whatwg-url

whatwg-url is a full implementation of the WHATWG URL Standard. It can be used standalone, but it also exposes a lot of the internal algorithms that are useful for integrating a URL parser into a project like jsdom.

Current status

whatwg-url is currently up to date with the URL spec up to commit 2eef975.

API

The URL and URLSearchParams classes

The main API is provided by the URL and URLSearchParams exports, which follows the spec's behavior in all ways (including e.g. USVString conversion). Most consumers of this library will want to use these.

Low-level URL Standard API

The following methods are exported for use by places like jsdom that need to implement things like HTMLHyperlinkElementUtils. They mostly operate on or return an "internal URL" or "URL record" type.

• URL parser: parseURL(input, { baseURL, encodingOverride })
• Basic URL parser: basicURLParse(input, { baseURL, encodingOverride, url, stateOverride })
• URL serializer: serializeURL(urlRecord, excludeFragment)
• Host serializer: serializeHost(hostFromURLRecord)
• Serialize an integer: serializeInteger(number)
• Origin serializer: serializeURLOrigin(urlRecord)
• Set the username: setTheUsername(urlRecord, usernameString)
• Set the password: setThePassword(urlRecord, passwordString)
• Cannot have a username/password/port: cannotHaveAUsernamePasswordPort(urlRecord)
• Percent decode: percentDecode(buffer)

The stateOverride parameter is one of the following strings:

• "scheme start"
• "scheme"
• "no scheme"
• "special relative or authority"
• "path or authority"
• "relative"
• "relative slash"
• "special authority slashes"
• "special authority ignore slashes"
• "authority"
• "host"
• "hostname"
• "port"
• "file"
• "file slash"
• "file host"
• "path start"
• "path"
• "cannot-be-a-base-URL path"
• "query"
• "fragment"

The URL record type has the following API:

• scheme
• username
• password
• host
• port
• path (as an array)
• query
• fragment
• cannotBeABaseURL (as a boolean)

These properties should be treated with care, as in general changing them will cause the URL record to be in an inconsistent state until the appropriate invocation of basicURLParse is used to fix it up. You can see examples of this in the URL Standard, where there are many step sequences like "4. Set context object’s url’s fragment to the empty string. 5. Basic URL parse input with context object’s url as url and fragment state as state override." In between those two steps, a URL record is in an unusable state.

The return value of "failure" in the spec is represented by null. That is, functions like parseURL and basicURLParse can return either a URL record or null.