url.bs

<pre class=metadata>
Group: WHATWG
H1: URL
Shortname: url
Text Macro: TWITTER urlstandard
Text Macro: LATESTRD 2026-02
Abstract: The URL Standard defines URLs, domains, IP addresses, the <code>application/x-www-form-urlencoded</code> format, and their API.
Translation: ja https://triple-underscore.github.io/URL-ja.html
Translation: zh-Hans https://htmlspecs.com/url/
Translation: ko https://ko.htmlspecs.com/url/
Required IDs: application/x-www-form-urlencoded,urlencoded-parsing
</pre>

<pre class=anchors>
spec: ECMA-262; url: https://tc39.es/ecma262/#sec-encodeuricomponent-uricomponent; text: "encodeURIComponent() [sic]"; type: method
spec: UTS46; urlPrefix: https://www.unicode.org/reports/tr46/
    type: abstract-op; text: ToASCII; url: #ToASCII
    type: abstract-op; text: ToUnicode; url: #ToUnicode
</pre>



<h2 id=goals class=no-num>Goals</h2>

<p>The URL standard takes the following approach towards making URLs fully interoperable:

<ul>
 <li><p>Align RFC 3986 and RFC 3987 with contemporary implementations and
 obsolete the RFCs in the process. (E.g., spaces, other "illegal" code points,
 query encoding, equality, canonicalization, are all concepts not entirely
 shared, or defined.) URL parsing needs to become as solid as HTML parsing.
 [[RFC3986]]
 [[RFC3987]]

 <li><p>Standardize on the term URL. URI and IRI are just confusing. In
 practice a single algorithm is used for both so keeping them distinct is
 not helping anyone. URL also easily wins the
 <a href="https://trends.google.com/trends/explore?q=url,uri">search result popularity contest</a>.

 <li><p>Supplanting <a href="https://tools.ietf.org/html/rfc6454#section-4">Origin of a URI [sic]</a>.
 [[RFC6454]]

 <li><p>Define URL's existing JavaScript API in full detail and add
 enhancements to make it easier to work with. Add a new <code><a interface>URL</a></code>
 object as well for URL manipulation without usage of HTML elements. (Useful
 for JavaScript worker environments.)

 <li><p>Ensure the combination of parser, serializer, and API guarantee idempotence. For example, a
 non-failure result of a parse-then-serialize operation will not change with any further
 parse-then-serialize operations applied to it. Similarly, manipulating a non-failure result through
 the API will not change from applying any number of serialize-then-parse operations to it.
</ul>

<p class=note>As the editors learn more about the subject matter the goals
might increase in scope somewhat.



<h2 id=infrastructure>Infrastructure</h2>

<p>This specification depends on <cite>Infra</cite>. [[!INFRA]]

<p>Some terms used in this specification are defined in the following standards and specifications:

<ul class=brief>
 <li><cite>Encoding</cite> [[!ENCODING]]
 <li><cite>File API</cite> [[!FILEAPI]]
 <li><cite>HTML</cite> [[!HTML]]
 <li><cite>Unicode IDNA Compatibility Processing</cite> [[!UTS46]]
 <li><cite>Web IDL</cite> [[!WEBIDL]]
</ul>

<hr>

<p>To <dfn>serialize an integer</dfn>, represent it as the shortest possible decimal
number.


<h3 id=writing>Writing</h3>

<p>A <dfn oldids=syntax-violation>validation error</dfn> indicates a mismatch between input and
valid input. User agents, especially conformance checkers, are encouraged to report them somewhere.

<div class=note>
 <p>A <a>validation error</a> does not mean that the parser terminates. Termination of a parser is
 always stated explicitly, e.g., through a return statement.

 <p>It is useful to signal <a>validation errors</a> as error-handling can be non-intuitive, legacy
 user agents might not implement correct error-handling, and the intent of what is written might be
 unclear to other developers.
</div>

<table class=yesno>
 <thead>
  <tr>
   <th>Error type
   <th>Error description
   <th>Failure
 <!-- The rows inside the <tbody>s are generally sorted by first occurrence. However, where logical
      groupings exist those override that sorting:
      - domain- and host- stay together
      - IPv6- stays together
      - IPv4-in-IPv6- stays together -->
 <tbody>
  <tr>
   <th colspan=3 scope=rowgroup><a href=#idna>IDNA</a>
  <tr>
   <td><dfn id=validation-error-domain-to-ascii>domain-to-ASCII</dfn>
   <td>
    <p><a abstract-op lt=ToASCII>Unicode ToASCII</a> records an error or returns the empty string.
    [[UTS46]]
    <p class=note>If details about <a abstract-op lt=ToASCII>Unicode ToASCII</a> errors are
    recorded, user agents are encouraged to pass those along.
   <td class=yes>Yes
  <tr>
   <td><dfn>domain-invalid-code-point</dfn>
   <td>
    <p>The input's <a for=/>host</a> contains a <a>forbidden domain code point</a>.
    <div class=example id=example-domain-invalid-code-point>
     <p>Hosts are <a for=string>percent-decoded</a> before being processed when the URL
     <a>is special</a>, which would result in the following host portion becoming
     "<code>exa#mple.org</code>" and thus triggering this error.
     <p>"<code>https://exa%23mple.org</code>"
    </div>
   <td class=yes>Yes
  <tr>
   <td><dfn>domain-to-Unicode</dfn>
   <td>
    <p><a abstract-op lt=ToUnicode>Unicode ToUnicode</a> records an error. [[UTS46]]
    <p class=note>The same considerations as with <a>domain-to-ASCII</a> apply.
   <td class=no>·
 <tbody>
  <tr>
   <th colspan=3 scope=rowgroup><a href=#host-parsing>Host parsing</a>
  <!-- opaque-host parser -->
  <tr>
   <td><dfn>host-invalid-code-point</dfn>
   <td>
    <p>An <a>opaque host</a> (in a URL that <a>is not special</a>) contains a
    <a>forbidden host code point</a>.
    <p class=example id=example-host-invalid-code-point>"<code>foo://exa[mple.org</code>"
   <td class=yes>Yes
  <!-- IPv4 parser -->
  <tr>
   <td><dfn>IPv4-empty-part</dfn>
   <td>
    <p>An <a for=/>IPv4 address</a> ends with a U+002E (.).
    <p class=example id=example-ipv4-empty-part>"<code>https://127.0.0.1./</code>"
   <td class=no>·
  <tr>
   <td><dfn>IPv4-too-many-parts</dfn>
   <td>
    <p>An <a for=/>IPv4 address</a> does not consist of exactly 4 parts.
    <p class=example id=example-ipv4-too-many-parts>"<code>https://1.2.3.4.5/</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv4-non-numeric-part</dfn>
   <td>
    <p>An <a for=/>IPv4 address</a> part is not numeric.
    <p class=example id=example-ipv4-non-numeric-part>"<code>https://test.42</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv4-non-decimal-part</dfn>
   <td>
    <p>The <a for=/>IPv4 address</a> contains numbers expressed using hexadecimal or octal digits.
    <p class=example id=example-ipv4-non-decimal-part>"<code>https://127.0.0x0.1</code>"
   <td class=no>·
  <tr>
   <td><dfn>IPv4-out-of-range-part</dfn>
   <td>
    <p>An <a for=/>IPv4 address</a> part exceeds 255.
    <p class=example id=example-ipv4-out-of-range-part>"<code>https://255.255.4000.1</code>"
   <td class=yes>Yes<br>(only if applicable to the last part)
  <!-- host parser, but grouped with IPv6- -->
  <tr>
   <td><dfn>IPv6-unclosed</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> is missing the closing U+005D (]).
    <p class=example id=example-ipv6-unclosed>"<code>https://\[::1</code>"
   <td class=yes>Yes
  <!-- IPv6 parser -->
  <tr>
   <td><dfn>IPv6-invalid-compression</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> begins with improper compression.
    <p class=example id=example-ipv6-invalid-compression>"<code>https://\[:1]</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv6-too-many-pieces</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> contains more than 8 pieces.
    <p class=example id=example-ipv6-too-many-pieces>"<code>https://[1:2:3:4:5:6:7:8:9]</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv6-multiple-compression</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> is compressed in more than one spot.
    <p class=example id=example-ipv6-multiple-compression>"<code>https://[1::1::1]</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv6-invalid-code-point</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> contains a code point that is neither an <a>ASCII hex digit</a>
    nor a U+003A (:). Or it unexpectedly ends.
    <div class=example id=example-ipv6-invalid-code-point>
     <p>"<code>https://[1:2:3!:4]</code>"
     <p>"<code>https://[1:2:3:]</code>"
    </div>
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv6-too-few-pieces</dfn>
   <td>
    <p>An uncompressed <a for=/>IPv6 address</a> contains fewer than 8 pieces.
    <p class=example id=example-ipv6-too-few-pieces>"<code>https://[1:2:3]</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv4-in-IPv6-too-many-pieces</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> with <a for=/>IPv4 address</a> syntax: the IPv6 address has more
    than 6 pieces.
    <p class=example id=example-ipv4-in-ipv6-too-many-pieces>"<code>https://[1:1:1:1:1:1:1:127.0.0.1]</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv4-in-IPv6-invalid-code-point</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> with <a for=/>IPv4 address</a> syntax:
    <ul>
     <li>An IPv4 part is empty or contains a non-<a>ASCII digit</a>.
     <li>An IPv4 part contains a leading 0.
     <li>There are too many IPv4 parts.
    </ul>
    <div class=example id=example-ipv4-in-ipv6-invalid-code-point>
     <p>"<code>https://[ffff::.0.0.1]</code>"
     <p>"<code>https://[ffff::127.0.xyz.1]</code>"
     <p>"<code>https://[ffff::127.0xyz]</code>"
     <p>"<code>https://[ffff::127.00.0.1]</code>"
     <p>"<code>https://[ffff::127.0.0.1.2]</code>"
    </div>
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv4-in-IPv6-out-of-range-part</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> with <a for=/>IPv4 address</a> syntax: an IPv4 part exceeds 255.
    <p class=example id=example-ipv4-in-ipv6-out-of-range-part>"<code>https://[ffff::127.0.0.4000]</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>IPv4-in-IPv6-too-few-parts</dfn>
   <td>
    <p>An <a for=/>IPv6 address</a> with <a for=/>IPv4 address</a> syntax: an IPv4 address contains
    too few parts.
    <p class=example id=example-ipv4-in-ipv6-too-few-parts>"<code>https://[ffff::127.0.0]</code>"
   <td class=yes>Yes
 <tbody>
  <tr>
   <th colspan=3 scope=rowgroup><a href=#url-parsing>URL parsing</a>
  <!-- invalid-URL-unit is also present in the opaque-host parser, but this is a more logical place.
       -->
  <tr>
   <td><dfn>invalid-URL-unit</dfn>
   <td>
    <p>A code point is found that is not a <a>URL unit</a>.
    <div class=example id=example-invalid-url-unit>
     <p>"<code>https://example.org/></code>"
     <p>"<code> https://example.org </code>"
     <p>"<code>ht<br>tps://example.org</code>"
     <p>"<code>https://example.org/%s</code>"
    </div>
   <td class=no>·
  <tr>
   <td><dfn>special-scheme-missing-following-solidus</dfn>
   <td>
    <p>The input's scheme is not followed by "<code>//</code>".
    <div class=example id=example-special-scheme-missing-following-solidus>
     <p>"<code>file:c:/my-secret-folder</code>"
     <p>"<code>https:example.org</code>"
     <pre><code class="lang-javascript">
const url = new URL("https:foo.html", "https://example.org/");</code></pre>
    </div>
   <td class=no>·
  <tr>
   <td><dfn>missing-scheme-non-relative-URL</dfn>
   <td>
    <p>The input is missing a <a for=url>scheme</a>, because it does not begin with an
    <a>ASCII alpha</a>, and either no <a>base URL</a> was provided or the <a>base URL</a> cannot be
    used as a <a>base URL</a> because it has an <a for=url>opaque path</a>.
    <div class=example id=example-missing-scheme-non-relative-url>
     <p>Input's <a for=url>scheme</a> is missing and no <a>base URL</a> is given:
     <pre><code class=lang-javascript>
const url = new URL("💩");</code></pre>
     <p>Input's <a for=url>scheme</a> is missing, but the <a>base URL</a> has an
     <a for=url>opaque path</a>.
     <pre><code class=lang-javascript>
const url = new URL("💩", "mailto:user@example.org");</code></pre>
    </div>
   <td class=yes>Yes
  <tr>
   <td><dfn>invalid-reverse-solidus</dfn>
   <td>
    <p>The URL has a <a>special scheme</a> and it uses U+005C (\) instead of U+002F (/).
    <p class=example id=example-invalid-reverse-solidus>"<code>https://example.org\path\to\file</code>"
   <td class=no>·
  <tr>
   <td><dfn>invalid-credentials</dfn>
   <td>
    <p>The input <a>includes credentials</a>.
    <div class=example id=example-invalid-credentials>
     <p>"<code>https://user@example.org</code>"
     <p>"<code>ssh://user@example.org</code>"
    </div>
   <td class=no>·
  <tr>
   <td><dfn>host-missing</dfn>
   <td>
    <p>The input has a <a>special scheme</a>, but does not contain a <a for=/>host</a>.
    <div class=example id=example-host-missing>
     <p>"<code>https://#fragment</code>"
     <p>"<code>https://:443</code>"
     <p>"<code>https://user:pass@</code>"
    </div>
   <td class=yes>Yes
  <tr>
   <td><dfn>port-out-of-range</dfn>
   <td>
    <p>The input's port is too big.
    <p class=example id=example-port-out-of-range>"<code>https://example.org:70000</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>port-invalid</dfn>
   <td>
    <p>The input's port is invalid.
    <p class=example id=example-port-invalid>"<code>https://example.org:7z</code>"
   <td class=yes>Yes
  <tr>
   <td><dfn>file-invalid-Windows-drive-letter</dfn>
   <td>
    <p>The input is a <a>relative-URL string</a> that <a>starts with a Windows drive letter</a> and
    the <a>base URL</a>'s <a for=url>scheme</a> is "<code>file</code>".
    <pre class=example id=example-file-invalid-windows-drive-letter><code class=lang-javascript>
const url = new URL("/c:/path/to/file", "file:///c:/");</code></pre>
   <td class=no>·
  <tr>
   <td><dfn>file-invalid-Windows-drive-letter-host</dfn>
   <td>
    <p>A <code>file:</code> URL's host is a Windows drive letter.
    <p class=example id=example-file-invalid-windows-drive-letter-host>"<code>file://c:</code>"
   <td class=no>·
</table>


<h3 id=parsers>Parsers</h3>

<p>The <dfn>EOF code point</dfn> is a conceptual code point that signifies the end of a string or
code point stream.

<p>A <dfn>pointer</dfn> for a <a for=/>string</a> <var>input</var> is an integer that points to a
<a for=/>code point</a> within <var>input</var>. Initially it points to the start of
<var>input</var>. If it is &minus;1 it points nowhere. If it is greater than or equal to
<var>input</var>'s <a for=string>code point length</a>, it points to the <a>EOF code point</a>.

<p>When a <a>pointer</a> is used, <dfn>c</dfn> references the <a for=/>code point</a> the
<a>pointer</a> points to as long as it does not point nowhere. When the <a>pointer</a> points to
nowhere <a>c</a> cannot be used.

<p>When a <a>pointer</a> is used, <dfn>remaining</dfn> references the
<a lt="code point substring to the end of the string">code point substring</a> from the
<a>pointer</a> + 1 to the end of the string, as long as <a>c</a> is not the <a>EOF code point</a>.
When <a>c</a> is the <a>EOF code point</a> <a>remaining</a> cannot be used.

<p class=example id=example-12672b6a>If "<code>mailto:username@example</code>" is a <a>string</a>
being processed and a <a>pointer</a> points to @, <a>c</a> is U+0040 (@) and <a>remaining</a> is
"<code>example</code>".

<p class=example id=example-empty-string>If the empty string is being processed and a <a>pointer</a>
points to the start and is then decreased by 1, using <a>c</a> or <a>remaining</a> would be an
error.


<h3 id=percent-encoded-bytes>Percent-encoded bytes</h3>

<p>A <dfn>percent-encoded byte</dfn> is U+0025 (%), followed by two <a>ASCII hex digits</a>.

<p class=note>It is generally a good idea for sequences of <a>percent-encoded bytes</a> to be such
that, when <a for=string>percent-decoded</a> and then passed to
<a>UTF-8 decode without BOM or fail</a>, they do not end up as failure. How important this is
depends on where the <a>percent-encoded bytes</a> are used. E.g., for the <a>host parser</a> not
following this advice is fatal, whereas for <a href="#url-rendering-i18n">URL rendering</a> the
<a>percent-encoded bytes</a> would not be rendered <a for=string>percent-decoded</a>.

<div algorithm>
<p>To <dfn for=byte id=percent-encode>percent-encode</dfn> a <a for=/>byte</a> <var>byte</var>,
return a <a for=/>string</a> consisting of U+0025 (%), followed by two <a>ASCII upper hex digits</a>
representing <var>byte</var>.
</div>

<div algorithm>
<p>To <dfn export for="byte sequence" id=percent-decode>percent-decode</dfn> a
<a for=/>byte sequence</a> <var>input</var>, run these steps:

<p class=warning>Using anything but <a>UTF-8 decode without BOM</a> when <var>input</var> contains
bytes that are not <a>ASCII bytes</a> might be insecure and is not recommended.

<ol>
 <li><p>Let <var>output</var> be an empty <a>byte sequence</a>.

 <li>
  <p>For each byte <var>byte</var> in <var>input</var>:

  <ol>
   <li><p>If <var>byte</var> is not 0x25 (%), then append <var>byte</var> to <var>output</var>.

   <li><p>Otherwise, if <var>byte</var> is 0x25 (%) and the next two bytes after
   <var>byte</var> in <var>input</var> are not in the ranges 0x30 (0) to 0x39 (9),
   0x41 (A) to 0x46 (F), and 0x61 (a) to 0x66 (f), all inclusive, append <var>byte</var> to
   <var>output</var>.

   <li>
    <p>Otherwise:

    <ol>
     <li><p>Let <var>bytePoint</var> be the two bytes after <var>byte</var> in <var>input</var>,
     <a lt="isomorphic decode">decoded</a>, and then interpreted as hexadecimal number.
     <!-- We should have a better definition for this. -->

     <li><p>Append a byte whose value is <var>bytePoint</var> to
     <var>output</var>.

