path
function wp_parse_url( $url, $component = -1 ) {
$parts = @parse_url( $url, $component );
if ( version_compare( PHP_VERSION, '5.4.7', '>=' ) ) {
return $parts;
}
if ( false === $parts ) {
// < PHP 5.4.7 compat, trouble with relative paths including a scheme break in the path.
if ( '/' == $url[0] && false !== strpos( $url, '://' ) ) {
// Since we know it's a relative path, prefix with a scheme/host placeholder and try again
if ( ! $parts = @parse_url( 'placeholder://placeholder' . $url ) ) {
if ( in_array( $component, array( PHP_URL_SCHEME, PHP_URL_HOST ), true ) ) {
return null;
}
// Since we know it's a relative path, prefix with a scheme/host placeholder and try again.
if ( ! $parts = @parse_url( 'placeholder://placeholder' . $url, $component ) ) {
return $parts;
}
// Remove the placeholder values
unset( $parts['scheme'], $parts['host'] );
// Remove the placeholder values.
if ( -1 === $component ) {
unset( $parts['scheme'], $parts['host'] );
}
} else {
return $parts;
}
}
// < PHP 5.4.7 compat, doesn't detect schemeless URL's host field
if ( '//' == substr( $url, 0, 2 ) && ! isset( $parts['host'] ) ) {
$path_parts = explode( '/', substr( $parts['path'], 2 ), 2 );
$parts['host'] = $path_parts[0];
if ( isset( $path_parts[1] ) ) {
$parts['path'] = '/' . $path_parts[1];
} else {
unset( $parts['path'] );
// < PHP 5.4.7 compat, doesn't detect a schemeless URL's host field.
if ( '//' == substr( $url, 0, 2 ) ) {
if ( -1 === $component && ! isset( $parts['host'] ) ) {
$path_parts = explode( '/', substr( $parts['path'], 2 ), 2 );
$parts['host'] = $path_parts[0];
if ( isset( $path_parts[1] ) ) {
$parts['path'] = '/' . $path_parts[1];
} else {
unset( $parts['path'] );
}
} elseif ( PHP_URL_HOST === $component || PHP_URL_PATH === $component ) {
$all_parts = @parse_url( $url );
if ( ! isset( $all_parts['host'] ) ) {
$path_parts = explode( '/', substr( $all_parts['path'], 2 ), 2 );
if ( PHP_URL_PATH === $component ) {
if ( isset( $path_parts[1] ) ) {
$parts = '/' . $path_parts[1];
} else {
$parts = null;
}
} elseif ( PHP_URL_HOST === $component ) {
$parts = $path_parts[0];
}
}
}
}

View File

@ -6,6 +6,8 @@
*/
class Tests_HTTP_HTTP extends WP_UnitTestCase {
const FULL_TEST_URL = 'http://username:password@host.name:9090/path?arg1=value1&arg2=value2#anchor';
/**
* @dataProvider make_absolute_url_testcases
*/
@ -78,6 +80,16 @@ class Tests_HTTP_HTTP extends WP_UnitTestCase {
function parse_url_testcases() {
// 0: The URL, 1: The expected resulting structure
return array(
array( self::FULL_TEST_URL, array(
'scheme' => 'http',
'host' => 'host.name',
'port' => 9090,
'user' => 'username',
'pass' => 'password',
'path' => '/path',
'query' => 'arg1=value1&arg2=value2',
'fragment' => 'anchor',
) ),
array( 'http://example.com/', array( 'scheme' => 'http', 'host' => 'example.com', 'path' => '/' ) ),
// < PHP 5.4.7: Schemeless URL
@ -85,7 +97,7 @@ class Tests_HTTP_HTTP extends WP_UnitTestCase {
array( '//example.com/', array( 'host' => 'example.com', 'path' => '/' ) ),
array( 'http://example.com//path/', array( 'scheme' => 'http', 'host' => 'example.com', 'path' => '//path/' ) ),
// < PHP 5.4.7: Scheme seperator in the URL
// < PHP 5.4.7: Scheme separator in the URL.
array( 'http://example.com/http://example.net/', array( 'scheme' => 'http', 'host' => 'example.com', 'path' => '/http://example.net/' ) ),
array( '/path/http://example.net/', array( 'path' => '/path/http://example.net/' ) ),
@ -103,6 +115,69 @@ class Tests_HTTP_HTTP extends WP_UnitTestCase {
*/
}
/**
* @ticket 36356
*/
function test_wp_parse_url_with_default_component() {
$actual = wp_parse_url( self::FULL_TEST_URL, -1 );
$this->assertEquals( array(
'scheme' => 'http',
'host' => 'host.name',
'port' => 9090,
'user' => 'username',
'pass' => 'password',
'path' => '/path',
'query' => 'arg1=value1&arg2=value2',
'fragment' => 'anchor',
), $actual );
}
/**
* @ticket 36356
*
* @dataProvider parse_url_component_testcases
*/
function test_wp_parse_url_with_component( $url, $component, $expected ) {
$actual = wp_parse_url( $url, $component );
$this->assertSame( $expected, $actual );
}
function parse_url_component_testcases() {
// 0: The URL, 1: The requested component, 2: The expected resulting structure.
return array(
array( self::FULL_TEST_URL, PHP_URL_SCHEME, 'http' ),
array( self::FULL_TEST_URL, PHP_URL_USER, 'username' ),
array( self::FULL_TEST_URL, PHP_URL_PASS, 'password' ),
array( self::FULL_TEST_URL, PHP_URL_HOST, 'host.name' ),
array( self::FULL_TEST_URL, PHP_URL_PORT, 9090 ),
array( self::FULL_TEST_URL, PHP_URL_PATH, '/path' ),
array( self::FULL_TEST_URL, PHP_URL_QUERY, 'arg1=value1&arg2=value2' ),
array( self::FULL_TEST_URL, PHP_URL_FRAGMENT, 'anchor' ),
// < PHP 5.4.7: Schemeless URL.
array( '//example.com/path/', PHP_URL_HOST, 'example.com' ),
array( '//example.com/path/', PHP_URL_PATH, '/path/' ),
array( '//example.com/', PHP_URL_HOST, 'example.com' ),
array( '//example.com/', PHP_URL_PATH, '/' ),
array( 'http://example.com//path/', PHP_URL_HOST, 'example.com' ),
array( 'http://example.com//path/', PHP_URL_PATH, '//path/' ),
// < PHP 5.4.7: Scheme separator in the URL.
array( 'http://example.com/http://example.net/', PHP_URL_HOST, 'example.com' ),
array( 'http://example.com/http://example.net/', PHP_URL_PATH, '/http://example.net/' ),
array( '/path/http://example.net/', PHP_URL_HOST, null ),
array( '/path/http://example.net/', PHP_URL_PATH, '/path/http://example.net/' ),
// < PHP 5.4.7: IPv6 literals in schemeless URLs are handled incorrectly.
array( '//[::FFFF::127.0.0.1]/', PHP_URL_HOST, '[::FFFF::127.0.0.1]' ),
array( '//[::FFFF::127.0.0.1]/', PHP_URL_PATH, '/' ),
// PHP's parse_url() calls this an invalid URL, we handle it as a path.
array( '/://example.com/', PHP_URL_PATH, '/://example.com/' ),
);
}
/**
* @ticket 35426
*/