URL

Added in: v0.10.0

源代码: lib/url.js

The node:url module provides utilities for URL resolution and parsing. It can be accessed using:

MJScopy

CJScopy

URL strings and URL objects

A URL string is a structured string containing multiple meaningful components. When parsed, a URL object is returned containing properties for each of these components.

The node:url module provides two APIs for working with URLs: a legacy API that is Node.js specific, and a newer API that implements the same WHATWG URL Standard used by web browsers.

A comparison between the WHATWG and Legacy APIs is provided below. Above the URL 'https://user:pass@sub.example.com:8080/p/a/t/h?query=string#hash', properties of an object returned by the legacy url.parse() are shown. Below it are properties of a WHATWG URL object.

WHATWG URL's origin property includes protocol and host, but not username or password.

TEXTcopy

Parsing the URL string using the WHATWG API:

JScopy

Parsing the URL string using the Legacy API:

MJScopy

CJScopy

Constructing a URL from component parts and getting the constructed string

It is possible to construct a WHATWG URL from component parts using either the property setters or a template literal string:

JScopy

JScopy

To get the constructed URL string, use the href property accessor:

JScopy

The WHATWG URL API

C URL

Browser-compatible URL class, implemented by following the WHATWG URL Standard. Examples of parsed URLs may be found in the Standard itself. The URL class is also available on the global object.

In accordance with browser conventions, all properties of URL objects are implemented as getters and setters on the class prototype, rather than as data properties on the object itself. Thus, unlike legacy urlObjects, using the delete keyword on any properties of URL objects (e.g. delete myURL.protocol, delete myURL.pathname, etc) has no effect but will still return true.

M new URL(input[, base])

• input string The absolute or relative input URL to parse. If input is relative, then base is required. If input is absolute, the base is ignored. If input is not a string, it is converted to a string first.
• base string The base URL to resolve against if the input is not absolute. If base is not a string, it is converted to a string first.

Creates a new URL object by parsing the input relative to the base. If base is passed as a string, it will be parsed equivalent to new URL(base).

JScopy

The URL constructor is accessible as a property on the global object. It can also be imported from the built-in url module:

MJScopy

CJScopy

A TypeError will be thrown if the input or base are not valid URLs. Note that an effort will be made to coerce the given values into strings. For instance:

JScopy

Unicode characters appearing within the host name of input will be automatically converted to ASCII using the Punycode algorithm.

JScopy

This feature is only available if the node executable was compiled with ICU enabled. If not, the domain names are passed through unchanged.

In cases where it is not known in advance if input is an absolute URL and a base is provided, it is advised to validate that the origin of the URL object is what is expected.

JScopy

M url.hash

• string

Gets and sets the fragment portion of the URL.

JScopy

Invalid URL characters included in the value assigned to the hash property are percent-encoded. The selection of which characters to percent-encode may vary somewhat from what the url.parse() and url.format() methods would produce.

M url.host

• string

Gets and sets the host portion of the URL.

JScopy

Invalid host values assigned to the host property are ignored.

M url.hostname

• string

Gets and sets the host name portion of the URL. The key difference between url.host and url.hostname is that url.hostname does not include the port.

JScopy

Invalid host name values assigned to the hostname property are ignored.

M url.href

• string

Gets and sets the serialized URL.

JScopy

Getting the value of the href property is equivalent to calling url.toString().

Setting the value of this property to a new value is equivalent to creating a new URL object using new URL(value). Each of the URL object's properties will be modified.

If the value assigned to the href property is not a valid URL, a TypeError will be thrown.

M url.origin

• string

Gets the read-only serialization of the URL's origin.

JScopy

JScopy

M url.password

• string

Gets and sets the password portion of the URL.

JScopy

Invalid URL characters included in the value assigned to the password property are percent-encoded. The selection of which characters to percent-encode may vary somewhat from what the url.parse() and url.format() methods would produce.

M url.pathname

• string

Gets and sets the path portion of the URL.

JScopy

Invalid URL characters included in the value assigned to the pathname property are percent-encoded. The selection of which characters to percent-encode may vary somewhat from what the url.parse() and url.format() methods would produce.

M url.port

• string

Gets and sets the port portion of the URL.

The port value may be a number or a string containing a number in the range 0 to 65535 (inclusive). Setting the value to the default port of the URL objects given protocol will result in the port value becoming the empty string ('').

The port value can be an empty string in which case the port depends on the protocol/scheme:

Upon assigning a value to the port, the value will first be converted to a string using .toString().