     <li><p>Skip the next two bytes in <var>input</var>.
    </ol>
  </ol>

 <li><p>Return <var>output</var>.
</ol>
</div>

<div algorithm>
<p>To <dfn export for=string>percent-decode</dfn> a <a for=/>scalar value string</a>
<var>input</var>:

<ol>
 <li><p>Let <var>bytes</var> be the <a>UTF-8 encoding</a> of <var>input</var>.

 <li><p>Return the <a for="byte sequence">percent-decoding</a> of <var>bytes</var>.
</ol>

<p class=note>In general, percent-encoding results in a string with more U+0025 (%) code points than
the input, and percent-decoding results in a byte sequence with less 0x25 (%) bytes than the input.
</div>

<hr>

<p>A <dfn>percent-encode set</dfn> is a <a for=/>set</a> of <a for=/>code points</a>.

<p>The <dfn oldids=simple-encode-set>C0 control percent-encode set</dfn> is a
<a>percent-encode set</a> consisting of <a>C0 controls</a> and all <a>code points</a> greater than
U+007E (~).

<p>The <dfn>fragment percent-encode set</dfn> is a <a>percent-encode set</a> consisting of the
<a>C0 control percent-encode set</a> and U+0020 SPACE, U+0022 ("), U+003C (&lt;), U+003E (>), and
U+0060 (`).

<p>The <dfn>query percent-encode set</dfn> is a <a>percent-encode set</a> consisting of the
<a>C0 control percent-encode set</a> and U+0020 SPACE, U+0022 ("), U+0023 (#), U+003C (&lt;), and
U+003E (>).

<p class=note>The <a>query percent-encode set</a> cannot be defined in terms of the
<a>fragment percent-encode set</a> due to the omission of U+0060 (`).

<p>The <dfn>special-query percent-encode set</dfn> is a <a>percent-encode set</a> consisting of
the <a>query percent-encode set</a> and U+0027 (').

<p>The <dfn oldids=default-encode-set>path percent-encode set</dfn> is a <a>percent-encode set</a>
consisting of the <a>query percent-encode set</a> and U+003F (?), U+005E (^), U+0060 (`),
U+007B ({), and U+007D (}).

<p>The <dfn oldids=userinfo-encode-set>userinfo percent-encode set</dfn> is a
<a>percent-encode set</a> consisting of the <a>path percent-encode set</a> and U+002F (/),
U+003A (:), U+003B (;), U+003D (=), U+0040 (@), U+005B ([) to U+005D (]), inclusive, and U+007C (|).

<p>The <dfn export>component percent-encode set</dfn> is a <a>percent-encode set</a> consisting of
the <a>userinfo percent-encode set</a> and U+0024 ($) to U+0026 (&amp;), inclusive, U+002B (+), and
U+002C (,).

<p class=note>This is used by <cite>HTML</cite> for
{{NavigatorContentUtils/registerProtocolHandler()}}, and could also be used by other standards to
percent-encode data that can then be embedded in a <a for=/>URL</a>'s <a for=url>path</a>,
<a for=url>query</a>, or <a for=url>fragment</a>; or in an <a for=/>opaque host</a>. Using it with
<a for=string>UTF-8 percent-encode</a> gives identical results to JavaScript's
<a method><code>encodeURIComponent()</code> [sic]</a>. [[HTML]] [[ECMA-262]]

<p>The <dfn><code>application/x-www-form-urlencoded</code> percent-encode set</dfn> is a
<a>percent-encode set</a> consisting of the <a>component percent-encode set</a> and U+0021 (!),
U+0027 (') to U+0029 RIGHT PARENTHESIS, inclusive, and U+007E (~).

<p class=note>The <a><code>application/x-www-form-urlencoded</code> percent-encode set</a> contains
all code points, except the <a>ASCII alphanumeric</a>, U+002A (*), U+002D (-), U+002E (.), and
U+005F (_).

<div algorithm>
<p>To <dfn for=string>percent-encode after encoding</dfn>, given an <a for=/>encoding</a>
<var>encoding</var>, <a for=/>scalar value string</a> <var>input</var>, and a
<a>percent-encode set</a> <var>percentEncodeSet</var>:

<ol>
 <li><p><a for=/>Assert</a>: <var>encoding</var> is <a for=/>UTF-8</a> or
 <var>percentEncodeSet</var> is <a>special-query percent-encode set</a> or
 <a><code>application/x-www-form-urlencoded</code> percent-encode set</a>.

 <li><p>Let <var>spaceAsPlus</var> be true if <var>percentEncodeSet</var> is
 <a><code>application/x-www-form-urlencoded</code> percent-encode set</a>; otherwise false.

 <li><p>Let <var>encoder</var> be the result of <a>getting an encoder</a> from <var>encoding</var>.

 <li><p>Let <var>inputQueue</var> be <var>input</var> converted to an <a for=/>I/O queue</a>.

 <li><p>Let <var>output</var> be the empty string.

 <li>
  <p>Let <var>potentialError</var> be 0.

  <p class=note>This needs to be a non-null value to initiate the subsequent while loop.

 <li>
  <p>While <var>potentialError</var> is non-null:

  <ol>
   <li><p>Let <var>encodeOutput</var> be an empty <a for=/>I/O queue</a>.

   <li><p>Set <var>potentialError</var> to the result of running <a>encode or fail</a> with
   <var>inputQueue</var>, <var>encoder</var>, and <var>encodeOutput</var>.

   <li>
    <p>For each <var>byte</var> of <var>encodeOutput</var> converted to a byte sequence:

    <ol>
     <li><p>If <var>spaceAsPlus</var> is true and <var>byte</var> is 0x20 (SP), then append
     U+002B (+) to <var>output</var> and <a for=iteration>continue</a>.

     <li><p>Let <var>isomorph</var> be a <a for=/>code point</a> whose <a for="code point">value</a>
     is <var>byte</var>'s <a for=byte>value</a>.

     <li><p>Assert: <var>percentEncodeSet</var> includes all non-<a>ASCII code points</a>.

     <li><p>If <var>isomorph</var> is not in <var>percentEncodeSet</var>, then append
     <var>isomorph</var> to <var>output</var>.

     <li><p>Otherwise, <a for=byte>percent-encode</a> <var>byte</var> and append the result to
     <var>output</var>.
    </ol>

   <li>
    <p>If <var>potentialError</var> is non-null, then append "<code>%26%23</code>", followed by the
    shortest sequence of <a for=/>ASCII digits</a> representing <var>potentialError</var> in base
    ten, followed by "<code>%3B</code>", to <var>output</var>.

    <p class=note>This can happen when <var>encoding</var> is not <a>UTF-8</a>.
  </ol>

 <li><p>Return <var>output</var>.
</ol>

<p class=note>Of the possible values for the <var>percentEncodeSet</var> argument only two end up
encoding U+0025 (%) and thus give “roundtripable data”: <a>component percent-encode set</a> and
<a><code>application/x-www-form-urlencoded</code> percent-encode set</a>. The other values for the
<var>percentEncodeSet</var> argument — which happen to be used by the <a>URL parser</a> — leave
U+0025 (%) untouched and as such it needs to be
<a for="code point" lt="UTF-8 percent-encode">percent-encoded</a> first in order to be properly
represented.
</div>

<div algorithm>
<p>To <dfn for="code point" id=utf-8-percent-encode>UTF-8 percent-encode</dfn> a
<a for=/>scalar value</a> <var>scalarValue</var> using a <var>percentEncodeSet</var>, return the
result of running <a for=string>percent-encode after encoding</a> with <a for=/>UTF-8</a>,
<var>scalarValue</var> as a <a for=/>string</a>, and <var>percentEncodeSet</var>.
</div>

<div algorithm>
<p>To <dfn export for=string>UTF-8 percent-encode</dfn> a <a for=/>scalar value string</a>
<var>input</var> using a <var>percentEncodeSet</var>, return the result of running
<a for=string>percent-encode after encoding</a> with <a for=/>UTF-8</a>, <var>input</var>, and
<var>percentEncodeSet</var>.
</div>

<hr>

<div class=example id=example-percent-encode-operations>
 <p>Here is a summary, by way of example, of the operations defined above:

 <table>
  <tr>
   <th>Operation
   <th>Input
   <th>Output
  <tr>
   <td rowspan=2><a for=byte>Percent-encode</a> <var>input</var>
   <td>0x23
   <td>"<code>%23</code>"
  <tr>
   <td>0x7F
   <td>"<code>%7F</code>"
  <tr>
   <td><a for="byte sequence">Percent-decode</a> <var>input</var>
   <td>`<code>%25%s%1G</code>`
   <td>`<code>%%s%1G</code>`
  <tr>
   <td><a for=string>Percent-decode</a> <var>input</var>
   <td>"<code>‽%25%2E</code>"
   <td>0xE2 0x80 0xBD 0x25 0x2E
  <tr>
   <td rowspan=3><a for=string>Percent-encode after encoding</a> with <a>Shift_JIS</a>,
   <var>input</var>, and the <a>special-query percent-encode set</a>
   <td>"<code> </code>"
   <td>"<code>%20</code>"
  <tr>
   <td>"<code>≡</code>"
   <td>"<code>%81%DF</code>"
  <tr>
   <td>"<code>‽</code>"
   <td>"<code>%26%238253%3B</code>"
  <tr>
   <td><a for=string>Percent-encode after encoding</a> with <a>ISO-2022-JP</a>, <var>input</var>,
   and the <a>special-query percent-encode set</a>
   <td>"<code>¥</code>"
   <td>"<code>%1B(J\%1B(B</code>"
  <tr>
   <td><a for=string>Percent-encode after encoding</a> with <a>Shift_JIS</a>, <var>input</var>, and
   the <a><code>application/x-www-form-urlencoded</code> percent-encode set</a>
   <td>"<code>1+1 ≡ 2%20‽</code>"
   <td>"<code>1%2B1+%81%DF+2%2520%26%238253%3B</code>"
  <tr>
   <td rowspan=2><a for="code point">UTF-8 percent-encode</a> <var>input</var> using the
   <a>userinfo percent-encode set</a>
   <td>U+2261 (≡)
   <td>"<code>%E2%89%A1</code>"
  <tr>
   <td>U+203D (‽)
   <td>"<code>%E2%80%BD</code>"
  <tr>
   <td><a for=string>UTF-8 percent-encode</a> <var>input</var> using the
   <a>userinfo percent-encode set</a>
   <td>"<code>Say what‽</code>"
   <td>"<code>Say%20what%E2%80%BD</code>"
 </table>
</div>



<h2 id=security-considerations>Security considerations</h2>

<p>The security of a <a for=/>URL</a> is a function of its environment. Care is to be
taken when rendering, interpreting, and passing <a for=/>URLs</a> around.

<p>When rendering and allocating new <a for=/>URLs</a> "spoofing" needs to be considered. An attack
whereby one <a for=/>host</a> or <a for=/>URL</a> can be confused for another. For instance,
consider how 1/l/I, m/rn/rri, 0/O, and а/a can all appear eerily similar. Or worse, consider how
U+202A LEFT-TO-RIGHT EMBEDDING and similar <a>code points</a> are invisible. [[UTR36]]

<p>When passing a <a for=/>URL</a> from party <var>A</var> to <var>B</var>, both need to
carefully consider what is happening. <var>A</var> might end up leaking data it does not
want to leak. <var>B</var> might receive input it did not expect and take an action that
harms the user. In particular, <var>B</var> should never trust <var>A</var>, as at some
point <a for=/>URLs</a> from <var>A</var> can come from untrusted sources.



<h2 id="hosts-(domains-and-ip-addresses)">Hosts (domains and IP addresses)</h2>

<p>At a high level, a <a for=/>host</a>, <a>valid host string</a>, <a>host parser</a>, and
<a>host serializer</a> relate as follows:

<ul>
 <li><p>The <a>host parser</a> takes an arbitrary <a>scalar value string</a> and returns either
 failure or a <a for=/>host</a>.

 <li><p>A <a for=/>host</a> can be seen as the in-memory representation.

 <li><p>A <a>valid host string</a> defines what input would not trigger a <a>validation error</a>
 or failure when given to the <a>host parser</a>. I.e., input that would be considered conforming or
 valid.

 <li><p>The <a>host serializer</a> takes a <a for=/>host</a> and returns an <a>ASCII string</a>. (If
 that string is then <a lt="host parser">parsed</a>, the result will <a for=host>equal</a> the
 <a for=/>host</a> that was <a lt="host serializer">serialized</a>.)
</ul>

<div class=example id=example-host-parsing>
 <p>A <a lt="host parser">parse</a>-<a lt="host serializer">serialize</a> roundtrip gives the
 following results, depending on the <var ignore>isOpaque</var> argument to the <a>host parser</a>:

 <table>
  <tr>
   <th>Input
   <th>Output (<var ignore>isOpaque</var> = false)
   <th>Output (<var ignore>isOpaque</var> = true)
  <tr>
   <td><code>EXAMPLE.COM</code>
   <td rowspan=2><code>example.com</code> (<a for=/>domain</a>)
   <td><code>EXAMPLE.COM</code> (<a>opaque host</a>)
  <tr>
   <td><code>example%2Ecom</code>
   <td><code>example%2Ecom</code> (<a>opaque host</a>)
  <tr>
   <td><code>faß.example</code>
   <td><code>xn--fa-hia.example</code> (<a for=/>domain</a>)
   <td><code>fa%C3%9F.example</code> (<a>opaque host</a>)
  <tr>
   <td><code>0</code>
   <td rowspan=3><code>0.0.0.0</code> (<a for=/ lt="IPv4 address">IPv4</a>)
   <td><code>0</code> (<a>opaque host</a>)
  <tr>
   <td><code>%30</code>
   <td><code>%30</code> (<a>opaque host</a>)
  <tr>
   <td><code>0x</code>
   <td><code>0x</code> (<a>opaque host</a>)
  <tr>
   <td><code>0xffffffff</code>
   <td><code>255.255.255.255</code> (<a for=/ lt="IPv4 address">IPv4</a>)
   <td><code>0xffffffff</code> (<a>opaque host</a>)
  <tr>
   <td><code>[0:0::1]</code>
   <td colspan=2><code>\[::1]</code> (<a for=/ lt="IPv6 address">IPv6</a>)
  <tr>
   <td><code>[0:0::1%5D</code>
   <td colspan=2 rowspan=2>Failure
  <tr>
   <td><code>[0:0::%31]</code>
  <tr>
   <td><code>09</code>
   <td rowspan=3>Failure
   <td><code>09</code> (<a>opaque host</a>)
  <tr>
   <td><code>example.255</code>
   <td><code>example.255</code> (<a>opaque host</a>)
  <tr>
   <td><code>example^example</code>
   <td>Failure
 </table>
</div>


<h3 id=host-representation>Host representation</h3>

<p>A <dfn export id=concept-host>host</dfn> is a <a>domain</a>, an <a>IP address</a>, an
<a>opaque host</a>, or an <a>empty host</a>. Typically a <a for=/>host</a> serves as a network
address, but it is sometimes used as opaque identifier in <a for=/>URLs</a> where a network address
is not necessary.

<p class=example id=example-opaque-host-url>A typical <a for=/>URL</a> whose <a for=url>host</a> is
an <a>opaque host</a> is <code>git://github.com/whatwg/url.git</code>.

<p class=note>The RFCs referenced in the paragraphs below are for informative purposes only. They
have no influence on <a for=/>host</a> writing, parsing, and serialization. Unless stated otherwise
in the sections that follow.

<p>A <dfn export id=concept-domain>domain</dfn> is a non-empty <a>ASCII string</a> that identifies a
realm within a network.
[[RFC1034]]

<p>The <dfn export lt="domain label">domain labels</dfn> of a <a>domain</a> <var>domain</var> are
the result of <a>strictly splitting</a> <var>domain</var> on U+002E (.).

<p class=note>The <code>example.com</code> and <code>example.com.</code> <a for=/>domains</a> are
not equivalent and typically treated as distinct.

<p>An <dfn export>IP address</dfn> is an <a>IPv4 address</a> or an <a>IPv6 address</a>.

<p>An <dfn export id=concept-ipv4>IPv4 address</dfn> is a <a for=/>32-bit unsigned integer</a> that
identifies a network address.
[[RFC791]]

<p>An <dfn export id=concept-ipv6>IPv6 address</dfn> is a <a for=/>128-bit unsigned integer</a> that
identifies a network address. This integer is composed of a <a for=/>list</a> of 8
<a for=/>16-bit unsigned integers</a>, also known as an <a for=/>IPv6 address</a>'s
<dfn export for="IPv6 address" id=concept-ipv6-piece>pieces</dfn>.
[[RFC4291]]

<p class="note">Support for <code>&lt;zone_id></code> is
<a href="https://www.w3.org/Bugs/Public/show_bug.cgi?id=27234#c2">intentionally omitted</a>.

<p>An <dfn export>opaque host</dfn> is a non-empty <a>ASCII string</a> that can be used for further
processing.

<p>An <dfn export>empty host</dfn> is the empty string.


<h3 id=host-miscellaneous>Host miscellaneous</h3>

<p>A <dfn export>forbidden host code point</dfn> is U+0000 NULL, U+0009 TAB, U+000A LF, U+000D CR,
U+0020 SPACE, U+0023 (#), U+002F (/), U+003A (:), U+003C (&lt;), U+003E (>), U+003F (?), U+0040 (@),
U+005B ([), U+005C (\), U+005D (]), U+005E (^), or U+007C (|).

<p>A <dfn export>forbidden domain code point</dfn> is a <a>forbidden host code point</a>,
a <a>C0 control</a>, U+0025 (%), or U+007F DELETE.

<div algorithm>
<p>To obtain the <dfn export for=host>public suffix</dfn> of a <a for=/>host</a> <var>host</var>,
run these steps. They return null or a <a for=/>domain</a> representing a portion of <var>host</var>
that is included on the <cite>Public Suffix List</cite>. [[!PSL]]

<ol>
 <li><p>If <var>host</var> is not a <a>domain</a>, then return null.

 <li><p>Let <var>trailingDot</var> be "<code>.</code>" if <var>host</var>
 <a for=string>ends with</a> "<code>.</code>"; otherwise the empty string.

 <li><p>Let <var>publicSuffix</var> be the public suffix determined by running the
 <a href="https://github.com/publicsuffix/list/wiki/Format#formal-algorithm">Public Suffix List algorithm</a>
 with <var>host</var> as domain. [[!PSL]]

 <li><p><a for=/>Assert</a>: <var>publicSuffix</var> is an <a>ASCII string</a> that
 <a for=string>ends with</a> <var>trailingDot</var>.

