PHP RFC: New Curl URL API

• Version: 0.9
• Date: 2022-06-21
• Author: Pierrick Charron pierrick@php.net
• Status: Under discussion
• First Published at: https://wiki.php.net/rfc/curl-url-api

Since its version 7.62.0 [1], libcurl features a brand new URL API [2] that can be used to parse and generate URLs, using libcurl’s own parser. One of the goal of this API is to tighten a problematic vulnerable area for applications where the URL parser library would believe one thing and libcurl another. This could and has sometimes led to security problems [3].

The current RFC proposes the addition of 2 new classes CurlUrl and CurlUrlException. These two classes will only exist if the version of libcurl installed on the system is greater than or equal to 7.62. If the version is older, these 2 classes will not exist.

One new Curl option will also be available CURLOPT_CURLU. This option will tell curl to work with this URL and will overwrite the CURLOPT_URL option.

To avoid confusing behavior where the CurlUrl object would be unintentionally modified after being attached to a CurlHandle, the CurlUrl object will be immutable

$url = new CurlUrl("https://www.php.net");
$ch  = curl_init();
curl_setopt($ch, CURLOPT_CURLU, $url);
curl_exec($ch);
curl_close($ch);

/* libcurl >= 7.62.0 */
final class CurlUrl implements Stringable
{
    public const APPEND_QUERY = UNKNOWN;
    public const DEFAULT_PORT = UNKNOWN;
    public const DEFAULT_SCHEME = UNKNOWN;
    public const DISALLOW_USER = UNKNOWN;
    public const GUESS_SCHEME = UNKNOWN;
    public const NO_DEFAULT_PORT = UNKNOWN;
    public const ALLOW_UNSUPPORTED_SCHEME = UNKNOWN;
    public const PATH_AS_IS = UNKNOWN;
    public const URL_DECODE = UNKNOWN;
    public const URL_ENCODE = UNKNOWN;
 
    /* libcurl >= 7.67.0 */
    public const NO_AUTHORITY = UNKNOWN;
 
    /* libcurl >= 7.78.0 */
    public const ALLOW_SPACE = UNKNOWN;
 
    public function __construct(?string $url = null, int $flags = 0) {}
    public function get(int $flags = 0): string {}
    public function with(string $url, int $flags = 0): CurlUrl {}
    public function getHost(): ?string {}
    public function withHost(?string $host): CurlUrl {}
    public function getScheme(): ?string {}
    public function withScheme(?string $scheme, int $flags = 0): CurlUrl {}
    public function getPort(int $flags = 0): ?int {}
    public function withPort(?int $port): CurlUrl {}
    public function getPath(int $flags = 0): string {}
    public function withPath(?string $scheme, int $flags = 0): CurlUrl {}
    public function getQuery(int $flags = 0): ?string {}
    public function withQuery(?string $query, int $flags = 0): CurlUrl {}
    public function getFragment(int $flags = 0): ?string {}
    public function withFragment(?string $fragment, int $flags = 0): CurlUrl {}
    public function getUser(int $flags = 0): ?string {}
    public function withUser(?string $user, int $flags = 0): CurlUrl {}
    public function getPassword(int $flags = 0): ?string {}
    public function withPassword(?string $password, int $flags = 0): CurlUrl {}
    public function getOptions(int $flags = 0): ?string {}
    public function withOptions(?string $options, int $flags = 0): CurlUrl {}
    public function __toString(): string {}
 
    /* libcurl >= 7.65.0 */
    public function getZoneId(int $flags = 0): ?string {}
    public function withZoneId(?string $zoneid, int $flags = 0): CurlUrl {}
}

Create a new CurlUrl object.

Return the full, normalized, and possibly cleaned up URL version of what was previously parsed.

Return a new CurlUrl object with the new URL. If the source object is already populated with a URL, the new URL can be relative to the previous.

Get the Host part of the URL.

Return a new CurlUrl object with the host par set to the given host. If it is IDNA, the string must then be encoded as your locale says or UTF-8. If it is a bracketed IPv6 numeric address, it may contain a zone id (or you can use CurlUrl::withZoneid()).

Get the scheme part of the URL.

Return a new CurlUrl object with the scheme part set to the given scheme. Libcurl only accepts setting schemes up to 40 bytes long.

Get the port part of the URL.

Return a new CurlUrl object with the port set to the given one. The given port number must be between 1 and 65535. Anything else will throw an exception.

Get the path part of the URL.

Return a new CurlUrl object with the path set to the given path. If a path is set in the URL without a leading slash, a slash will be inserted automatically when this URL is read from the handle.

Get the query part of the URL.

Return a new CurlUrl object with the query set to the given query. The question mark in the URL is not part of the actual query contents.

Get the fragment part of the URL. The hash sign in the URL is not part of the actual fragment contents.

Return a new CurlUrl object with the fragment set to the given fragment.

Get the user part of the URL.

Return a new CurlUrl object with the user set to the given user.

Get the password part of the URL.

Return a new CurlUrl object with the password set to the given password.

Get the options part of the URL.

Return a new CurlUrl object with the options set to the given options.

Get the zone id part of the URL.

Return a new CurlUrl object with the zoneid set to the given zoneid.

Same as calling CurlUrl::get()

The CurlUrlException class represents an error raised by libcurl. The constants exposed in this class are all the code that CurlUrlException::getCode() could return. Those code are internally mapped to CURLUE_* error codes that libcurl could raise. Those constants may vary depending on the version of libcurl ext/curl was compiled with.

If ext/curl was compiled with libcurl > 7.80 then CurlUrlException::getMessage() will return a user friendly message that will discriber the problem. (Exemple: Malformed input to a URL function).

/* libcurl >= 7.62.0 */
final class CurlUrlException extends Exception
{
    public const BAD_PORT_NUMBER = UNKNOWN;
    public const MALFORMED_INPUT = UNKNOWN;
    public const OUT_OF_MEMORY = UNKNOWN;
    public const UNSUPPORTED_SCHEME = UNKNOWN;
    public const URL_DECODING_FAILED = UNKNOWN;
    public const USER_NOT_ALLOWED = UNKNOWN;
 
    /* libcurl >= 7.81.0 */
    public const BAD_FILE_URL = UNKNOWN;
    public const BAD_FRAGMENT = UNKNOWN;
    public const BAD_HOSTNAME = UNKNOWN;
    public const BAD_IPV6 = UNKNOWN;
    public const BAD_LOGIN = UNKNOWN;
    public const BAD_PASSWORD = UNKNOWN;
    public const BAD_PATH = UNKNOWN;
    public const BAD_QUERY = UNKNOWN;
    public const BAD_SCHEME = UNKNOWN;
    public const BAD_SLASHES = UNKNOWN;
    public const BAD_USER = UNKNOWN;
}

None, except that the class names CurlUrl and CurlException will be declared by PHP and conflict with applications declaring the same class names in the global namespace.

8.2

As per the voting RFC a yes/no vote with a 2/3 majority is needed for this proposal to be accepted.

https://github.com/php/php-src/pull/8770

N/A

1. https://daniel.haxx.se/blog/2018/10/31/curl-7-62-0-moar-stuff/
2. https://daniel.haxx.se/blog/2018/09/09/libcurl-gets-a-url-api/
3. https://www.blackhat.com/docs/us-17/thursday/us-17-Tsai-A-New-Era-Of-SSRF-Exploiting-URL-Parser-In-Trending-Programming-Languages.pdf