If that string is invalid but it begins with a number, the leading number is assigned to port. If the number lies outside the range denoted above, it is ignored.

JScopy

Numbers which contain a decimal point, such as floating-point numbers or numbers in scientific notation, are not an exception to this rule. Leading numbers up to the decimal point will be set as the URL's port, assuming they are valid:

JScopy

M url.protocol

• string

Gets and sets the protocol portion of the URL.

JScopy

Invalid URL protocol values assigned to the protocol property are ignored.

Special schemes

The WHATWG URL Standard considers a handful of URL protocol schemes to be special in terms of how they are parsed and serialized. When a URL is parsed using one of these special protocols, the url.protocol property may be changed to another special protocol but cannot be changed to a non-special protocol, and vice versa.

For instance, changing from http to https works:

JScopy

However, changing from http to a hypothetical fish protocol does not because the new protocol is not special.

JScopy

Likewise, changing from a non-special protocol to a special protocol is also not permitted:

JScopy

According to the WHATWG URL Standard, special protocol schemes are ftp, file, http, https, ws, and wss.

M url.search

• string

Gets and sets the serialized query portion of the URL.

JScopy

Any invalid URL characters appearing in the value assigned the search property will be percent-encoded. The selection of which characters to percent-encode may vary somewhat from what the url.parse() and url.format() methods would produce.

M url.searchParams

• URLSearchParams

Gets the URLSearchParams object representing the query parameters of the URL. This property is read-only but the URLSearchParams object it provides can be used to mutate the URL instance; to replace the entirety of query parameters of the URL, use the url.search setter. See URLSearchParams documentation for details.

Use care when using .searchParams to modify the URL because, per the WHATWG specification, the URLSearchParams object uses different rules to determine which characters to percent-encode. For instance, the URL object will not percent encode the ASCII tilde (~) character, while URLSearchParams will always encode it:

JScopy

M url.username

• string

Gets and sets the username portion of the URL.

JScopy

Any invalid URL characters appearing in the value assigned the username property will be percent-encoded. The selection of which characters to percent-encode may vary somewhat from what the url.parse() and url.format() methods would produce.

M url.toString()

• Returns: string

The toString() method on the URL object returns the serialized URL. The value returned is equivalent to that of url.href and url.toJSON().

M url.toJSON()

• Returns: string

The toJSON() method on the URL object returns the serialized URL. The value returned is equivalent to that of url.href and url.toString().

This method is automatically called when an URL object is serialized with JSON.stringify().

JScopy

M URL.createObjectURL(blob)

Added in: v16.7.0

• blob Blob
• Returns: string

Creates a 'blob:nodedata:...' URL string that represents the given Blob object and can be used to retrieve the Blob later.

JScopy

The data stored by the registered Blob will be retained in memory until URL.revokeObjectURL() is called to remove it.

Blob objects are registered within the current thread. If using Worker Threads, Blob objects registered within one Worker will not be available to other workers or the main thread.

M URL.revokeObjectURL(id)

Added in: v16.7.0

• id string A 'blob:nodedata:... URL string returned by a prior call to URL.createObjectURL().

Removes the stored Blob identified by the given ID. Attempting to revoke a ID that isn't registered will silently fail.

C URLSearchParams

The URLSearchParams API provides read and write access to the query of a URL. The URLSearchParams class can also be used standalone with one of the four following constructors. The URLSearchParams class is also available on the global object.

The WHATWG URLSearchParams interface and the querystring module have similar purpose, but the purpose of the querystring module is more general, as it allows the customization of delimiter characters (& and =). On the other hand, this API is designed purely for URL query strings.

JScopy

M new URLSearchParams()

Instantiate a new empty URLSearchParams object.

M new URLSearchParams(string)

• string string A query string

Parse the string as a query string, and use it to instantiate a new URLSearchParams object. A leading '?', if present, is ignored.

JScopy

M new URLSearchParams(obj)

Added in: v7.10.0, v6.13.0

• obj Object An object representing a collection of key-value pairs

Instantiate a new URLSearchParams object with a query hash map. The key and value of each property of obj are always coerced to strings.

Unlike querystring module, duplicate keys in the form of array values are not allowed. Arrays are stringified using array.toString(), which simply joins all array elements with commas.

JScopy

M new URLSearchParams(iterable)

Added in: v7.10.0, v6.13.0

• iterable Iterable An iterable object whose elements are key-value pairs

Instantiate a new URLSearchParams object with an iterable map in a way that is similar to Map's constructor. iterable can be an Array or any iterable object. That means iterable can be another URLSearchParams, in which case the constructor will simply create a clone of the provided URLSearchParams. Elements of iterable are key-value pairs, and can themselves be any iterable object.