 <li><p>Return <var>publicSuffix</var>.
</ol>
</div>

<div algorithm>
<p>To obtain the <dfn export for=host>registrable domain</dfn> of a <a for=/>host</a>
<var>host</var>, run these steps. They return null or a <a for=/>domain</a> formed by
<var>host</var>'s <a for=host>public suffix</a> and the <a for=/>domain label</a> preceding it, if
any.

<ol>
 <li><p>If <var>host</var>'s <a for=host>public suffix</a> is null or <var>host</var>'s
 <a for=host>public suffix</a> <a for=host>equals</a> <var>host</var>, then return null.

 <li><p>Let <var>trailingDot</var> be "<code>.</code>" if <var>host</var>
 <a for=string>ends with</a> "<code>.</code>"; otherwise the empty string.

 <li><p>Let <var>registrableDomain</var> be the registrable domain determined by running the
 <a href="https://github.com/publicsuffix/list/wiki/Format#formal-algorithm">Public Suffix List algorithm</a>
 with <var>host</var> as domain. [[!PSL]]

 <li><p><a for=/>Assert</a>: <var>registrableDomain</var> is an <a>ASCII string</a> that
 <a for=string>ends with</a> <var>trailingDot</var>.

 <li><p>Return <var>registrableDomain</var>.
</ol>
</div>

<div class=example id=example-host-psl>
 <table>
  <tr>
   <th>Host input
   <th>Public suffix
   <th>Registrable domain
  <tr>
   <td><code>com</code>
   <td><code>com</code>
   <td>null
  <tr>
   <td><code>example.com</code>
   <td><code>com</code>
   <td><code>example.com</code>
  <tr>
   <td><code>www.example.com</code>
   <td><code>com</code>
   <td><code>example.com</code>
  <tr>
   <td><code>sub.www.example.com</code>
   <td><code>com</code>
   <td><code>example.com</code>
  <tr>
   <td><code>EXAMPLE.COM</code>
   <td><code>com</code>
   <td><code>example.com</code>
  <tr>
   <td><code>example.com.</code>
   <td><code>com.</code>
   <td><code>example.com.</code>
  <tr>
   <td><code>github.io</code>
   <td><code>github.io</code>
   <td>null
  <tr>
   <td><code>whatwg.github.io</code>
   <td><code>github.io</code>
   <td><code>whatwg.github.io</code>
  <tr>
   <td><code>إختبار</code>
   <td><code>xn--kgbechtv</code>
   <td>null
  <tr>
   <td><code>example.إختبار</code>
   <td><code>xn--kgbechtv</code>
   <td><code>example.xn--kgbechtv</code>
  <tr>
   <td><code>sub.example.إختبار</code>
   <td><code>xn--kgbechtv</code>
   <td><code>example.xn--kgbechtv</code>
  <tr>
   <td><code>[2001:0db8:85a3:0000:0000:8a2e:0370:7334]</code>
   <td>null
   <td>null
 </table>
</div>

<p class=warning id=warning-avoid-psl>Specifications should prefer the <a for=/>origin</a> concept
for security decisions. The notion of "<a for=host>public suffix</a>" and
"<a for=host>registrable domain</a>" cannot be relied-upon to provide a hard security boundary, as
the public suffix list will diverge from client to client. Specifications which ignore this advice
are encouraged to carefully consider whether URLs' schemes ought to be incorporated into any
decisions made, i.e. whether to use the <a for=/>same site</a> or <a>schemelessly same site</a>
concepts.


<h3 id=idna>IDNA</h3>

<div algorithm>
<p>The <dfn id=concept-domain-to-ascii>domain to ASCII</dfn> algorithm, given a <a for=/>string</a>
<var>domain</var> and a boolean <var>beStrict</var>, runs these steps:

<ol>
 <li>
  <p>Let <var>result</var> be the result of running <a abstract-op lt=ToASCII>Unicode ToASCII</a>
  with <i>domain_name</i> set to <var>domain</var>, <i>CheckHyphens</i> set to <var>beStrict</var>,
  <i>CheckBidi</i> set to true, <i>CheckJoiners</i> set to true, <i>UseSTD3ASCIIRules</i> set to
  <var>beStrict</var>, <i>Transitional_Processing</i> set to false, <i>VerifyDnsLength</i> set to
  <var>beStrict</var>, and <i>IgnoreInvalidPunycode</i> set to false. [[!UTS46]]

  <p class=note>If <var>beStrict</var> is false, <var>domain</var> is an <a>ASCII string</a>, and
  <a>strictly splitting</a> <var>domain</var> on U+002E (.) does not produce any
  <a for=list>item</a> that <a for=string>starts with</a> an <a>ASCII case-insensitive</a> match for
  "<code>xn--</code>", this step is equivalent to <a>ASCII lowercasing</a> <var>domain</var>.

 <li><p>If <var>result</var> is a failure value, <a>domain-to-ASCII</a> <a>validation error</a>,
 return failure.

 <li>
  <p>If <var>beStrict</var> is false:

  <ol>
   <li><p>If <var>result</var> is the empty string, <a>domain-to-ASCII</a> <a>validation error</a>,
   return failure.

   <li>
    <p>If <var>result</var> contains a <a>forbidden domain code point</a>,
    <a>domain-invalid-code-point</a> <a>validation error</a>, return failure.

    <p class=note>Due to web compatibility and compatibility with non-DNS-based systems the
    <a>forbidden domain code points</a> are a subset of those disallowed when
    <i>UseSTD3ASCIIRules</i> is true. See also
    <a href="https://github.com/whatwg/url/issues/397">issue #397</a>.
  </ol>

 <li>
  <p><a for=/>Assert</a>: <var>result</var> is not the empty string and does not contain a
  <a>forbidden domain code point</a>.

  <p class=note><cite>Unicode IDNA Compatibility Processing</cite> guarantees this holds when
  <var>beStrict</var> is true. [[UTS46]]

 <li><p>Return <var>result</var>.
</ol>

<p class=note>This document and the web platform at large use
<cite>Unicode IDNA Compatibility Processing</cite> and not IDNA2008. For instance,
<code>☕.example</code> becomes <code>xn--53h.example</code> and not failure. [[UTS46]] [[RFC5890]]
</div>

<div algorithm>
<p>The <dfn id=concept-domain-to-unicode>domain to Unicode</dfn> algorithm, given a <a>domain</a>
<var>domain</var> and a boolean <var>beStrict</var>, runs these steps:

<ol>
 <li><p>Let <var>result</var> be the result of running
 <a abstract-op lt=ToUnicode>Unicode ToUnicode</a> with <i>domain_name</i> set to <var>domain</var>,
 <i>CheckHyphens</i> set to <var>beStrict</var>, <i>CheckBidi</i> set to true, <i>CheckJoiners</i>
 set to true, <i>UseSTD3ASCIIRules</i> set to <var>beStrict</var>, <i>Transitional_Processing</i>
 set to false, and <i>IgnoreInvalidPunycode</i> set to false. [[!UTS46]]

 <li><p>Signify <a>domain-to-Unicode</a> <a>validation errors</a> for any returned errors, and then,
 return <var>result</var>.
</ol>
</div>


<h3 id=host-writing oldids=host-syntax>Host writing</h3>

<p>A <dfn export oldids=syntax-host>valid host string</dfn> must be a <a>valid domain string</a>, a
<a>valid IPv4-address string</a>, or: U+005B ([), followed by a
<a>valid IPv6-address string</a>, followed by U+005D (]).

<p>A <a for=/>string</a> <var>input</var> is a <dfn>valid domain</dfn> if these steps return true:

<ol>
 <li><p>Let <var>domain</var> be the result of running <a>domain to ASCII</a> with <var>input</var>
 and true.

 <li><p>Return false if <var>domain</var> is failure; otherwise true.
</ol>

<p class=XXX>Ideally we define this in terms of a sequence of code points that make up a
<a>valid domain</a> rather than through a whack-a-mole:
<a href=https://github.com/whatwg/url/issues/245>issue 245</a>.

<p>A <dfn export oldids=syntax-host-domain>valid domain string</dfn> must be a string that is a
<a>valid domain</a>.

<p>A <dfn export oldids=syntax-host-ipv4>valid IPv4-address string</dfn> must be four shortest
possible strings of <a>ASCII digits</a>, representing a decimal number in the range 0 to 255,
inclusive, separated from each other by U+002E (.).

<p>A <dfn export oldids=syntax-host-ipv6>valid IPv6-address string</dfn> is defined in the
<a href="https://tools.ietf.org/html/rfc4291#section-2.2">"Text Representation of Addresses" chapter of IP Version 6 Addressing Architecture</a>.
[[!RFC4291]]
<!-- https://tools.ietf.org/html/rfc5952 updates that RFC, but it seems as
     far as what developers can do we should be liberal

     XXX should we define the format inline instead just like STD 66? -->

<p>A <dfn export>valid opaque-host string</dfn> must be one of the following:

<ul class=brief>
 <li><p>one or more <a>URL units</a> excluding <a>forbidden host code points</a>
 <li><p>U+005B ([), followed by a <a>valid IPv6-address string</a>, followed by U+005D (]).
</ul>

<p class=note>This is not part of the definition of <a>valid host string</a> as it requires context
to be distinguished.


<h3 id=host-parsing>Host parsing</h3>

<div algorithm>
<p>The <dfn export id=concept-host-parser lt="host parser|host parsing">host parser</dfn> takes a
<a>scalar value string</a> <var>input</var> with an optional boolean <var>isOpaque</var> (default
false), and then runs these steps. They return failure or a <a for=/>host</a>.

<ol>
 <li>
  <p>If <var>input</var> starts with U+005B ([), then:

  <ol>
   <li><p>If <var>input</var> does not end with U+005D (]), <a>IPv6-unclosed</a>
   <a>validation error</a>, return failure.

   <li><p>Return the result of <a lt="IPv6 parser">IPv6 parsing</a> <var>input</var> with its
   leading U+005B ([) and trailing U+005D (]) removed.
  </ol>

 <li><p>If <var>isOpaque</var> is true, then return the result of
 <a lt="opaque-host parser">opaque-host parsing</a> <var>input</var>.

 <li><p>Assert: <var>input</var> is not the empty string.

 <li>
  <p>Let <var>domain</var> be the result of running <a>UTF-8 decode without BOM</a> on the
  <a for=string>percent-decoding</a> of <var>input</var>.

  <p class=note>Alternatively <a>UTF-8 decode without BOM or fail</a> can be used, coupled with an
  early return for failure, as <a>domain to ASCII</a> fails on U+FFFD (�).

 <li><p>Let <var>asciiDomain</var> be the result of running <a>domain to ASCII</a> with
 <var>domain</var> and false.

 <li><p>If <var>asciiDomain</var> is failure, then return failure.

 <li><p>If <var>asciiDomain</var> <a lt="ends in a number checker">ends in a number</a>, then return
 the result of <a lt="IPv4 parser">IPv4 parsing</a> <var>asciiDomain</var>.

 <li><p>Return <var>asciiDomain</var>.
</ol>
</div>

<hr>

<div algorithm>
<p>The <dfn>ends in a number checker</dfn> takes an <a>ASCII string</a> <var>input</var> and then
runs these steps. They return a boolean.

<ol>
 <li><p>Let <var>parts</var> be the result of <a>strictly splitting</a> <var>input</var> on
 U+002E (.).

 <li>
  <p>If the last <a for=list>item</a> in <var>parts</var> is the empty string, then:

  <ol>
   <li><p>If <var>parts</var>'s <a for=list>size</a> is 1, then return false.

   <li><p><a for=list>Remove</a> the last <a for=list>item</a> from <var>parts</var>.
  </ol>

 <li><p>Let <var>last</var> be the last <a for=list>item</a> in <var>parts</var>.

 <li>
  <p>If <var>last</var> is non-empty and contains only <a>ASCII digits</a>, then return true.

  <p class=note>The erroneous input "<code>09</code>" will be caught by the <a>IPv4 parser</a> at a
  later stage.

 <li>
  <p>If parsing <var>last</var> as an <a lt="IPv4 number parser">IPv4 number</a> does not return
  failure, then return true.

  <p class=note>This is equivalent to checking that <var>last</var> is "<code>0X</code>" or
  "<code>0x</code>", followed by zero or more <a>ASCII hex digits</a>.

 <li><p>Return false.
</ol>
</div>

<div algorithm>
<p>The <dfn id=concept-ipv4-parser>IPv4 parser</dfn> takes an <a>ASCII string</a> <var>input</var>
and then runs these steps. They return failure or an <a for=/>IPv4 address</a>.

<p class=note>The <a for=/>IPv4 parser</a> is not to be invoked directly. Instead check that the
return value of the <a for=/>host parser</a> is an <a for=/>IPv4 address</a>.

<ol>
 <li><p>Let <var>parts</var> be the result of <a>strictly splitting</a> <var>input</var> on
 U+002E (.).

 <li>
  <p>If the last <a for=list>item</a> in <var>parts</var> is the empty string, then:

  <ol>
   <li><p><a>IPv4-empty-part</a> <a>validation error</a>.

   <li><p>If <var>parts</var>'s <a for=list>size</a> is greater than 1, then <a for=list>remove</a>
   the last <a for=list>item</a> from <var>parts</var>.
   <!-- Since the IPv4 parser is not to be invoked directly the input cannot be the empty string,
        but if it somehow is this conditional makes sure we can keep going. -->
  </ol>

 <li><p>If <var>parts</var>'s <a for=list>size</a> is greater than 4, <a>IPv4-too-many-parts</a>
 <a>validation error</a>, return failure.

 <li><p>Let <var>numbers</var> be an empty <a for=/>list</a>.

 <li>
  <p><a for=list>For each</a> <var>part</var> of <var>parts</var>:

  <ol>
   <li><p>Let <var>result</var> be the result of <a lt="IPv4 number parser">parsing</a>
   <var>part</var>.

   <li><p>If <var>result</var> is failure, <a>IPv4-non-numeric-part</a> <a>validation error</a>,
   return failure.

   <li><p>If <var>result</var>[1] is true, <a>IPv4-non-decimal-part</a> <a>validation error</a>.

   <li><p><a for=list>Append</a> <var>result</var>[0] to <var>numbers</var>.
  </ol>

 <li><p>If any item in <var>numbers</var> is greater than 255, <a>IPv4-out-of-range-part</a>
 <a>validation error</a>.

 <li><p>If any but the last <a for=list>item</a> in <var>numbers</var> is greater than 255, then
 return failure.

 <li><p>If the last <a for=list>item</a> in <var>numbers</var> is greater than or equal to
 256<sup>(5 &minus; <var>numbers</var>'s <a for=list>size</a>)</sup>, then return failure.

 <li><p>Let <var>ipv4</var> be the last <a for=list>item</a> in <var>numbers</var>.

 <li><p><a for=list>Remove</a> the last <a for=list>item</a> from <var>numbers</var>.

 <li><p>Let <var>counter</var> be 0.

 <li>
  <p><a for=list>For each</a> <var>n</var> of <var>numbers</var>:

  <ol>
   <li><p>Increment <var>ipv4</var> by <var>n</var> &times;
   256<sup>(3 &minus; <var>counter</var>)</sup>.

   <li><p>Increment <var>counter</var> by 1.
  </ol>

 <li><p>Return <var>ipv4</var>.
</ol>
</div>

<div algorithm>
<p>The <dfn>IPv4 number parser</dfn> takes an <a>ASCII string</a> <var>input</var> and then runs
these steps. They return failure or a <a for=/>tuple</a> of a number and a boolean.

<ol>
 <li><p>If <var>input</var> is the empty string, then return failure.

 <li><p>Let <var>validationError</var> be false.

 <li><p>Let <var>R</var> be 10.

 <li>
  <p>If <var>input</var> contains at least two code points and the first two code points are either
  "<code>0X</code>" or "<code>0x</code>", then:

  <ol>
   <li><p>Set <var>validationError</var> to true.

   <li><p>Remove the first two code points from <var>input</var>.

   <li><p>Set <var>R</var> to 16.
  </ol>

 <li>
  <p>Otherwise, if <var>input</var> contains at least two code points and the first code point is
  U+0030 (0), then:
  <!-- Needs to be at least two code points. Otherwise "0" as input fails to parse. -->

  <ol>
   <li><p>Set <var>validationError</var> to true.

   <li><p>Remove the first code point from <var>input</var>.

   <li><p>Set <var>R</var> to 8.
  </ol>

 <li><p>If <var>input</var> is the empty string, then return (0, true).
 <!-- 0x/0X is an IPv4 number apparently -->

 <li><p>If <var>input</var> contains a code point that is not a radix-<var>R</var> digit, then
 return failure.
 <!-- There is no need to set validationError here.
      XXX radix-R digit, hahaha, that's not a thing -->

 <li><p>Let <var>output</var> be the mathematical integer value that is represented by
 <var>input</var> in radix-<var>R</var> notation, using <a>ASCII hex digits</a> for digits with
 values 0 through 15.
 <!-- XXX well, you know, it works for ECMAScript, kinda -->

 <li><p>Return (<var>output</var>, <var>validationError</var>).
</ol>
</div>

<hr>

<div algorithm>
<p>The <dfn id=concept-ipv6-parser>IPv6 parser</dfn> takes a <a>scalar value string</a>
<var>input</var> and then runs these steps. They return failure or an <a for=/>IPv6 address</a>.

<p class=note>The <a for=/>IPv6 parser</a> could in theory be invoked directly, but please discuss
actually doing that with the editors of this document first.

<ol>
 <li><p>Let <var>address</var> be a new <a>IPv6 address</a> whose <a for="IPv6 address">pieces</a>
 are all 0.

 <li><p>Let <var>pieceIndex</var> be 0.

 <li><p>Let <var>compress</var> be null.

 <li><p>Let <var>pointer</var> be a <a>pointer</a> for <var>input</var>.

 <li>
  <p>If <a>c</a> is U+003A (:), then:

  <ol>
   <li><p>If <a>remaining</a> does not start with U+003A (:), <a>IPv6-invalid-compression</a>
   <a>validation error</a>, return failure.

   <li><p>Increase <var>pointer</var> by 2.

   <li><p>Increase <var>pieceIndex</var> by 1 and then set <var>compress</var> to
   <var>pieceIndex</var>.
  </ol>

 <li>
  <p>While <a>c</a> is not the <a>EOF code point</a>:

  <ol>
   <li><p>If <var>pieceIndex</var> is 8, <a>IPv6-too-many-pieces</a> <a>validation error</a>, return
   failure.

   <li>
    <p>If <a>c</a> is U+003A (:), then:

    <ol>
     <li><p>If <var>compress</var> is non-null, <a>IPv6-multiple-compression</a>
     <a>validation error</a>, return failure.