Duplicate keys are allowed.

JScopy

M urlSearchParams.append(name, value)

• name string
• value string

Append a new name-value pair to the query string.

M urlSearchParams.delete(name)

• name string

Remove all name-value pairs whose name is name.

M urlSearchParams.entries()

• Returns: Iterator

Returns an ES6 Iterator over each of the name-value pairs in the query. Each item of the iterator is a JavaScript Array. The first item of the Array is the name, the second item of the Array is the value.

Alias for urlSearchParams[@@iterator]().

M urlSearchParams.forEach(fn[, thisArg])

• fn Function Invoked for each name-value pair in the query
• thisArg Object To be used as this value for when fn is called

Iterates over each name-value pair in the query and invokes the given function.

JScopy

M urlSearchParams.get(name)

• name string
• Returns: string or null if there is no name-value pair with the given name.

Returns the value of the first name-value pair whose name is name. If there are no such pairs, null is returned.

M urlSearchParams.getAll(name)

• name string
• Returns: string[]

Returns the values of all name-value pairs whose name is name. If there are no such pairs, an empty array is returned.

M urlSearchParams.has(name)

• name string
• Returns: boolean

Returns true if there is at least one name-value pair whose name is name.

M urlSearchParams.keys()

• Returns: Iterator

Returns an ES6 Iterator over the names of each name-value pair.

JScopy

M urlSearchParams.set(name, value)

• name string
• value string

Sets the value in the URLSearchParams object associated with name to value. If there are any pre-existing name-value pairs whose names are name, set the first such pair's value to value and remove all others. If not, append the name-value pair to the query string.

JScopy

M urlSearchParams.sort()

Added in: v7.7.0, v6.13.0

Sort all existing name-value pairs in-place by their names. Sorting is done with a stable sorting algorithm, so relative order between name-value pairs with the same name is preserved.

This method can be used, in particular, to increase cache hits.

JScopy

M urlSearchParams.toString()

• Returns: string

Returns the search parameters serialized as a string, with characters percent-encoded where necessary.

M urlSearchParams.values()

• Returns: Iterator

Returns an ES6 Iterator over the values of each name-value pair.

M urlSearchParams[Symbol.iterator]()

• Returns: Iterator

Returns an ES6 Iterator over each of the name-value pairs in the query string. Each item of the iterator is a JavaScript Array. The first item of the Array is the name, the second item of the Array is the value.

Alias for urlSearchParams.entries().

JScopy

M url.domainToASCII(domain)

Added in: v7.4.0, v6.13.0

• domain string
• Returns: string

Returns the Punycode ASCII serialization of the domain. If domain is an invalid domain, the empty string is returned.

It performs the inverse operation to url.domainToUnicode().

This feature is only available if the node executable was compiled with ICU enabled. If not, the domain names are passed through unchanged.

MJScopy

CJScopy

M url.domainToUnicode(domain)

Added in: v7.4.0, v6.13.0

• domain string
• Returns: string

Returns the Unicode serialization of the domain. If domain is an invalid domain, the empty string is returned.

It performs the inverse operation to url.domainToASCII().

This feature is only available if the node executable was compiled with ICU enabled. If not, the domain names are passed through unchanged.

MJScopy

CJScopy

M url.fileURLToPath(url)

Added in: v10.12.0

• url URL | string The file URL string or URL object to convert to a path.
• Returns: string The fully-resolved platform-specific Node.js file path.

This function ensures the correct decodings of percent-encoded characters as well as ensuring a cross-platform valid absolute path string.

MJScopy

CJScopy

M url.format(URL[, options])

Added in: v7.6.0

• URL URL A WHATWG URL object
• options Object
  • auth boolean true if the serialized URL string should include the username and password, false otherwise. Default: true.
  • fragment boolean true if the serialized URL string should include the fragment, false otherwise. Default: true.
  • search boolean true if the serialized URL string should include the search query, false otherwise. Default: true.
  • unicode boolean true if Unicode characters appearing in the host component of the URL string should be encoded directly as opposed to being Punycode encoded. Default: false.
• Returns: string

Returns a customizable serialization of a URL String representation of a WHATWG URL object.

The URL object has both a toString() method and href property that return string serializations of the URL. These are not, however, customizable in any way. The url.format(URL[, options]) method allows for basic customization of the output.

MJScopy

CJScopy

M url.pathToFileURL(path)

Added in: v10.12.0

• path string The path to convert to a File URL.
• Returns: URL The file URL object.