     <li>Increase <var>pointer</var> and <var>pieceIndex</var> by 1, set <var>compress</var> to
     <var>pieceIndex</var>, and then <a for=iteration>continue</a>.
    </ol>

   <li><p>Let <var>value</var> and <var>length</var> be 0.

   <li><p>While <var>length</var> is less than 4 and <a>c</a> is an <a>ASCII hex digit</a>, set
   <var>value</var> to <var>value</var> &times; 0x10 + <a>c</a> interpreted as hexadecimal number,
   and increase <var>pointer</var> and <var>length</var> by 1.

   <li>
    <p>If <a>c</a> is U+002E (.), then:

    <ol>
     <li><p>If <var>length</var> is 0, <a>IPv4-in-IPv6-invalid-code-point</a>
     <a>validation error</a>, return failure.

     <li><p>Decrease <var>pointer</var> by <var>length</var>.

     <li><p>If <var>pieceIndex</var> is greater than 6, <a>IPv4-in-IPv6-too-many-pieces</a>
     <a>validation error</a>, return failure.

     <li><p>Let <var>numbersSeen</var> be 0.

     <li>
      <p>While <a>c</a> is not the <a>EOF code point</a>:

      <ol>
       <li><p>Let <var>ipv4Piece</var> be null.

       <li>
        <p>If <var>numbersSeen</var> is greater than 0, then:

        <ol>
         <li><p>If <a>c</a> is a U+002E (.) and <var>numbersSeen</var> is less than 4, then increase
         <var>pointer</var> by 1.

         <li>Otherwise, <a>IPv4-in-IPv6-invalid-code-point</a> <a>validation error</a>, return
         failure.
        </ol>

       <li><p>If <a>c</a> is not an <a>ASCII digit</a>, <a>IPv4-in-IPv6-invalid-code-point</a>
       <a>validation error</a>, return failure.
       <!-- prevent the empty string -->

       <li>
        <p>While <a>c</a> is an <a>ASCII digit</a>:

        <ol>
         <li><p>Let <var>number</var> be <a>c</a> interpreted as decimal number.

         <li><p>If <var>ipv4Piece</var> is null, then set <var>ipv4Piece</var> to <var>number</var>.

         <li><p>Otherwise, if <var>ipv4Piece</var> is 0, <a>IPv4-in-IPv6-invalid-code-point</a>
         <a>validation error</a>, return failure.

         <li><p>Otherwise, set <var>ipv4Piece</var> to <var>ipv4Piece</var> &times; 10 +
         <var>number</var>.

         <li><p>If <var>ipv4Piece</var> is greater than 255, <a>IPv4-in-IPv6-out-of-range-part</a>
         <a>validation error</a>, return failure.

         <li><p>Increase <var>pointer</var> by 1.
        </ol>

       <li><p>Set <var>address</var>[<var>pieceIndex</var>] to
       <var>address</var>[<var>pieceIndex</var>] &times; 0x100 + <var>ipv4Piece</var>.

       <li><p>Increase <var>numbersSeen</var> by 1.

       <li><p>If <var>numbersSeen</var> is 2 or 4, then increase <var>pieceIndex</var> by 1.
      </ol>

     <li><p>If <var>numbersSeen</var> is not 4, <a>IPv4-in-IPv6-too-few-parts</a>
     <a>validation error</a>, return failure.

     <li><p><a for=iteration>Break</a>.
    </ol>

   <li>
    <p>Otherwise, if <a>c</a> is U+003A (:):

    <ol>
     <li><p>Increase <var>pointer</var> by 1.

     <li><p>If <a>c</a> is the <a>EOF code point</a>, <a>IPv6-invalid-code-point</a>
     <a>validation error</a>, return failure.
    </ol>

   <li><p>Otherwise, if <a>c</a> is not the <a>EOF code point</a>, <a>IPv6-invalid-code-point</a>
   <a>validation error</a>, return failure.

   <li><p>Set <var>address</var>[<var>pieceIndex</var>] to <var>value</var>.

   <li><p>Increase <var>pieceIndex</var> by 1.
  </ol>

 <li>
  <p>If <var>compress</var> is non-null, then:

  <ol>
   <li><p>Let <var>swaps</var> be <var>pieceIndex</var> &minus; <var>compress</var>.

   <li><p>Set <var>pieceIndex</var> to 7.

   <li><p>While <var>pieceIndex</var> is not 0 and <var>swaps</var> is greater than 0, swap
   <var>address</var>[<var>pieceIndex</var>] with
   <var>address</var>[<var>compress</var> + <var>swaps</var> &minus; 1], and then decrease both
   <var>pieceIndex</var> and <var>swaps</var> by 1.
  </ol>

 <li><p>Otherwise, if <var>compress</var> is null and <var>pieceIndex</var> is not 8,
 <a>IPv6-too-few-pieces</a> <a>validation error</a>, return failure.

 <li><p>Return <var>address</var>.
</ol>
</div>

<hr>

<div algorithm>
<p>The <dfn export id=concept-opaque-host-parser>opaque-host parser</dfn> takes a
<a>scalar value string</a> <var>input</var>, and then runs these steps. They return failure or an
<a for=/>opaque host</a>.

<ol>
 <li><p>If <var>input</var> contains a <a>forbidden host code point</a>,
 <a>host-invalid-code-point</a> <a>validation error</a>, return failure.

 <li><p>If <var>input</var> contains a <a>code point</a> that is not a <a>URL code point</a> and not
 U+0025 (%), <a>invalid-URL-unit</a> <a>validation error</a>.

 <li><p>If <var>input</var> contains a U+0025 (%) and the two <a>code points</a> following it are
 not <a>ASCII hex digits</a>, <a>invalid-URL-unit</a> <a>validation error</a>.

 <li><p>Return the result of running <a for=string>UTF-8 percent-encode</a> on <var>input</var>
 using the <a>C0 control percent-encode set</a>.
</ol>
</div>


<h3 id=host-serializing>Host serializing</h3>

<div algorithm>
<p>The <dfn id=concept-host-serializer lt="host serializer">host serializer</dfn> takes a
<a for=/>host</a> <var>host</var> and then runs these steps. They return an <a>ASCII string</a>.

<ol>
 <li><p>If <var>host</var> is an <a>IPv4 address</a>, return the result of
 running the <a>IPv4 serializer</a> on <var>host</var>.

 <li><p>Otherwise, if <var>host</var> is an <a>IPv6 address</a>, return U+005B ([), followed by the
 result of running the <a>IPv6 serializer</a> on <var>host</var>, followed by U+005D (]).

 <li><p>Otherwise, <var>host</var> is a <a>domain</a>, <a>opaque host</a>, or <a>empty host</a>,
 return <var>host</var>.
</ol>
</div>

<div algorithm>
<p>The <dfn id=concept-ipv4-serializer>IPv4 serializer</dfn> takes an <a>IPv4 address</a>
<var>address</var> and then runs these steps. They return an <a>ASCII string</a>.

<ol>
 <li><p>Let <var>output</var> be the empty string.

 <li><p>Let <var>n</var> be the value of <var>address</var>.

 <li>
  <p><a for=set>For each</a> <var>i</var> in the range 1 to 4, inclusive:

  <ol>
   <li><p>Prepend <var>n</var> % 256, <a lt="serialize an integer">serialized</a>, to
   <var>output</var>.

   <li><p>If <var>i</var> is not 4, then prepend U+002E (.) to <var>output</var>.

   <li><p>Set <var>n</var> to floor(<var>n</var> / 256).
  </ol>

 <li><p>Return <var>output</var>.
</ol>
</div>

<div algorithm>
<p>The <dfn id=concept-ipv6-serializer>IPv6 serializer</dfn> takes an <a>IPv6 address</a>
<var>address</var> and then runs these steps. They return an <a>ASCII string</a>.

<ol>
 <li><p>Let <var>output</var> be the empty string.

 <li><p>Let <var>compress</var> be the result of
 <a for=/>finding the IPv6 address compressed piece index</a> given <var>address</var>.

 <li><p>Let <var>ignore0</var> be false.

 <li>
  <p><a for=set>For each</a> <var>pieceIndex</var> of <var>address</var>'s
  <a for="IPv6 address">pieces</a>'s <a for=list>indices</a>:

  <ol>
   <li><p>If <var>ignore0</var> is true and <var>address</var>[<var>pieceIndex</var>] is 0, then
   <a for=iteration>continue</a>.

   <li><p>Otherwise, if <var>ignore0</var> is true, set <var>ignore0</var> to false.

   <li>
    <p>If <var>compress</var> is <var>pieceIndex</var>, then:

    <ol>
     <li><p>Let <var>separator</var> be "<code>::</code>" if <var>pieceIndex</var> is 0; otherwise
     U+003A (:).

     <li><p>Append <var>separator</var> to <var>output</var>.

     <li><p>Set <var>ignore0</var> to true and <a for=iteration>continue</a>.
    </ol>

   <li><p>Append <var>address</var>[<var>pieceIndex</var>], represented as the shortest possible
   lowercase hexadecimal number, to <var>output</var>.

   <li><p>If <var>pieceIndex</var> is not 7, then append U+003A (:) to <var>output</var>.
  </ol>

 <li><p>Return <var>output</var>.
</ol>

<p class=note>This algorithm requires the recommendation from
A Recommendation for IPv6 Address Text Representation.
[[RFC5952]]
</div>

<div algorithm>
<p>To <dfn>find the IPv6 address compressed piece index</dfn> given an <a for=/>IPv6 address</a>
<var>address</var>:

<ol>
 <li><p>Let <var>longestIndex</var> be null.

 <li><p>Let <var>longestSize</var> be 1.

 <li><p>Let <var>foundIndex</var> be null.

 <li><p>Let <var>foundSize</var> be 0.

 <li>
  <p><a for=list>For each</a> <var>pieceIndex</var> of <var>address</var>'s
  <a for="IPv6 address">pieces</a>'s <a for=list>indices</a>:

  <ol>
   <li>
    <p>If <var>address</var>'s <a for="IPv6 address">pieces</a>[<var>pieceIndex</var>] is not 0:

    <ol>
     <li><p>If <var>foundSize</var> is greater than <var>longestSize</var>, then set
     <var>longestIndex</var> to <var>foundIndex</var> and <var>longestSize</var> to
     <var>foundSize</var>.

     <li>Set <var>foundIndex</var> to null.

     <li>Set <var>foundSize</var> to 0.
    </ol>

   <li>
    <p>Otherwise:

    <ol>
     <li><p>If <var>foundIndex</var> is null, then set <var>foundIndex</var> to
     <var>pieceIndex</var>.

     <li><p>Increment <var>foundSize</var> by 1.
    </ol>
  </ol>

 <li><p>If <var>foundSize</var> is greater than <var>longestSize</var>, then return
 <var>foundIndex</var>.

 <li><p>Return <var>longestIndex</var>.
</ol>

<p class=example id=example-e2b3492e>In <code>0:f:0:0:f:f:0:0</code> it would point to the second 0.
</div>


<h3 id=host-equivalence>Host equivalence</h3>

<div algorithm>
<p>To determine whether a <a for=/>host</a> <var>A</var>
<dfn export for=host id=concept-host-equals lt=equal>equals</dfn> <a for=/>host</a> <var>B</var>,
return true if <var>A</var> is <var>B</var>, and false otherwise.
</div>

<p class=XXX>Certificate comparison requires a host equivalence check that ignores the
trailing dot of a domain (if any). However, those hosts have also various other facets
enforced, such as DNS length, that are not enforced here, as URLs do not enforce them. If
anyone has a good suggestion for how to bring these two closer together, or what a good
unified model would be, please file an issue.



<h2 id=urls>URLs</h2>

<!-- History behind URL as term:
     https://lists.w3.org/Archives/Public/uri/2012Oct/0080.html -->

<p>At a high level, a <a for=/>URL</a>, <a>valid URL string</a>, <a>URL parser</a>, and
<a>URL serializer</a> relate as follows:

<ul>
 <li><p>The <a>URL parser</a> takes an arbitrary <a>scalar value string</a> and returns either
 failure or a <a for=/>URL</a>. It might also record zero or more <a>validation errors</a>.

 <li><p>A <a for=/>URL</a> can be seen as the in-memory representation.

 <li><p>A <a>valid URL string</a> defines what input would not trigger a <a>validation error</a> or
 failure when given to the <a>URL parser</a>. I.e., input that would be considered conforming or
 valid.

 <li><p>The <a>URL serializer</a> takes a <a for=/>URL</a> and returns an <a>ASCII string</a>. (If
 that string is then <a lt="URL parser">parsed</a>, the result will <a for=url>equal</a> the <a
 for=/>URL</a> that was <a lt="URL serializer">serialized</a>.) The output of the
 <a>URL serializer</a> is not always a <a>valid URL string</a>.
</ul>

<div class=example id=example-url-parsing>
 <table>
  <tr>
   <th>Input
   <th>Base
   <th>Valid
   <th>Output
  <tr>
   <td><code>https:example.org</code>
   <td>
   <td>❌
   <td><code>https://example.org/</code>
  <tr>
   <td><code>https://////example.com///</code>
   <td>
   <td>❌
   <td><code>https://example.com///</code>
  <tr>
   <td><code>https://example.com/././foo</code>
   <td>
   <td>✅
   <td><code>https://example.com/foo</code>
  <tr>
   <td><code>hello:world</code>
   <td><code>https://example.com/</code>
   <td>✅
   <td><code>hello:world</code>
  <tr>
   <td><code>https:example.org</code>
   <td><code>https://example.com/</code>
   <td>❌
   <td><code>https://example.com/example.org</code>
  <tr>
   <td><code>\example\..\demo/.\</code>
   <td><code>https://example.com/</code>
   <td>❌
   <td><code>https://example.com/demo/</code>
  <tr>
   <td><code>example</code>
   <td><code>https://example.com/demo</code>
   <td>✅
   <td><code>https://example.com/example</code>
  <tr>
   <td><code>file:///C|/demo</code>
   <td>
   <td>❌
   <td><code>file:///C:/demo</code>
  <tr>
   <td><code>..</code>
   <td><code>file:///C:/demo</code>
   <td>✅
   <td><code>file:///C:/</code>
  <tr>
   <td><code>file://loc%61lhost/</code>
   <td>
   <td>✅
   <td><code>file:///</code>
  <tr>
   <td><code>https://user:password@example.org/</code>
   <td>
   <td>❌
   <td><code>https://user:password@example.org/</code>
  <tr>
   <td><code>https://example.org/foo bar</code>
   <td>
   <td>❌
   <td><code>https://example.org/foo%20bar</code>
  <tr>
   <td><code>https://EXAMPLE.com/../x</code>
   <td>
   <td>✅
   <td><code>https://example.com/x</code>
  <tr>
   <td><code>https://ex ample.org/</code>
   <td>
   <td>❌
   <td>Failure
  <tr>
   <td><code>example</code>
   <td>
   <td>❌, due to lack of base
   <td>Failure
  <tr>
   <td><code>https://example.com:demo</code>
   <td>
   <td>❌
   <td>Failure
  <tr>
   <td><code>http://[www.example.com]/</code>
   <td>
   <td>❌
   <td>Failure
  <tr>
   <td><code>https://example.org//</code>
   <td>
   <td>✅
   <td><code>https://example.org//</code>
  <tr>
   <td><code>https://example.com/[]?[]#[]</code>
   <td>
   <td>❌
   <td><code>https://example.com/[]?[]#[]</code>
  <tr>
   <td><code>https://example/%?%#%</code>
   <td>
   <td>❌
   <td><code>https://example/%?%#%</code>
  <tr>
   <td><code>https://example/%25?%25#%25</code>
   <td>
   <td>✅
   <td><code>https://example/%25?%25#%25</code>
 </table>

 <p>The base and output <a lt="URL record">URL</a> are represented in
 <a lt="URL serializer">serialized</a> form for brevity.
</div>


<h3 id=url-representation>URL representation</h3>

<p>A <dfn export id=concept-url lt="URL|URL record">URL</dfn> is a <a for=/>struct</a> that
represents a universal identifier. To disambiguate from a <a>valid URL string</a> it can also be
referred to as a <a for=/>URL record</a>.

<p>A <a for=/>URL</a>'s <dfn export for=url id=concept-url-scheme>scheme</dfn> is an
<a>ASCII string</a> that identifies the type of <a for=/>URL</a> and can be used to
dispatch a <a for=/>URL</a> for further processing after <a lt="URL parser">parsing</a>.
It is initially the empty string.

<p>A <a for=/>URL</a>'s <dfn export for=url id=concept-url-username>username</dfn> is an
<a>ASCII string</a> identifying a username. It is initially the empty string.

<p>A <a for=/>URL</a>'s <dfn export for=url id=concept-url-password>password</dfn> is an
<a>ASCII string</a> identifying a password. It is initially the empty string.

<p>A <a for=/>URL</a>'s <dfn export for=url id=concept-url-host>host</dfn> is null or a
<a for=/>host</a>. It is initially null.

<div class="note">
 <p>The following table lists allowed <a for=/>URL</a>'s <a for=url>scheme</a> /
 <a for=url>host</a> combinations.

 <table>
  <tr>
   <th rowspan=2><a for=url>scheme</a>
   <th colspan=6><a for=url>host</a>
  <tr>
   <th><a>domain</a>
   <th><a>IPv4 address</a>
   <th><a>IPv6 address</a>
   <th><a>opaque host</a>
   <th><a>empty host</a>
   <th>null
  <tr>
   <td><a>Special schemes</a> excluding "<code>file</code>"
   <td>✅
   <td>✅
   <td>✅
   <td>❌
   <td>❌
   <td>❌
  <tr>
   <td>"<code>file</code>"
   <td>✅
   <td>✅
   <td>✅
   <td>❌
   <td>✅
   <td>❌
  <tr>
   <td>Others
   <td>❌
   <td>❌
   <td>✅
   <td>✅
   <td>✅
   <td>✅
 </table>
</div>

<p>A  <a for=/>URL</a>'s <dfn export for=url id=concept-url-port>port</dfn> is either null or a
<a for=/>16-bit unsigned integer</a> that identifies a networking port. It is initially null.

<p>A <a for=/>URL</a>'s
<dfn export for=url id=concept-url-path oldids=non-relative-flag,url-cannot-be-a-base-url-flag>path</dfn>
is a <a for=/>URL path</a>, usually identifying a location. It is initially « ».

<p class=note>A <a lt="is special">special</a> <a for=/>URL</a>'s <a for=url>path</a> is always a
<a for=/>list</a>, i.e., it is never <a for=url lt="opaque path">opaque</a>.

<p>A  <a for=/>URL</a>'s <dfn export for=url id=concept-url-query>query</dfn> is either
null or an <a>ASCII string</a>. It is initially null.

<p>A <a for=/>URL</a>'s <dfn export for=url id=concept-url-fragment>fragment</dfn> is either null or
an <a>ASCII string</a> that can be used for further processing on the resource the
<a for=/>URL</a>'s other components identify. It is initially null.

<p>A <a for=/>URL</a> also has an associated
<dfn export for=url id=concept-url-blob-entry>blob URL entry</dfn> that is either null or a
<a for=/>blob URL entry</a>. It is initially null.

<p class=note>This is used to support caching the object a "<code>blob</code>" URL refers to as well
as its origin. It is important that these are cached as the <a for=/>URL</a> might be removed from
the <a>blob URL store</a> between parsing and fetching, while fetching will still need to succeed.

<div class=example id=example-url-components>
 <p>The following table lists how <a>valid URL strings</a>, when <a lt="URL parser">parsed</a>, map
 to a <a for=/>URL</a>'s components. <a for=url>Username</a>, <a for=url>password</a>, and
 <a for=url>blob URL entry</a> are omitted; in the examples below they are the empty string, the
 empty string, and null, respectively.

 <table>
  <tr>
   <th>Input
   <th><a for=url>Scheme</a>
   <th><a for=url>Host</a>
   <th><a for=url>Port</a>
   <th><a for=url>Path</a>
   <th><a for=url>Query</a>
   <th><a for=url>Fragment</a>
  <tr>
   <td><code>https://example.com/</code>
   <td>"<code>https</code>"
   <td>"<code>example.com</code>"
   <td>null
   <td>« the empty string »
   <td>null
   <td>null
  <tr>
   <td><code>https://localhost:8000/search?q=text#hello</code>
   <td>"<code>https</code>"
   <td>"<code>localhost</code>"
   <td>8000
   <td>« "<code>search</code>" »
   <td>"<code>q=text</code>"
   <td>"<code>hello</code>"
  <tr>
   <td><code>urn:isbn:9780307476463</code>
   <td>"<code>urn</code>"
   <td>null
   <td>null
   <td>"<code>isbn:9780307476463</code>"
   <td>null
   <td>null
  <tr>
   <td><code>file:///ada/Analytical%20Engine/README.md</code>
   <td>"<code>file</code>"
   <td>null
   <td>null
   <td>« "<code>ada</code>", "<code>Analytical%20Engine</code>", "<code>README.md</code>" »
   <td>null
   <td>null
 </table>
</div>

<hr>

<p>A <dfn export>URL path</dfn> is either a <a>URL path segment</a> or a <a for=/>list</a> of zero
or more <a>URL path segments</a>.

<p>A <dfn export>URL path segment</dfn> is an <a for=/>ASCII string</a>. It commonly refers to a
directory or a file, but has no predefined meaning.

<p>A
<dfn export id=single-dot-path-segment oldids=syntax-url-path-segment-dot>single-dot URL path segment</dfn>
is a <a for=/>URL path segment</a> that is "<code>.</code>" or an <a>ASCII case-insensitive</a>
match for "<code>%2e</code>".
<!-- "." is not a code point here -->

<p>A
<dfn export id=double-dot-path-segment oldids=syntax-url-path-segment-dotdot>double-dot URL path segment</dfn>
is a <a for=/>URL path segment</a> that is "<code>..</code>" or an <a>ASCII case-insensitive</a>
match for "<code>.%2e</code>", "<code>%2e.</code>", or "<code>%2e%2e</code>".
<!-- Referenced by EPUB. -->


<h3 id=url-miscellaneous>URL miscellaneous</h3>

<p>A <dfn export>special scheme</dfn> is an <a>ASCII string</a> that is listed in the first column
of the following table. The <dfn export>default port</dfn> for a <a>special scheme</a> is listed in
the second column on the same row. The <a>default port</a> for any other <a>ASCII string</a> is
null.

<table>
 <tr><th><a>Special scheme</a>
     <th><a>Default port</a>
 <tr><td>"<code>ftp</code>"<td>21
 <tr><td>"<code>file</code>"<td>null
 <tr><td>"<code>http</code>"<td>80
 <tr><td>"<code>https</code>"<td>443
 <tr><td>"<code>ws</code>"<td>80
 <tr><td>"<code>wss</code>"<td>443
</table>

<p>A <a for=/>URL</a> <dfn export>is special</dfn> if its <a for=url>scheme</a> is a
<a>special scheme</a>. A <a for=/>URL</a> <dfn>is not special</dfn> if its <a for=url>scheme</a> is
not a <a>special scheme</a>.

<p>A <a for=/>URL</a>
<dfn export lt="include credentials|includes credentials">includes credentials</dfn> if its
<a for=url>username</a> or <a for=url>password</a> is not the empty string.
<!-- also used by Fetch -->

<p>A <a for=/>URL</a> has an <dfn export for=url>opaque path</dfn> if its <a for=url>path</a> is a
<a for=/>URL path segment</a>.

<p>A <a for=/>URL</a> <dfn export>cannot have a username/password/port</dfn> if its
<a for=url>host</a> is null or the empty string, or its <a for=url>scheme</a> is
"<code>file</code>".

<p>A <a for=/>URL</a> can be designated as <dfn id=concept-base-url>base URL</dfn>.

<p class=note>A <a>base URL</a> is useful for the <a>URL parser</a> when the input might be a
<a>relative-URL string</a>.

<hr>

<p>A <dfn>Windows drive letter</dfn> is two code points, of which the first is an <a>ASCII alpha</a>
and the second is either U+003A (:) or U+007C (|).

<p>A <dfn>normalized Windows drive letter</dfn> is a <a>Windows drive letter</a> of which the second
code point is U+003A (:).

<p class="note">As per the <a href=#url-writing>URL writing</a> section, only a
<a>normalized Windows drive letter</a> is conforming.

<p>A string
<dfn lt="start with a Windows drive letter|starts with a Windows drive letter">starts with a Windows drive letter</dfn>
if all of the following are true:

<ul class=brief>
 <li>its <a for=string>length</a> is greater than or equal to 2
 <li>its first two code points are a <a>Windows drive letter</a>
 <li>its <a for=string>length</a> is 2 or its third code point is U+002F (/), U+005C (\),
 U+003F (?), or U+0023 (#).
</ul>

<div class=example id=example-start-with-a-widows-drive-letter>
 <table>
  <tr>
   <th>String
   <th>Starts with a Windows drive letter
  <tr>
   <td>"<code>c:</code>"
   <td>✅
  <tr>
   <td>"<code>c:/</code>"
   <td>✅
  <tr>
   <td>"<code>c:a</code>"
   <td>❌
 </table>
</div>

<div algorithm>
<p id=pop-a-urls-path>To <dfn local-lt=shorten>shorten a <var>url</var>'s path</dfn>:

<ol>
 <li><p><a for=/>Assert</a>: <var>url</var> does not have an <a for=url>opaque path</a>.

 <li><p>Let <var>path</var> be <var>url</var>'s <a for=url>path</a>.

 <li><p>If <var>url</var>'s <a for=url>scheme</a> is "<code>file</code>", <var>path</var>'s
 <a for=list>size</a> is 1, and <var>path</var>[0] is a <a>normalized Windows drive letter</a>, then
 return.

 <li><p><a for=list>Remove</a> <var>path</var>'s last item, if any.
</ol>
</div>


<h3 id=url-writing oldids=url-syntax>URL writing</h3>

<!-- http://tantek.com/2011/238/b1/many-ways-slice-url-name-pieces -->

<p>A <dfn export oldids=syntax-url>valid URL string</dfn> must be either a
<a>relative-URL-with-fragment string</a> or an <a>absolute-URL-with-fragment string</a>.

<p>An
<dfn export oldids=syntax-url-absolute-with-fragment>absolute-URL-with-fragment string</dfn> must be
an <a>absolute-URL string</a>, optionally followed by U+0023 (#) and a <a>URL-fragment string</a>.

<p>An <dfn export oldids=syntax-url-absolute>absolute-URL string</dfn> must be one of the following:

<ul class=brief>
 <li><p>a <a>URL-scheme string</a> that is an <a>ASCII case-insensitive</a> match for a
 <a>special scheme</a> and not an <a>ASCII case-insensitive</a> match for "<code>file</code>",
 followed by U+003A (:) and a <a>scheme-relative-special-URL string</a>
 <li><p>a <a>URL-scheme string</a> that is <em>not</em> an <a>ASCII case-insensitive</a> match for a
 <a>special scheme</a>, followed by U+003A (:) and a <a>relative-URL string</a>
 <li><p>a <a>URL-scheme string</a> that is an <a>ASCII case-insensitive</a> match for
 "<code>file</code>", followed by U+003A (:) and a <a>scheme-relative-file-URL string</a>
</ul>

<p>any optionally followed by U+003F (?) and a <a>URL-query string</a>.

<p>A <dfn export oldids=syntax-url-scheme>URL-scheme string</dfn> must be one <a>ASCII alpha</a>,
followed by zero or more of <a>ASCII alphanumeric</a>, U+002B (+), U+002D (-), and U+002E (.).
<a lt="URL-scheme string">Schemes</a> should be registered in the
<cite>IANA URI [sic] Schemes</cite> registry.
[[!IANA-URI-SCHEMES]]
[[RFC7595]]

<p>A <dfn export oldids=syntax-url-relative-with-fragment>relative-URL-with-fragment string</dfn>
must be a <a>relative-URL string</a>, optionally followed by U+0023 (#) and a
<a>URL-fragment string</a>.

<p>A <dfn export oldids=syntax-url-relative>relative-URL string</dfn> must be one of the following,
switching on <a>base URL</a>'s <a for=url>scheme</a>:

<dl class=switch>
 <dt>A <a>special scheme</a> that is not "<code>file</code>"
 <dd><p>a <a>scheme-relative-special-URL string</a>
 <dd><p>a <a>path-absolute-URL string</a>
 <dd><p>a <a>path-relative-scheme-less-URL string</a>
 <dt>"<code>file</code>"
 <dd><p>a <a>scheme-relative-file-URL string</a>
 <dd><p>a <a>path-absolute-URL string</a> if <a>base URL</a>'s <a for=url>host</a> is an
 <a>empty host</a>
 <dd><p>a <a>path-absolute-non-Windows-file-URL string</a> if <a>base URL</a>'s <a for=url>host</a>
 is not an <a>empty host</a>
 <dd><p>a <a>path-relative-scheme-less-URL string</a>
 <dt>Otherwise
 <dd><p>a <a>scheme-relative-URL string</a>
 <dd><p>a <a>path-absolute-URL string</a>
 <dd><p>a <a>path-relative-scheme-less-URL string</a>
</dl>

<p>any optionally followed by U+003F (?) and a <a>URL-query string</a>.

<p class=note>A non-null <a>base URL</a> is necessary when <a lt="URL parser">parsing</a> a
<a>relative-URL string</a>.

<p>A <dfn export>scheme-relative-special-URL string</dfn> must be "<code>//</code>", followed by a
<a>valid host string</a>, optionally followed by U+003A (:) and a <a>URL-port string</a>, optionally
followed by a <a>path-absolute-URL string</a>.

<p>A <dfn export oldids=syntax-url-port>URL-port string</dfn> must be one of the following:

<ul class=brief>
 <li><p>the empty string
 <li><p>one or more <a>ASCII digits</a> representing a decimal number that is a
 <a for=/>16-bit unsigned integer</a>.
</ul>

<p>A <dfn export oldids=syntax-url-scheme-relative>scheme-relative-URL string</dfn> must be
"<code>//</code>", followed by an <a>opaque-host-and-port string</a>, optionally followed by a
<a>path-absolute-URL string</a>.

<p>An <dfn export>opaque-host-and-port string</dfn> must be either the empty string or: a
<a>valid opaque-host string</a>, optionally followed by U+003A (:) and a <a>URL-port string</a>.

<p>A <dfn export oldids=syntax-url-file-scheme-relative>scheme-relative-file-URL string</dfn> must
be "<code>//</code>", followed by one of the following:

<ul class=brief>
 <li><p>a <a>valid host string</a>, optionally followed by a
 <a>path-absolute-non-Windows-file-URL string</a>
 <li><p>a <a>path-absolute-URL string</a>.
</ul>

<p>A <dfn export oldids=syntax-url-path-absolute>path-absolute-URL string</dfn> must be U+002F (/)
followed by a <a>path-relative-URL string</a>.

<p>A <dfn export oldids=syntax-url-file-path-absolute>path-absolute-non-Windows-file-URL string</dfn>
must be a <a>path-absolute-URL string</a> that does not start with: U+002F (/), followed by a
<a>Windows drive letter</a>, followed by U+002F (/).

<p>A <dfn export oldids=syntax-url-path-relative>path-relative-URL string</dfn> must be zero or more
<a>URL-path-segment strings</a>, separated from each other by U+002F (/), and not start with
U+002F (/).

<p>A
<dfn export oldids=syntax-url-path-relative-scheme-less>path-relative-scheme-less-URL string</dfn>
must be a <a>path-relative-URL string</a> that does not start with: a <a>URL-scheme string</a>,
followed by U+003A (:).

<p>A <dfn export oldids=syntax-url-path-segment>URL-path-segment string</dfn> must be one of the
following:

<ul class=brief>
 <li><p>zero or more <a>URL units</a> excluding U+002F (/) and U+003F (?), that together are not a
 <a>single-dot URL path segment</a> or a <a>double-dot URL path segment</a>.
 <li><p>a <a>single-dot URL path segment</a>
 <li><p>a <a>double-dot URL path segment</a>.
</ul>

<p>A <dfn export oldids=syntax-url-query>URL-query string</dfn> must be zero or more <a>URL units</a>.

<p>A <dfn export oldids=syntax-url-fragment>URL-fragment string</dfn> must be zero or more
<a>URL units</a>.

<p>The <dfn export lt="URL code point" id=url-code-points>URL code points</dfn> are
<a>ASCII alphanumeric</a>,
U+0021 (!),<!-- sub-delims -->
U+0024 ($),<!-- sub-delims -->
U+0026 (&amp;),<!-- sub-delims -->
U+0027 ('),<!-- sub-delims -->
U+0028 LEFT PARENTHESIS,<!-- sub-delims -->
U+0029 RIGHT PARENTHESIS,<!-- sub-delims -->
U+002A (*),<!-- sub-delims -->
U+002B (+),<!-- sub-delims -->
U+002C (,),<!-- sub-delims -->
U+002D (-),<!-- iunreserved -->
U+002E (.),<!-- iunreserved -->
U+002F (/),<!-- iquery/ifragment -->
U+003A (:),<!-- ipchar -->
U+003B (;),<!-- sub-delims -->
U+003D (=),<!-- sub-delims -->
U+003F (?),<!-- iquery/ifragment -->
U+0040 (@),<!-- ipchar -->
U+005F (_),<!-- iunreserved -->
U+007E (~),<!-- iunreserved -->
and <a>code points</a> in the range U+00A0 to U+10FFFD, inclusive, excluding <a>surrogates</a> and
<a>noncharacters</a>.
<!-- IRI also excludes the ranges U+E000 to U+F8FF, U+FFF0 to U+FFFD, and U+E0000 to U+E09FF, all
     inclusive. We don't to align with HTML. -->

<p class=note>Code points greater than U+007F DELETE will be converted to
<a lt="percent-encoded byte">percent-encoded bytes</a> by the <a>URL parser</a>.

<p class=note>In HTML, when the document encoding is a legacy encoding, code points in the
<a>URL-query string</a> that are higher than U+007F DELETE will be converted to
<a lt="percent-encoded byte">percent-encoded bytes</a> <em>using the document's encoding</em>. This
can cause problems if a URL that works in one document is copied to another document that uses a
different document encoding. Using the <a>UTF-8</a> encoding everywhere solves this problem.

<div class=example id=query-encoding-example>
 <p>For example, consider this HTML document:

 <pre><code class="lang-html">
 &lt;!doctype html>
 &lt;meta charset="windows-1252">
 &lt;a href="?sm&amp;ouml;rg&amp;aring;sbord">Test&lt;/a></code></pre>

 <p>Since the document encoding is windows-1252, the link's <a for=/>URL</a>'s <a for=url>query</a>
 will be "<code>sm%F6rg%E5sbord</code>". If the document encoding had been UTF-8, it would instead
 be "<code>sm%C3%B6rg%C3%A5sbord</code>".
</div>

<p>The <dfn>URL units</dfn> are <a>URL code points</a> and <a>percent-encoded bytes</a>.

<p class=note><a>Percent-encoded bytes</a> can be used to encode code points that are not
<a>URL code points</a> or are excluded from being written.

<hr>

<p class=note>There is no way to express a <a for=url>username</a> or <a for=url>password</a> of a
<a for=/>URL record</a> within a <a>valid URL string</a>.


<h3 id=url-parsing>URL parsing</h3>

<div algorithm>
<p>The <dfn export id=concept-url-parser lt="URL parser">URL parser</dfn> takes a
<a>scalar value string</a> <var>input</var>, with an optional null or <a>base URL</a>
<var>base</var> (default null) and an optional <a for=/>encoding</a> <var>encoding</var> (default
<a>UTF-8</a>), and then runs these steps:

<p class=note>Non-web-browser implementations only need to implement the <a>basic URL parser</a>.

<p class=note>How user input in the web browser's address bar is converted to a
<a for=/>URL record</a> is out-of-scope of this standard. This standard does include
<a href="#url-rendering">URL rendering requirements</a> as they pertain trust decisions.

<ol>
 <li><p>Let <var>url</var> be the result of running the <a>basic URL parser</a> on <var>input</var>
 with <var>base</var> and <var>encoding</var>.

 <li><p>If <var>url</var> is failure, return failure.

 <li><p>If <var>url</var>'s <a for=url>scheme</a> is not
 "<code>blob</code>", return <var>url</var>.

 <li><p>Set <var>url</var>'s <a for=url>blob URL entry</a> to the result of
 <a lt="resolve a blob URL">resolving the blob URL</a> <var>url</var>, if that did not return
 failure, and null otherwise.

 <li><p>Return <var>url</var>.
</ol>
</div>

<hr>

<div algorithm="basic URL parser">
<p>The <dfn export id=concept-basic-url-parser lt="basic URL parser">basic URL parser</dfn> takes a
<a>scalar value string</a> <var>input</var>, with an optional null or <a>base URL</a>
<var>base</var> (default null), an optional <a for=/>encoding</a> <var>encoding</var> (default
<a>UTF-8</a>), an optional <a for=/>URL</a> <dfn export for="basic URL parser"><var>url</var></dfn>,
and an optional state override <dfn export for="basic URL parser"><var>state override</var></dfn>,
and then runs these steps:

<div class=note>
 <p>The <var>encoding</var> argument is a legacy concept only relevant for <cite>HTML</cite>. The
 <var>url</var> and <var>state override</var> arguments are only for use by various APIs. [[HTML]]
 <!-- HTMLHyperlinkElementUtils, Location, and URL -->

 <p>When the <var>url</var> and <var>state override</var> arguments are not passed, the
 <a>basic URL parser</a> returns either a new <a for=/>URL</a> or failure. If they are passed, the
 algorithm modifies the passed <var>url</var> and can terminate without returning anything.
</div>

<ol>
 <li>
  <p>If <var>url</var> is not given:

  <ol>
   <li><p>Set <var>url</var> to a new <a for=/>URL</a>.