This function ensures that path is resolved absolutely, and that the URL control characters are correctly encoded when converting into a File URL.

MJScopy

CJScopy

M url.urlToHttpOptions(url)

Added in: v15.7.0

• url URL The WHATWG URL object to convert to an options object.
• Returns: Object Options object
  • protocol string Protocol to use.
  • hostname string A domain name or IP address of the server to issue the request to.
  • hash string The fragment portion of the URL.
  • search string The serialized query portion of the URL.
  • pathname string The path portion of the URL.
  • path string Request path. Should include query string if any. E.G. '/index.html?page=12'. An exception is thrown when the request path contains illegal characters. Currently, only spaces are rejected but that may change in the future.
  • href string The serialized URL.
  • port number Port of remote server.
  • auth string Basic authentication i.e. 'user:password' to compute an Authorization header.

This utility function converts a URL object into an ordinary options object as expected by the http.request() and https.request() APIs.

MJScopy

CJScopy

Legacy URL API

Legacy urlObject

The legacy urlObject (require('node:url').Url or import { Url } from 'node:url') is created and returned by the url.parse() function.

M urlObject.auth

The auth property is the username and password portion of the URL, also referred to as userinfo. This string subset follows the protocol and double slashes (if present) and precedes the host component, delimited by @. The string is either the username, or it is the username and password separated by :.

For example: 'user:pass'.

M urlObject.hash

The hash property is the fragment identifier portion of the URL including the leading # character.

For example: '#hash'.

M urlObject.host

The host property is the full lower-cased host portion of the URL, including the port if specified.

For example: 'sub.example.com:8080'.

M urlObject.hostname

The hostname property is the lower-cased host name portion of the host component without the port included.

For example: 'sub.example.com'.

M urlObject.href

The href property is the full URL string that was parsed with both the protocol and host components converted to lower-case.

For example: 'http://user:pass@sub.example.com:8080/p/a/t/h?query=string#hash'.

M urlObject.path

The path property is a concatenation of the pathname and search components.

For example: '/p/a/t/h?query=string'.

No decoding of the path is performed.

M urlObject.pathname

The pathname property consists of the entire path section of the URL. This is everything following the host (including the port) and before the start of the query or hash components, delimited by either the ASCII question mark (?) or hash (#) characters.

For example: '/p/a/t/h'.

No decoding of the path string is performed.

M urlObject.port

The port property is the numeric port portion of the host component.

For example: '8080'.

M urlObject.protocol

The protocol property identifies the URL's lower-cased protocol scheme.

For example: 'http:'.

M urlObject.query

The query property is either the query string without the leading ASCII question mark (?), or an object returned by the querystring module's parse() method. Whether the query property is a string or object is determined by the parseQueryString argument passed to url.parse().

For example: 'query=string' or {'query': 'string'}.

If returned as a string, no decoding of the query string is performed. If returned as an object, both keys and values are decoded.

M urlObject.search

The search property consists of the entire "query string" portion of the URL, including the leading ASCII question mark (?) character.

For example: '?query=string'.

No decoding of the query string is performed.

M urlObject.slashes

The slashes property is a boolean with a value of true if two ASCII forward-slash characters (/) are required following the colon in the protocol.

M url.format(urlObject)

• urlObject Object | string A URL object (as returned by url.parse() or constructed otherwise). If a string, it is converted to an object by passing it to url.parse().

The url.format() method returns a formatted URL string derived from urlObject.

JScopy

If urlObject is not an object or a string, url.format() will throw a TypeError.

The formatting process operates as follows:

• A new empty string result is created.
• If urlObject.protocol is a string, it is appended as-is to result.
• Otherwise, if urlObject.protocol is not undefined and is not a string, an Error is thrown.
• For all string values of urlObject.protocol that do not end with an ASCII colon (:) character, the literal string : will be appended to result.
• If either of the following conditions is true, then the literal string // will be appended to result:
  • urlObject.slashes property is true;
  • urlObject.protocol begins with http, https, ftp, gopher, or file;
• If the value of the urlObject.auth property is truthy, and either urlObject.host or urlObject.hostname are not undefined, the value of urlObject.auth will be coerced into a string and appended to result followed by the literal string @.
• If the urlObject.host property is undefined then:
  • If the urlObject.hostname is a string, it is appended to result.
  • Otherwise, if urlObject.hostname is not undefined and is not a string, an Error is thrown.
  • If the urlObject.port property value is truthy, and urlObject.hostname is not undefined:
    • The literal string : is appended to result, and
    • The value of urlObject.port is coerced to a string and appended to result.
• Otherwise, if the urlObject.host property value is truthy, the value of urlObject.host is coerced to a string and appended to result.
• If the urlObject.pathname property is a string that is not an empty string:
  • If the urlObject.pathname does not start with an ASCII forward slash (/), then the literal string '/' is appended to result.
  • The value of urlObject.pathname is appended to result.
• Otherwise, if urlObject.pathname is not undefined and is not a string, an Error is thrown.
• If the urlObject.search property is undefined and if the urlObject.query property is an Object, the literal string ? is appended to result followed by the output of calling the querystring module's stringify() method passing the value of urlObject.query.
• Otherwise, if urlObject.search is a string:
  • If the value of urlObject.search does not start with the ASCII question mark (?) character, the literal string ? is appended to result.
  • The value of urlObject.search is appended to result.
• Otherwise, if urlObject.search is not undefined and is not a string, an Error is thrown.
• If the urlObject.hash property is a string:
  • If the value of urlObject.hash does not start with the ASCII hash (#) character, the literal string # is appended to result.
  • The value of urlObject.hash is appended to result.
• Otherwise, if the urlObject.hash property is not undefined and is not a string, an Error is thrown.
• result is returned.

M url.parse(urlString[, parseQueryString[, slashesDenoteHost]])

• urlString string The URL string to parse.
• parseQueryString boolean If true, the query property will always be set to an object returned by the querystring module's parse() method. If false, the query property on the returned URL object will be an unparsed, undecoded string. Default: false.
• slashesDenoteHost boolean If true, the first token after the literal string // and preceding the next / will be interpreted as the host. For instance, given //foo/bar, the result would be {host: 'foo', pathname: '/bar'} rather than {pathname: '//foo/bar'}. Default: false.

The url.parse() method takes a URL string, parses it, and returns a URL object.

A TypeError is thrown if urlString is not a string.

A URIError is thrown if the auth property is present but cannot be decoded.

url.parse() uses a lenient, non-standard algorithm for parsing URL strings. It is prone to security issues such as host name spoofing and incorrect handling of usernames and passwords.

url.parse() is an exception to most of the legacy APIs. Despite its security concerns, it is legacy and not deprecated because it is:

• Faster than the alternative WHATWG URL parser.
• Easier to use with regards to relative URLs than the alternative WHATWG URL API.
• Widely relied upon within the npm ecosystem.

Use with caution.

M url.resolve(from, to)

• from string The base URL to use if to is a relative URL.
• to string The target URL to resolve.

The url.resolve() method resolves a target URL relative to a base URL in a manner similar to that of a web browser resolving an anchor tag.

JScopy

To achieve the same result using the WHATWG URL API:

JScopy

Percent-encoding in URLs

URLs are permitted to only contain a certain range of characters. Any character falling outside of that range must be encoded. How such characters are encoded, and which characters to encode depends entirely on where the character is located within the structure of the URL.

Legacy API

Within the Legacy API, spaces (' ') and the following characters will be automatically escaped in the properties of URL objects:

TEXTcopy

For example, the ASCII space character (' ') is encoded as %20. The ASCII forward slash (/) character is encoded as %3C.

WHATWG API

The WHATWG URL Standard uses a more selective and fine grained approach to selecting encoded characters than that used by the Legacy API.

The WHATWG algorithm defines four "percent-encode sets" that describe ranges of characters that must be percent-encoded:

• The C0 control percent-encode set includes code points in range U+0000 to U+001F (inclusive) and all code points greater than U+007E.

• The fragment percent-encode set includes the C0 control percent-encode set and code points U+0020, U+0022, U+003C, U+003E, and U+0060.

• The path percent-encode set includes the C0 control percent-encode set and code points U+0020, U+0022, U+0023, U+003C, U+003E, U+003F, U+0060, U+007B, and U+007D.

• The userinfo encode set includes the path percent-encode set and code points U+002F, U+003A, U+003B, U+003D, U+0040, U+005B, U+005C, U+005D, U+005E, and U+007C.

The userinfo percent-encode set is used exclusively for username and passwords encoded within the URL. The path percent-encode set is used for the path of most URLs. The fragment percent-encode set is used for URL fragments. The C0 control percent-encode set is used for host and path under certain specific conditions, in addition to all other cases.

When non-ASCII characters appear within a host name, the host name is encoded using the Punycode algorithm. Note, however, that a host name may contain both Punycode encoded and percent-encoded characters:

JScopy