   <li><p>If <var>input</var> contains any leading or trailing <a>C0 control or space</a>,
   <a>invalid-URL-unit</a> <a>validation error</a>.

   <li><p>Remove any leading and trailing <a>C0 control or space</a> from <var>input</var>.
  </ol>

 <li><p>If <var>input</var> contains any <a>ASCII tab or newline</a>, <a>invalid-URL-unit</a>
 <a>validation error</a>.

 <li><p>Remove all <a>ASCII tab or newline</a> from <var>input</var>.

 <li><p>Let <var>state</var> be <var>state override</var>
 if given, or <a>scheme start state</a> otherwise.

 <li><p>Set <var>encoding</var> to the result of <a>getting an output encoding</a> from
 <var>encoding</var>.

 <li><p>Let <var>buffer</var> be the empty string.

 <li><p>Let <var>atSignSeen</var>, <var>insideBrackets</var>, and <var>passwordTokenSeen</var> be
 false.

 <li><p>Let <var>pointer</var> be a <a>pointer</a> for <var>input</var>.

 <li>
  <p>Keep running the following state machine by switching on <var>state</var>. If after a run
  <var>pointer</var> points to the <a>EOF code point</a>, go to the next step. Otherwise, increase
  <var>pointer</var> by 1 and continue with the state machine.

  <dl class=switch>
   <dt><dfn export for="basic URL parser" id=scheme-start-state>scheme start state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is an <a>ASCII alpha</a>,
     append <a>c</a>, <a lt="ASCII lowercase">lowercased</a>, to <var>buffer</var>, and
     set <var>state</var> to <a>scheme state</a>.

     <li><p>Otherwise, if <var>state override</var> is not given, set <var>state</var> to
     <a>no scheme state</a> and decrease <var>pointer</var> by 1.

     <li>
      <p>Otherwise, return failure.
      <!-- API validation error -->

      <p class=note>This indication of failure is used exclusively by the {{Location}} object's
      {{Location/protocol}} setter.
    </ol>

   <dt><dfn export for="basic URL parser" id=scheme-state>scheme state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is an <a>ASCII alphanumeric</a>, U+002B (+), U+002D (-), or U+002E (.),
     append <a>c</a>, <a lt="ASCII lowercase">lowercased</a>, to <var>buffer</var>.

     <li>
      <p>Otherwise, if <a>c</a> is U+003A (:), then:

      <ol>
       <li>
        <p>If <var>state override</var> is given, then:

        <ol>
         <li><p>If <var>url</var>'s <a for=url>scheme</a> is a <a>special scheme</a> and
         <var>buffer</var> is not a <a>special scheme</a>, then return.

         <li><p>If <var>url</var>'s <a for=url>scheme</a> is not a <a>special scheme</a> and
         <var>buffer</var> is a <a>special scheme</a>, then return.

         <li><p>If <var>url</var> <a>includes credentials</a> or has a non-null <a for=url>port</a>,
         and <var>buffer</var> is "<code>file</code>", then return.

         <li><p>If <var>url</var>'s <a for=url>scheme</a> is "<code>file</code>" and its
         <a for=url>host</a> is an <a>empty host</a>, then return.
        </ol>

       <li><p>Set <var>url</var>'s <a for=url>scheme</a> to <var>buffer</var>.

       <li>
         <p>If <var>state override</var> is given, then:

         <ol>
          <li><p>If <var>url</var>'s <a for=url>port</a> is <var>url</var>'s <a for=url>scheme</a>'s
          <a>default port</a>, then set <var>url</var>'s <a for=url>port</a> to null.

          <li><p>Return.
         </ol>

       <li><p>Set <var>buffer</var> to the empty string.

       <li>
        <p>If <var>url</var>'s <a for=url>scheme</a> is "<code>file</code>", then:

        <ol>
         <li><p>If <a>remaining</a> does not start with "<code>//</code>",
         <a>special-scheme-missing-following-solidus</a> <a>validation error</a>.

         <li><p>Set <var>state</var> to <a>file state</a>.
        </ol>

       <li>
        <p>Otherwise, if <var>url</var> <a>is special</a>, <var>base</var> is non-null, and
        <var>base</var>'s <a for=url>scheme</a> is <var>url</var>'s <a for=url>scheme</a>:

        <ol>
         <li><p><a for=/>Assert</a>: <var>base</var> <a>is special</a> (and therefore does not have
         an <a for=url>opaque path</a>).

         <li><p>Set <var>state</var> to <a>special relative or authority state</a>.
        </ol>

       <li><p>Otherwise, if <var>url</var> <a>is special</a>, set <var>state</var> to
       <a>special authority slashes state</a>.

       <li><p>Otherwise, if <a>remaining</a> starts with an U+002F (/), set <var>state</var> to
       <a>path or authority state</a> and increase <var>pointer</var> by 1.

       <li><p>Otherwise, set <var>url</var>'s <a for=url>path</a> to the empty string and set
       <var>state</var> to <a>opaque path state</a>.
      </ol>

     <li><p>Otherwise, if <var>state override</var> is not given, set
     <var>buffer</var> to the empty string, <var>state</var> to
     <a>no scheme state</a>, and start over (from the first code point
     in <var>input</var>).

     <li>
      <p>Otherwise, return failure.
      <!-- API validation error -->

      <p class=note>This indication of failure is used exclusively by the {{Location}} object's
      {{Location/protocol}} setter. Furthermore, the non-failure termination earlier in this state
      is an intentional difference for defining that setter.
    </ol>

   <dt><dfn export for="basic URL parser" id=no-scheme-state>no scheme state</dfn>
   <dd>
    <ol>
     <li><p>If <var>base</var> is null, or <var>base</var> has an <a for=url>opaque path</a> and
     <a>c</a> is not U+0023 (#), <a>missing-scheme-non-relative-URL</a> <a>validation error</a>,
     return failure.

     <li><p>Otherwise, if <var>base</var> has an <a for=url>opaque path</a> and <a>c</a> is
     U+0023 (#), set <var>url</var>'s <a for=url>scheme</a> to
     <var>base</var>'s <a for=url>scheme</a>,
     <var>url</var>'s <a for=url>path</a> to
     <var>base</var>'s <a for=url>path</a>,
     <var>url</var>'s <a for=url>query</a> to
     <var>base</var>'s <a for=url>query</a>,
     <var>url</var>'s <a for=url>fragment</a> to the empty string, and set <var>state</var> to
     <a>fragment state</a>.

     <li><p>Otherwise, if <var>base</var>'s <a for=url>scheme</a> is not "<code>file</code>", set
     <var>state</var> to <a>relative state</a> and decrease <var>pointer</var> by 1.

     <li><p>Otherwise, set <var>state</var> to <a>file state</a> and decrease <var>pointer</var>
     by 1.
    </ol>

   <dt><dfn export for="basic URL parser" id=special-relative-or-authority-state>special relative or authority state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is U+002F (/) and <a>remaining</a> starts with U+002F (/), then set
     <var>state</var> to <a>special authority ignore slashes state</a> and increase
     <var>pointer</var> by 1.

     <li><p>Otherwise, <a>special-scheme-missing-following-solidus</a> <a>validation error</a>, set
     <var>state</var> to <a>relative state</a> and decrease <var>pointer</var> by 1.
    </ol>

   <dt><dfn export for="basic URL parser" id=path-or-authority-state>path or authority state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is U+002F (/), then set <var>state</var> to <a>authority state</a>.

     <li><p>Otherwise, set <var>state</var> to <a>path state</a>, and decrease <var>pointer</var>
     by 1.
    </ol>

   <dt><dfn export for="basic URL parser" id=relative-state>relative state</dfn>
   <dd>
    <ol>
     <li><p>Assert: <var>base</var>'s <a for=url>scheme</a> is not "<code>file</code>".

     <li><p>Set <var>url</var>'s <a for=url>scheme</a> to <var>base</var>'s <a for=url>scheme</a>.

     <li><p>If <a>c</a> is U+002F (/), then set <var>state</var> to <a>relative slash state</a>.

     <li><p>Otherwise, if <var>url</var> <a>is special</a> and <a>c</a> is U+005C (\),
     <a>invalid-reverse-solidus</a> <a>validation error</a>, set <var>state</var> to
     <a>relative slash state</a>.

     <li>
      <p>Otherwise:

      <ol>
       <li><p>Set <var>url</var>'s <a for=url>username</a> to
       <var>base</var>'s <a for=url>username</a>,
       <var>url</var>'s <a for=url>password</a> to
       <var>base</var>'s <a for=url>password</a>,
       <var>url</var>'s <a for=url>host</a> to
       <var>base</var>'s <a for=url>host</a>,
       <var>url</var>'s <a for=url>port</a> to
       <var>base</var>'s <a for=url>port</a>,
       <var>url</var>'s <a for=url>path</a> to a <a for=list>clone</a> of
       <var>base</var>'s <a for=url>path</a>, and
       <var>url</var>'s <a for=url>query</a> to
       <var>base</var>'s <a for=url>query</a>.

       <li><p>If <a>c</a> is U+003F (?), then set <var>url</var>'s <a for=url>query</a> to the empty
       string, and <var>state</var> to <a>query state</a>.

       <li><p>Otherwise, if <a>c</a> is U+0023 (#), set <var>url</var>'s <a for=url>fragment</a> to
       the empty string and <var>state</var> to <a>fragment state</a>.

       <li>
        <p>Otherwise, if <a>c</a> is not the <a>EOF code point</a>:

        <ol>
         <li><p>Set <var>url</var>'s <a for=url>query</a> to null.

         <li><p><a>Shorten</a> <var>url</var>'s <a for=url>path</a>.

         <li><p>Set <var>state</var> to <a>path state</a> and decrease <var>pointer</var> by 1.
        </ol>
      </ol>
    </ol>

   <dt><dfn export for="basic URL parser" id=relative-slash-state>relative slash state</dfn>
   <dd>
    <ol>
     <li>
      <p>If <var>url</var> <a>is special</a> and <a>c</a> is U+002F (/) or U+005C (\), then:

      <ol>
       <li><p>If <a>c</a> is U+005C (\), <a>invalid-reverse-solidus</a>
       <a>validation error</a>.

       <li><p>Set <var>state</var> to <a>special authority ignore slashes state</a>.
      </ol>

     <li><p>Otherwise, if <a>c</a> is U+002F (/), then set <var>state</var> to
     <a>authority state</a>.

     <li><p>Otherwise, set
     <var>url</var>'s <a for=url>username</a> to
     <var>base</var>'s <a for=url>username</a>,
     <var>url</var>'s <a for=url>password</a> to
     <var>base</var>'s <a for=url>password</a>,
     <var>url</var>'s <a for=url>host</a> to
     <var>base</var>'s <a for=url>host</a>,
     <var>url</var>'s <a for=url>port</a> to
     <var>base</var>'s <a for=url>port</a>,
     <var>state</var> to <a>path state</a>, and then, decrease <var>pointer</var> by 1.
    </ol>

   <dt><dfn export for="basic URL parser" id=special-authority-slashes-state>special authority slashes state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is U+002F (/) and <a>remaining</a> starts with U+002F (/), then set
     <var>state</var> to <a>special authority ignore slashes state</a> and increase
     <var>pointer</var> by 1.

     <li><p>Otherwise, <a>special-scheme-missing-following-solidus</a> <a>validation error</a>, set
     <var>state</var> to <a>special authority ignore slashes state</a> and decrease
     <var>pointer</var> by 1.
    </ol>

   <dt><dfn export for="basic URL parser" id=special-authority-ignore-slashes-state>special authority ignore slashes state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is neither U+002F (/) nor U+005C (\), then set <var>state</var> to
     <a>authority state</a> and decrease <var>pointer</var> by 1.

     <li><p>Otherwise, <a>special-scheme-missing-following-solidus</a> <a>validation error</a>.
    </ol>

   <dt><dfn export for="basic URL parser" id=authority-state>authority state</dfn>
   <dd>
    <ol>
     <li>
      <p>If <a>c</a> is U+0040 (@), then:

      <ol>
       <li><p><a>Invalid-credentials</a> <a>validation error</a>.

       <li><p>If <var>atSignSeen</var> is true, then prepend "<code>%40</code>" to
       <var>buffer</var>.

       <li><p>Set <var>atSignSeen</var> to true.

       <li>
        <p>For each <var>codePoint</var> in <var>buffer</var>:

        <ol>
         <li><p>If <var>codePoint</var> is U+003A (:) and <var>passwordTokenSeen</var> is false,
         then set <var>passwordTokenSeen</var> to true and <a for=iteration>continue</a>.

         <li><p>Let <var>encodedCodePoints</var> be the result of running
         <a for="code point">UTF-8 percent-encode</a> <var>codePoint</var> using the
         <a>userinfo percent-encode set</a>.

         <li><p>If <var>passwordTokenSeen</var> is true, then append <var>encodedCodePoints</var> to
         <var>url</var>'s <a for=url>password</a>.

         <li><p>Otherwise, append <var>encodedCodePoints</var> to <var>url</var>'s
         <a for=url>username</a>.
        </ol>

       <li><p>Set <var>buffer</var> to the empty string.
      </ol>

     <li>
      <p>Otherwise, if one of the following is true:

      <ul class=brief>
       <li><p><a>c</a> is the <a>EOF code point</a>, U+002F (/), U+003F (?), or U+0023 (#)
       <li><p><var>url</var> <a>is special</a> and <a>c</a> is U+005C (\)
      </ul>

      <p>then:

      <ol>
       <li><p>If <var>atSignSeen</var> is true and <var>buffer</var> is the empty string,
       <a>host-missing</a> <a>validation error</a>, return failure.
       <!-- No URLs with userinfo, but without host. For special URLs it would also not be
            idempotent:
            https://@/example.org/ -> https:///example.org/ -> https://example.org/ -->

       <li><p>Decrease <var>pointer</var> by <var>buffer</var>'s
       <a for=string>code point length</a> + 1, set <var>buffer</var> to the empty string, and set
       <var>state</var> to <a>host state</a>.
      </ol>

     <li><p>Otherwise, append <a>c</a> to <var>buffer</var>.
    </ol>

   <dt><dfn export for="basic URL parser" id=host-state>host state</dfn>
   <dt><dfn export for="basic URL parser" id=hostname-state>hostname state</dfn>
   <dd>
    <ol>
     <li><p>If <var>state override</var> is given and <var>url</var>'s <a for=url>scheme</a> is
     "<code>file</code>", then decrease <var>pointer</var> by 1 and set <var>state</var> to
     <a>file host state</a>.

     <li>
      <p>Otherwise, if <a>c</a> is U+003A (:) and <var>insideBrackets</var> is false:

      <ol>
       <li><p>If <var>buffer</var> is the empty string, <a>host-missing</a> <a>validation error</a>,
       return failure.
       <!-- No URLs with port, but without host. -->

       <li><p>If <var>state override</var> is given and <var>state override</var> is
       <a>hostname state</a>, then return failure.
       <!-- API validation error -->

       <li><p>Let <var>host</var> be the result of <a>host parsing</a> <var>buffer</var> with
       <var>url</var> <a>is not special</a>.

       <li><p>If <var>host</var> is failure, then return failure.

       <li><p>Set <var>url</var>'s <a for=url>host</a> to
       <var>host</var>, <var>buffer</var> to the empty string,
       and <var>state</var> to <a>port state</a>.
      </ol>

     <li>
      <p>Otherwise, if one of the following is true:

      <ul class=brief>
       <li><p><a>c</a> is the <a>EOF code point</a>, U+002F (/), U+003F (?), or U+0023 (#)
       <li><p><var>url</var> <a>is special</a> and <a>c</a> is U+005C (\)
      </ul>

      <p>then decrease <var>pointer</var> by 1, and:

      <ol>
       <li><p>If <var>url</var> <a>is special</a> and <var>buffer</var> is the empty string,
       <a>host-missing</a> <a>validation error</a>, return failure.
       <!-- http://? -> failure
            test://? -> test://? -->

       <li><p>Otherwise, if <var>state override</var> is given, <var>buffer</var> is the empty
       string, and either <var>url</var> <a>includes credentials</a> or <var>url</var>'s
       <a for=url>port</a> is non-null, then return failure.
       <!-- API validation error -->

       <li><p>Let <var>host</var> be the result of <a>host parsing</a> <var>buffer</var> with
       <var>url</var> <a>is not special</a>.

       <li><p>If <var>host</var> is failure, then return failure.

       <li><p>Set <var>url</var>'s <a for=url>host</a> to
       <var>host</var>, <var>buffer</var> to the empty string,
       and <var>state</var> to <a>path start state</a>.

       <li><p>If <var>state override</var> is given, then return.
      </ol>

     <li>
      <p>Otherwise:

      <ol>
       <li><p>If <a>c</a> is U+005B ([), then set <var>insideBrackets</var> to true.

       <li><p>If <a>c</a> is U+005D (]), then set <var>insideBrackets</var> to false.

       <li><p>Append <a>c</a> to <var>buffer</var>.
      </ol>
    </ol>

   <dt><dfn export for="basic URL parser" id=port-state>port state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is an <a>ASCII digit</a>, append <a>c</a> to <var>buffer</var>.

     <li>
      <p>Otherwise, if one of the following is true:

      <ul class=brief>
       <li><p><a>c</a> is the <a>EOF code point</a>, U+002F (/), U+003F (?), or U+0023 (#);
       <li><p><var>url</var> <a>is special</a> and <a>c</a> is U+005C (\); or
       <li><p><var>state override</var> is given,
      </ul>

      <p>then:

      <ol>
       <li>
        <p>If <var>buffer</var> is not the empty string:

        <ol>
         <li><p>Let <var>port</var> be the mathematical integer value that is represented
         by <var>buffer</var> in radix-10 using <a>ASCII digits</a> for digits with values
         0 through 9.

         <li><p>If <var>port</var> is not a <a for=/>16-bit unsigned integer</a>,
         <a>port-out-of-range</a> <a>validation error</a>, return failure.

         <li><p>Set <var>url</var>'s <a for=url>port</a> to null, if <var>port</var> is
         <var>url</var>'s <a for=url>scheme</a>'s <a>default port</a>; otherwise to <var>port</var>.

         <li><p>Set <var>buffer</var> to the empty string.

         <li><p>If <var>state override</var> is given, then return.
         <!-- API validation error: we'd have to check what c is at this point if we cared about
              this. In particular when it's not EOF it's a bit dodgy. -->
        </ol>

       <li><p>If <var>state override</var> is given, then return failure.
       <!-- API validation error -->

       <li><p>Set <var>state</var> to <a>path start state</a> and decrease <var>pointer</var> by 1.
      </ol>

     <li><p>Otherwise, <a>port-invalid</a> <a>validation error</a>, return failure.
    </ol>

   <dt><dfn export for="basic URL parser" id=file-state>file state</dfn>
   <dd>
    <ol>
     <li><p>Set <var>url</var>'s <a for=url>scheme</a> to "<code>file</code>".

     <li><p>Set <var>url</var>'s <a for=url>host</a> to the empty string.

     <li>
      <p>If <a>c</a> is U+002F (/) or U+005C (\), then:

      <ol>
       <li><p>If <a>c</a> is U+005C (\), <a>invalid-reverse-solidus</a> <a>validation error</a>.

       <li><p>Set <var>state</var> to <a>file slash state</a>.
      </ol>

     <li>
      <p>Otherwise, if <var>base</var> is non-null and <var>base</var>'s <a for=url>scheme</a> is
      "<code>file</code>":

      <ol>
       <li><p>Set <var>url</var>'s <a for=url>host</a> to <var>base</var>'s <a for=url>host</a>,
       <var>url</var>'s <a for=url>path</a> to a <a for=list>clone</a> of <var>base</var>'s
       <a for=url>path</a>, and <var>url</var>'s <a for=url>query</a> to <var>base</var>'s
       <a for=url>query</a>.

       <li><p>If <a>c</a> is U+003F (?), then set <var>url</var>'s <a for=url>query</a> to the empty
       string and <var>state</var> to <a>query state</a>.

       <li><p>Otherwise, if <a>c</a> is U+0023 (#), set <var>url</var>'s <a for=url>fragment</a> to
       the empty string and <var>state</var> to <a>fragment state</a>.

       <li>
        <p>Otherwise, if <a>c</a> is not the <a>EOF code point</a>:

        <ol>
         <li><p>Set <var>url</var>'s <a for=url>query</a> to null.

         <li><p>If the
         <a lt="code point substring to the end of the string">code point substring</a> from
         <var>pointer</var> to the end of <var>input</var> does not
         <a>start with a Windows drive letter</a>, then <a>shorten</a> <var>url</var>'s
         <a for=url>path</a>.

         <li>
          <p>Otherwise:

          <ol>
           <li><p><a>File-invalid-Windows-drive-letter</a> <a>validation error</a>.

           <li><p>Set <var>url</var>'s <a for=url>path</a> to « ».
          </ol>

          <p class=note>This is a (platform-independent) Windows drive letter quirk.

         <li><p>Set <var>state</var> to <a>path state</a> and decrease <var>pointer</var> by 1.
        </ol>
      </ol>

     <li><p>Otherwise, set <var>state</var> to <a>path state</a>, and decrease <var>pointer</var>
     by 1.
    </ol>

   <dt><dfn export for="basic URL parser" id=file-slash-state>file slash state</dfn>
   <dd>
    <ol>
     <li>
      <p>If <a>c</a> is U+002F (/) or U+005C (\), then:

      <ol>
       <li><p>If <a>c</a> is U+005C (\), <a>invalid-reverse-solidus</a> <a>validation error</a>.

       <li><p>Set <var>state</var> to <a>file host state</a>.
      </ol>

     <li>
      <p>Otherwise:

      <ol>
       <li>
        <p>If <var>base</var> is non-null and <var>base</var>'s <a for=url>scheme</a> is
        "<code>file</code>", then:

        <ol>
         <li><p>Set <var>url</var>'s <a for=url>host</a> to <var>base</var>'s <a for=url>host</a>.

         <li>
          <p>If the <a lt="code point substring to the end of the string">code point substring</a>
          from <var>pointer</var> to the end of <var>input</var> does not
          <a>start with a Windows drive letter</a> and <var>base</var>'s <a for=url>path</a>[0] is a
          <a>normalized Windows drive letter</a>, then <a for=list>append</a> <var>base</var>'s
          <a for=url>path</a>[0] to <var>url</var>'s <a for=url>path</a>.

          <p class=note>This is a (platform-independent) Windows drive letter quirk.

        </ol>

       <li><p>Set <var>state</var> to <a>path state</a>, and decrease <var>pointer</var> by 1.
      </ol>
    </ol>

   <dt><dfn export for="basic URL parser" id=file-host-state>file host state</dfn>
   <dd>
    <ol>
     <li>
      <p>If <a>c</a> is the <a>EOF code point</a>, U+002F (/), U+005C (\), U+003F (?), or
      U+0023 (#), then decrease <var>pointer</var> by 1 and then:

      <ol>
       <li>
        <p>If <var>state override</var> is not given and <var>buffer</var> is a
        <a>Windows drive letter</a>, <a>file-invalid-Windows-drive-letter-host</a>
        <a>validation error</a>, set <var>state</var> to <a>path state</a>.

        <p class=note>This is a (platform-independent) Windows drive letter quirk. <var>buffer</var>
        is not reset here and instead used in the <a>path state</a>.

       <li>
        <p>Otherwise, if <var>buffer</var> is the empty string, then:

        <ol>
         <li><p>Set <var>url</var>'s <a for=url>host</a> to the empty string.

         <li><p>If <var>state override</var> is given, then return.

         <li><p>Set <var>state</var> to <a>path start state</a>.
        </ol>

       <li>
        <p>Otherwise, run these steps:

        <ol>
         <li><p>Let <var>host</var> be the result of <a>host parsing</a> <var>buffer</var> with
         <var>url</var> <a>is not special</a>.

         <li><p>If <var>host</var> is failure, then return failure.

         <li><p>If <var>host</var> is "<code title>localhost</code>", then set <var>host</var> to
         the empty string.

         <li><p>Set <var>url</var>'s <a for=url>host</a> to <var>host</var>.

         <li><p>If <var>state override</var> is given, then return.

         <li><p>Set <var>buffer</var> to the empty string and <var>state</var> to
         <a>path start state</a>.
        </ol>
      </ol>

     <li><p>Otherwise, append <a>c</a> to <var>buffer</var>.
    </ol>

   <dt><dfn export for="basic URL parser" id=path-start-state>path start state</dfn>
   <dd>
    <ol>
     <li>
      <p>If <var>url</var> <a>is special</a>, then:

      <ol>
       <li><p>If <a>c</a> is U+005C (\), <a>invalid-reverse-solidus</a> <a>validation error</a>.

       <li><p>Set <var>state</var> to <a>path state</a>.

       <li><p>If <a>c</a> is neither U+002F (/) nor U+005C (\), then decrease <var>pointer</var>
       by 1.
      </ol>

     <li><p>Otherwise, if <var>state override</var> is not given and <a>c</a> is U+003F (?), set
     <var>url</var>'s <a for=url>query</a> to the empty string and <var>state</var> to
     <a>query state</a>.

     <li><p>Otherwise, if <var>state override</var> is not given and <a>c</a> is U+0023 (#), set
     <var>url</var>'s <a for=url>fragment</a> to the empty string and <var>state</var> to
     <a>fragment state</a>.

     <li>
      <p>Otherwise, if <a>c</a> is not the <a>EOF code point</a>:

      <ol>
       <li><p>Set <var>state</var> to <a>path state</a>.

       <li><p>If <a>c</a> is not U+002F (/), then decrease <var>pointer</var> by 1.
      </ol>

     <li><p>Otherwise, if <var>state override</var> is given and <var>url</var>'s
     <a for=url>host</a> is null, <a for=list>append</a> the empty string to <var>url</var>'s
     <a for=url>path</a>.
    </ol>

   <dt><dfn export for="basic URL parser" id=path-state>path state</dfn>
   <dd>
    <ol>
     <li>
      <p>If one of the following is true:

      <ul class=brief>
       <li><p><a>c</a> is the <a>EOF code point</a> or U+002F (/)
       <li><p><var>url</var> <a>is special</a> and <a>c</a> is U+005C (\)
       <li><p><var>state override</var> is not given and <a>c</a> is U+003F (?) or U+0023 (#)
      </ul>

      <p>then:

      <ol>
       <li><p>If <var>url</var> <a>is special</a> and <a>c</a> is U+005C (\),
       <a>invalid-reverse-solidus</a> <a>validation error</a>.

       <li>
        <p>If <var>buffer</var> is a <a>double-dot URL path segment</a>, then:

        <ol>
         <li><p><a>Shorten</a> <var>url</var>'s <a for=url>path</a>.

         <li>
          <p>If neither <a>c</a> is U+002F (/), nor <var>url</var> <a>is special</a> and <a>c</a> is
          U+005C (\), <a for=list>append</a> the empty string to <var>url</var>'s
          <a for=url>path</a>.

          <p class=note>This means that for input <code>/usr/..</code> the result is <code>/</code>
          and not a lack of a path.
        </ol>

       <li><p>Otherwise, if <var>buffer</var> is a <a>single-dot URL path segment</a> and if neither
       <a>c</a> is U+002F (/), nor <var>url</var> <a>is special</a> and <a>c</a> is U+005C (\),
       <a for=list>append</a> the empty string to <var>url</var>'s <a for=url>path</a>.

       <li>
        <p>Otherwise, if <var>buffer</var> is not a <a>single-dot URL path segment</a>, then:

        <ol>
         <li>
          <p>If <var>url</var>'s <a for=url>scheme</a> is "<code>file</code>", <var>url</var>'s
          <a for=url>path</a> <a for=list>is empty</a>, and <var>buffer</var> is a
          <a>Windows drive letter</a>, then replace the second code point in <var>buffer</var> with
          U+003A (:).

          <p class=note>This is a (platform-independent) Windows drive letter quirk.

         <li><p><a for=list>Append</a> <var>buffer</var> to <var>url</var>'s <a for=url>path</a>.
        </ol>

       <li><p>Set <var>buffer</var> to the empty string.

       <li><p>If <a>c</a> is U+003F (?), then set <var>url</var>'s <a for=url>query</a> to the empty
       string and <var>state</var> to <a>query state</a>.

       <li><p>If <a>c</a> is U+0023 (#), then set <var>url</var>'s <a for=url>fragment</a> to the
       empty string and <var>state</var> to <a>fragment state</a>.
      </ol>

     <li>
      <p>Otherwise, run these steps:

      <ol>
       <li><p>If <a>c</a> is not a <a>URL code point</a> and not U+0025 (%),
       <a>invalid-URL-unit</a> <a>validation error</a>.

       <li><p>If <a>c</a> is U+0025 (%) and <a>remaining</a> does not start with two
       <a>ASCII hex digits</a>, <a>invalid-URL-unit</a> <a>validation error</a>.

       <li><p><a for="code point">UTF-8 percent-encode</a> <a>c</a> using the
       <a>path percent-encode set</a> and append the result to <var>buffer</var>.
      </ol>
    </ol>

   <dt><dfn export for="basic URL parser" id=cannot-be-a-base-url-path-state>opaque path state</dfn>
   <dd>
    <ol>
     <li><p>If <a>c</a> is U+003F (?), then set <var>url</var>'s <a for=url>query</a> to the empty
     string and <var>state</var> to <a>query state</a>.

     <li><p>Otherwise, if <a>c</a> is U+0023 (#), then set <var>url</var>'s <a for=url>fragment</a>
     to the empty string and <var>state</var> to <a>fragment state</a>.

     <li>
      <p>Otherwise, if <a>c</a> is U+0020 SPACE:

      <ol>
       <li><p>If <a>remaining</a> starts with U+003F (?) or U+003F (#), then append
       "<code>%20</code>" to <var>url</var>'s <a for=url>path</a>.

       <li><p>Otherwise, append U+0020 SPACE to <var>url</var>'s <a for=url>path</a>.
      </ol>

     <li>
      <p>Otherwise, if <a>c</a> is not the <a>EOF code point</a>:

      <ol>
       <li><p>If <a>c</a> is not a <a>URL code point</a> and not U+0025 (%), <a>invalid-URL-unit</a>
       <a>validation error</a>.

       <li><p>If <a>c</a> is U+0025 (%) and <a>remaining</a> does not start with two
       <a>ASCII hex digits</a>, <a>invalid-URL-unit</a> <a>validation error</a>.

       <li><p><a for="code point">UTF-8 percent-encode</a> <a>c</a> using the
       <a>C0 control percent-encode set</a> and append the result to <var>url</var>'s
       <a for=url>path</a>.
      </ol>
    </ol>

   <dt><dfn export for="basic URL parser" id=query-state>query state</dfn>
   <dd>
    <ol>
     <li>
      <p>If <var>encoding</var> is not <a>UTF-8</a> and one of the following is true:

      <ul class=brief>
       <li><p><var>url</var> <a>is not special</a>
       <li><p><var>url</var>'s <a for=url>scheme</a> is "<code>ws</code>" or "<code>wss</code>"
      </ul>

      <p>then set <var>encoding</var> to <a>UTF-8</a>.
      <!-- https://simon.html5.org/test/url/url-encoding.html -->

     <li>
      <p>If one of the following is true:

      <ul class=brief>
       <li><p><var>state override</var> is not given and <a>c</a> is U+0023 (#)
       <li><p><a>c</a> is the <a>EOF code point</a>
      </ul>

      <p>then:

      <ol>
       <li><p>Let <var>queryPercentEncodeSet</var> be the <a>special-query percent-encode set</a> if
       <var>url</var> <a>is special</a>; otherwise the <a>query percent-encode set</a>.

       <li>
        <p><a for=string>Percent-encode after encoding</a>, with <var>encoding</var>,
        <var>buffer</var>, and <var>queryPercentEncodeSet</var>, and append the result to
        <var>url</var>'s <a for=url>query</a>.

        <p class=note>This operation cannot be invoked code-point-for-code-point due to the stateful
        <a>ISO-2022-JP encoder</a>.

       <li><p>Set <var>buffer</var> to the empty string.

       <li><p>If <a>c</a> is U+0023 (#), then set <var>url</var>'s <a for=url>fragment</a> to
       the empty string and state to <a>fragment state</a>.
      </ol>

     <li>
      <p>Otherwise, if <a>c</a> is not the <a>EOF code point</a>:

      <ol>
       <li><p>If <a>c</a> is not a <a>URL code point</a> and not U+0025 (%),
       <a>invalid-URL-unit</a> <a>validation error</a>.

       <li><p>If <a>c</a> is U+0025 (%) and <a>remaining</a> does not start with two
       <a>ASCII hex digits</a>, <a>invalid-URL-unit</a> <a>validation error</a>.

       <li><p>Append <a>c</a> to <var>buffer</var>.
      </ol>
    </ol>

   <dt><dfn export for="basic URL parser" id=fragment-state>fragment state</dfn>
   <dd>
    <ol>
     <li>
      <p>If <a>c</a> is not the <a>EOF code point</a>, then:

      <ol>
       <li><p>If <a>c</a> is not a <a>URL code point</a> and not U+0025 (%),
       <a>invalid-URL-unit</a> <a>validation error</a>.

       <li><p>If <a>c</a> is U+0025 (%) and <a>remaining</a> does not start with two
       <a>ASCII hex digits</a>, <a>invalid-URL-unit</a> <a>validation error</a>.

       <li><p><a for="code point">UTF-8 percent-encode</a> <a>c</a> using the
       <a>fragment percent-encode set</a> and append the result to <var>url</var>'s
       <a for=url>fragment</a>.
      </ol>
    </ol>
  </dl>

 <li><p>Return <var>url</var>.
</ol>
</div>

<hr>

<div algorithm>
<p>To <dfn export id=set-the-username for=url>set the username</dfn> given a <var>url</var> and
<var>username</var>, set <var>url</var>'s <a for=url>username</a> to the result of running
<a for=string>UTF-8 percent-encode</a> on <var>username</var> using the
<a>userinfo percent-encode set</a>.
</div>

<div algorithm>
<p>To <dfn export id=set-the-password for=url>set the password</dfn> given a <var>url</var> and
<var>password</var>, set <var>url</var>'s <a for=url>password</a> to the result of running
<a for=string>UTF-8 percent-encode</a> on <var>password</var> using the
<a>userinfo percent-encode set</a>.
</div>


<h3 id=url-serializing>URL serializing</h3>

<div algorithm="URL serializer">
<p>The <dfn export id=concept-url-serializer lt="URL serializer">URL serializer</dfn> takes a
<a for=/>URL</a> <var>url</var>, with an optional boolean
<dfn export for="URL serializer"><var>exclude fragment</var></dfn> (default false), and then runs
these steps. They return an <a>ASCII string</a>.

<ol>
 <li><p>Let <var>output</var> be <var>url</var>'s <a for=url>scheme</a> and U+003A (:) concatenated.

 <li>
  <p>If <var>url</var>'s <a for=url>host</a> is non-null:

  <ol>
   <li><p>Append "<code>//</code>" to <var>output</var>.

   <li>
    <p>If <var>url</var> <a>includes credentials</a>, then:

    <ol>
     <li><p>Append <var>url</var>'s <a for=url>username</a> to
     <var>output</var>.

     <li><p>If <var>url</var>'s <a for=url>password</a> is not the empty string, then append
     U+003A (:), followed by <var>url</var>'s <a for=url>password</a>, to <var>output</var>.

     <li><p>Append U+0040 (@) to <var>output</var>.
    </ol>

   <li><p>Append <var>url</var>'s <a for=url>host</a>,
   <a lt="host serializer">serialized</a>, to <var>output</var>.

   <li><p>If <var>url</var>'s <a for=url>port</a> is non-null, append U+003A (:) followed by
   <var>url</var>'s <a for=url>port</a>, <a lt="serialize an integer">serialized</a>, to
   <var>output</var>.
  </ol>

 <li>
  <p>If <var>url</var>'s <a for=url>host</a> is null, <var>url</var> does not have an
  <a for=url>opaque path</a>, <var>url</var>'s <a for=url>path</a>'s <a for=list>size</a> is greater
  than 1, and <var>url</var>'s <a for=url>path</a>[0] is the empty string, then append U+002F (/)
  followed by U+002E (.) to <var>output</var>.

  <p class=note>This prevents <code>web+demo:/.//not-a-host/</code> or
  <code>web+demo:/path/..//not-a-host/</code>, when <a lt="URL parser">parsed</a> and then
  <a lt="URL serializer">serialized</a>, from ending up as <code>web+demo://not-a-host/</code> (they
  end up as <code>web+demo:/.//not-a-host/</code>).

 <li><p>Append the result of <a>URL path serializing</a> <var>url</var> to <var>output</var>.

 <li><p>If <var>url</var>'s <a for=url>query</a> is non-null, append
 U+003F (?), followed by <var>url</var>'s <a for=url>query</a>, to
 <var>output</var>.

 <li><p>If <var>exclude fragment</var> is false and <var>url</var>'s <a for=url>fragment</a> is
 non-null, then append U+0023 (#), followed by <var>url</var>'s <a for=url>fragment</a>, to
 <var>output</var>.

 <li><p>Return <var>output</var>.
</ol>
</div>

<div algorithm>
<p>The <dfn export lt="URL path serializer|URL path serializing">URL path serializer</dfn> takes a
<a for=/>URL</a> <var>url</var> and then runs these steps. They return an <a>ASCII string</a>.

<ol>
 <li><p>If <var>url</var> has an <a for=url>opaque path</a>, then return <var>url</var>'s
 <a for=url>path</a>.

 <li><p>Let <var>output</var> be the empty string.

 <li><p><a for=list>For each</a> <var>segment</var> of <var>url</var>'s <a for=url>path</a>: append
 U+002F (/) followed by <var>segment</var> to <var>output</var>.

 <li><p>Return <var>output</var>.
</ol>
</div>


<h3 id=url-equivalence>URL equivalence</h3>

<div algorithm="equal">
<p>To determine whether a <a for=/>URL</a> <var>A</var>
<dfn export for=url id=concept-url-equals lt=equal>equals</dfn> <a for=/>URL</a> <var>B</var>, with
an optional boolean <dfn export for="url/equals"><var>exclude fragments</var></dfn> (default false),
run these steps:

<ol>
 <li><p>Let <var>serializedA</var> be the result of <a lt="URL serializer">serializing</a>
 <var>A</var>, with <a for="URL serializer"><i>exclude fragment</i></a> set to
 <var>exclude fragments</var>.

 <li><p>Let <var>serializedB</var> be the result of <a lt="URL serializer">serializing</a>
 <var>B</var>, with <a for="URL serializer"><i>exclude fragment</i></a> set to
 <var>exclude fragments</var>.

 <li><p>Return true if <var>serializedA</var> is <var>serializedB</var>; otherwise false.
</ol>
</div>


<h3 id=origin>Origin</h3>
<!-- Still need to watch the final bits -->

<p class=note>See <a for=/>origin</a>'s definition in <cite>HTML</cite> for the necessary background
information. [[HTML]]

<div algorithm>
<p>The <dfn export for=url id=concept-url-origin>origin</dfn> of a <a for=/>URL</a> <var>url</var>
is the <a for=/>origin</a> returned by running these steps, switching on <var>url</var>'s
<a for=url>scheme</a>:

<dl class=switch>
 <dt>"<code>blob</code>"
 <dd>
  <ol>
   <li><p>If <var>url</var>'s <a for=url>blob URL entry</a> is non-null, then return
   <var>url</var>'s <a for=url>blob URL entry</a>'s <a for="blob URL entry">environment</a>'s
   <a for="environment settings object">origin</a>.

   <li><p>Let <var>pathURL</var> be the result of <a lt="basic URL parser">parsing</a> the result of
   <a>URL path serializing</a> <var>url</var>.

   <li><p>If <var>pathURL</var> is failure, then return a new <a>opaque origin</a>.

   <li><p>If <var>pathURL</var>'s <a for=url>scheme</a> is "<code>http</code>",
   "<code>https</code>", or "<code>file</code>", then return <var>pathURL</var>'s
   <a for=url>origin</a>.
   <!-- Did you mean: recursion -->

   <li><p>Return a new <a>opaque origin</a>.
  </ol>

  <p class=example id=example-43b5cea5>The <a for=url>origin</a> of
  <code>blob:https://whatwg.org/d0360e2f-caee-469f-9a2f-87d5b0456f6f</code> is the
  <a for=/>tuple origin</a> ("<code>https</code>", "<code>whatwg.org</code>", null, null).

 <dt>"<code>ftp</code>"
 <dt>"<code>http</code>"
 <dt>"<code>https</code>"
 <dt>"<code>ws</code>"
 <dt>"<code>wss</code>"
 <dd><p>Return the <a for=/>tuple origin</a> (<var>url</var>'s <a for=url>scheme</a>,
 <var>url</var>'s <a for=url>host</a>, <var>url</var>'s <a for=url>port</a>, null).

 <dt>"<code>file</code>"
 <dd><p>Unfortunate as it is, this is left as an exercise to the reader. When in doubt,
 return a new <a>opaque origin</a>.

 <dt>Otherwise
 <dd>
  <p>Return a new <a>opaque origin</a>.

  <p class=note>This does indeed mean that these <a for=/>URLs</a> cannot be <a>same origin</a> with
  themselves.
</dl>
</div>


<h3 id=url-rendering>URL rendering</h3>
<!-- See https://www.w3.org/Bugs/Public/show_bug.cgi?id=27641 for context -->

<p>A <a for=/>URL</a> should be rendered in its <a lt="URL serializer">serialized</a> form, with
modifications described below, when the primary purpose of displaying a URL is to have the user make
a security or trust decision. For example, users are expected to make trust decisions based on a URL
rendered in the browser address bar.

<h4 id=url-rendering-simplification>Simplify non-human-readable or irrelevant components</h4>

<p>Remove components that can provide opportunities for spoofing or distract from security-relevant
information:

<ul>
 <li><p>Browsers may render only a URL's <a for=url>host</a> in places where it is important for end
 users to distinguish between the host and other parts of the URL such as the <a for=url>path</a>.
 Browsers may consider simplifying the host further to draw attention to its
 <a for=host>registrable domain</a>. For example, browsers may omit a leading <code>www</code> or
 <code>m</code> <a for=/>domain label</a> to simplify the host, or display its registrable domain
 only to remove spoofing opportunities posted by subdomains (e.g.,
 <code>https://examplecorp.attacker.com/</code>).

 <li><p>Browsers should not render a <a for=/>URL</a>'s <a for=url>username</a> and <a
 for=url>password</a>, as they can be mistaken for a <a for=/>URL</a>'s <a for=url>host</a> (e.g.,
 <code>https://examplecorp.com@attacker.example/</code>).

 <li><p>Browsers may render a URL without its <a for=url>scheme</a> if the display surface only ever
 permits a single scheme (such as a browser feature that omits <code>https://</code> because it is
 only enabled for secure origins). Otherwise, the scheme may be replaced or supplemented with a
 human-readable string (e.g., "Not secure"), a security indicator icon, or both.
</ul>

<h4 id=url-rendering-elision>Elision</h4>

<p>In a space-constrained display, URLs should be elided carefully to avoid misleading the user when
making a security decision:

<ul>
 <li><p>Browsers should ensure that at least the <a for=host>registrable domain</a> can be shown
 when the URL is rendered (to avoid showing, e.g., <code>...examplecorp.com</code> when loading
 <code>https://not-really-examplecorp.com/</code>).

 <li><p>When the full <a for=url>host</a> cannot be rendered, browsers should elide
 <a for=/>domain labels</a> starting from the lowest-level domain label. For example,
 <code>examplecorp.com.evil.com</code> should be elided as <code>...com.evil.com</code>, not
 <code>examplecorp.com...</code>. (Note that bidirectional text means that the lowest-level domain
 label may not appear on the left.)
</ul>

<h4 id=url-rendering-i18n>Internationalization and special characters</h4>

<p>Internationalized domain names (IDNs), special characters, and bidirectional text should be
handled with care to prevent spoofing:

<ul>
 <li>
  <p>Browsers should render a <a for=/>URL</a>'s <a for=url>host</a> by running
  <a>domain to Unicode</a> with the <a for=/>URL</a>'s <a for=url>host</a> and false.

  <p class=note>Various characters can be used in homograph spoofing attacks. Consider detecting
  confusable characters and warning when they are in use. [[IDNFAQ]] [[UTS39]]

 <li><p>URLs are particularly prone to confusion between host and path when they contain
 bidirectional text, so in this case it is particularly advisable to only render a URL's
 <a for=url>host</a>. For readability, other parts of the <a for=/>URL</a>, if rendered, should have
 their sequences of <a>percent-encoded bytes</a> replaced with code points resulting from running
 <a>UTF-8 decode without BOM</a> on the <a for=string>percent-decoding</a> of those sequences,
 unless that renders those sequences invisible. Browsers may choose to not decode certain sequences
 that present spoofing risks (e.g., U+1F512 (🔒)).

 <li>
  <p>Browsers should render bidirectional text as if it were in a left-to-right embedding. [[!BIDI]]

  <p class=note>Unfortunately, as rendered <a for=/>URLs</a> are strings and can appear anywhere, a
  specific bidirectional algorithm for rendered <a for=/>URLs</a> would not see wide adoption.
  Bidirectional text interacts with the parts of a <a for=/>URL</a> in ways that can cause the
  rendering to be different from the model. Users of bidirectional languages can come to expect
  this, particularly in plain text environments.
</ul>



<h2 id="application/x-www-form-urlencoded"><code>application/x-www-form-urlencoded</code></h2>

<p>The <dfn export id=concept-urlencoded><code>application/x-www-form-urlencoded</code></dfn> format
provides a way to encode a <a for=/>list</a> of <a for=/>tuples</a>, each consisting of a name and a
value.

<p class=note>The <code>application/x-www-form-urlencoded</code> format is in many ways an aberrant
monstrosity, the result of many years of implementation accidents and compromises leading to a set
of requirements necessary for interoperability, but in no way representing good design practices. In
particular, readers are cautioned to pay close attention to the twisted details involving repeated
(and in some cases nested) conversions between character encodings and byte sequences. Unfortunately
the format is in widespread use due to the prevalence of HTML forms. [[HTML]]


<h3 id=urlencoded-parsing><code>application/x-www-form-urlencoded</code> parsing</h3>

<p class=note>A legacy server-oriented implementation might have to support <a for=/>encodings</a>
other than <a>UTF-8</a> as well as have special logic for tuples of which the name is
`<code>_charset</code>`. Such logic is not described here as only <a>UTF-8</a> is conforming.

<div algorithm>
<p>The
<dfn export id=concept-urlencoded-parser lt="urlencoded parser"><code>application/x-www-form-urlencoded</code> parser</dfn>
takes a byte sequence <var>input</var>, and then runs these steps:

<ol>
 <li><p>Let <var>sequences</var> be the result of splitting <var>input</var> on
 0x26 (&amp;).
 <!-- XXX either define strictly splitting for byte sequences in Infra, or investigate whether
      UTF-8 decoding can be done before this step rather than after. -->

 <li><p>Let <var>output</var> be an initially empty <a for=/>list</a> of name-value tuples where
 both name and value hold a string.

 <li>
  <p><a for=list>For each</a> byte sequence <var>bytes</var> in <var>sequences</var>:

  <ol>
   <li><p>If <var>bytes</var> is the empty byte sequence, then <a for=iteration>continue</a>.

   <li><p>If <var>bytes</var> contains a 0x3D (=), then let
   <var>name</var> be the bytes from the start of <var>bytes</var> up to but
   excluding its first 0x3D (=), and let <var>value</var> be the
   bytes, if any, after the first 0x3D (=) up to the end of
   <var>bytes</var>. If 0x3D (=) is the first byte, then
   <var>name</var> will be the empty byte sequence. If it is the last, then
   <var>value</var> will be the empty byte sequence.

   <li><p>Otherwise, let <var>name</var> have the value of <var>bytes</var>
   and let <var>value</var> be the empty byte sequence.

   <li><p>Replace any 0x2B (+) in <var>name</var> and <var>value</var> with 0x20 (SP).

   <li><p>Let <var>nameString</var> and <var>valueString</var> be the result of running <a>UTF-8
   decode without BOM</a> on the <a lt=percent-decode for="byte sequence">percent-decoding</a> of
   <var>name</var> and <var>value</var>, respectively.

   <li><p><a for=list>Append</a> (<var>nameString</var>, <var>valueString</var>) to
   <var>output</var>.
  </ol>

 <li><p>Return <var>output</var>.
</ol>
</div>


<h3 id=urlencoded-serializing><code>application/x-www-form-urlencoded</code> serializing</h3>

<div algorithm>
<p>The
<dfn export id=concept-urlencoded-serializer lt="urlencoded serializer"><code>application/x-www-form-urlencoded</code> serializer</dfn>
takes a list of name-value tuples <var>tuples</var>, with an optional <a for=/>encoding</a>
<var>encoding</var> (default <a>UTF-8</a>), and then runs these steps. They return an
<a>ASCII string</a>.

<ol>
 <li><p>Set <var>encoding</var> to the result of <a>getting an output encoding</a> from
 <var>encoding</var>.

 <li><p>Let <var>output</var> be the empty string.

 <li>
  <p><a for=list>For each</a> <var>tuple</var> of <var>tuples</var>:

  <ol>
   <li><p><a for=/>Assert</a>: <var>tuple</var>'s name and <var>tuple</var>'s value are
   <a for=/>scalar value strings</a>.

   <li><p>Let <var>name</var> be the result of running
   <a for=string>percent-encode after encoding</a> with <var>encoding</var>, <var>tuple</var>'s
   name, and the <a><code>application/x-www-form-urlencoded</code> percent-encode set</a>.

   <li><p>Let <var>value</var> be the result of running
   <a for=string>percent-encode after encoding</a> with <var>encoding</var>, <var>tuple</var>'s
   value, and the <a><code>application/x-www-form-urlencoded</code> percent-encode set</a>.

   <li><p>If <var>output</var> is not the empty string, then append U+0026 (&amp;) to
   <var>output</var>.

   <li>Append <var>name</var>, followed by U+003D (=), followed by <var>value</var>, to
   <var>output</var>.
  </ol>

 <li>Return <var>output</var>.
</ol>
</div>


<h3 id=urlencoded-hooks>Hooks</h3>

<p>The
<dfn id=concept-urlencoded-string-parser lt="urlencoded string parser"><code>application/x-www-form-urlencoded</code> string parser</dfn>
takes a <a>scalar value string</a> <var>input</var>, <a>UTF-8 encodes</a> it, and then returns the
result of <a lt="urlencoded parser"><code>application/x-www-form-urlencoded</code> parsing</a> it.



<h2 id=api>API</h2>

<p>This section uses terminology from <cite>Web IDL</cite>. Browser user agents must support this
API. JavaScript implementations should support this API. Other user agents or programming languages
are encouraged to use an API suitable to their needs, which might not be this one. [[!WEBIDL]]


<h3 id=url-class>URL class</h3>

<pre class=idl>
[Exposed=*,
 LegacyWindowAlias=webkitURL]
interface URL {
  constructor(USVString url, optional USVString base);

  static URL? parse(USVString url, optional USVString base);
  static boolean canParse(USVString url, optional USVString base);

  stringifier attribute USVString href;
  readonly attribute USVString origin;
           attribute USVString protocol;
           attribute USVString username;
           attribute USVString password;
           attribute USVString host;
           attribute USVString hostname;
           attribute USVString port;
           attribute USVString pathname;
           attribute USVString search;
  [SameObject] readonly attribute URLSearchParams searchParams;
           attribute USVString hash;

  USVString toJSON();
};
</pre>

<p>A {{URL}} object has an associated:

<ul class=brief>
 <li><dfn id=concept-url-url noexport for=URL>URL</dfn>: a <a for=/>URL</a>.
 <li><dfn id=concept-url-query-object noexport for=URL>query object</dfn>: a {{URLSearchParams}}
 object.
</ul>

<div algorithm>
<p>The <dfn>API URL parser</dfn> takes a <a>scalar value string</a> <var>url</var> and an optional
null-or-<a>scalar value string</a> <var>base</var> (default null), and then runs these steps:

<ol>
 <li><p>Let <var>parsedBase</var> be null.

 <li>
  <p>If <var>base</var> is non-null:

  <ol>
   <li><p>Set <var>parsedBase</var> to the result of running the <a>basic URL parser</a> on
   <var>base</var>.

   <li><p>If <var>parsedBase</var> is failure, then return failure.
  </ol>

 <li><p>Return the result of running the <a>basic URL parser</a> on <var>url</var> with
 <var>parsedBase</var>.
</ol>
</div>

<div algorithm>
<p>To <dfn for=URL>initialize</dfn> a {{URL}} object <var>url</var> with a <a for=/>URL</a>
<var>urlRecord</var>:

<ol>
 <li><p>Let <var>query</var> be <var>urlRecord</var>'s <a for=url>query</a>, if that is non-null;
 otherwise the empty string.

 <li><p>Set <var>url</var>'s <a for=URL>URL</a> to <var>urlRecord</var>.

 <li><p>Set <var>url</var>'s <a for=URL>query object</a> to a new {{URLSearchParams}} object.

 <li><p><a for=URLSearchParams>Initialize</a> <var>url</var>'s <a for=URL>query object</a> with
 <var>query</var>.

 <li><p>Set <var>url</var>'s <a for=URL>query object</a>'s <a for=URLSearchParams>URL object</a> to
 <var>url</var>.
</ol>
</div>

<div algorithm="URL/extract an origin">
<p>Objects implementing the {{URL}} interface's <a for="platform object">extract an origin</a> steps are
to return <a>this</a>'s <a for=URL>URL</a>'s <a for=url>origin</a>. [[!HTML]]
</div>

<hr>

<div algorithm>
<p id=constructors>The
<dfn constructor for=URL lt="URL(url, base)"><code>new URL(<var>url</var>, <var>base</var>)</code></dfn>
constructor steps are:

<ol>
 <li><p>Let <var>parsedURL</var> be the result of running the <a>API URL parser</a> on
 <var>url</var> with <var>base</var>, if given.

 <li><p>If <var>parsedURL</var> is failure, then <a>throw</a> a {{TypeError}}.

 <li><p><a for=URL>Initialize</a> <a>this</a> with <var>parsedURL</var>.
</ol>
</div>

<div class=example id=example-5434421b>
 <p>To <a lt="basic URL parser">parse</a> a string into a <a for=/>URL</a> without using a
 <a>base URL</a>, invoke the {{URL}} constructor with a single argument:

 <pre><code class="lang-javascript">
var input = "https://example.org/💩",
    url = new URL(input)
url.pathname // "/%F0%9F%92%A9"</code></pre>

 <p>This throws an exception if the input is a <a>relative-URL string</a>:

 <pre><code class="lang-javascript">
try {
  var url = new URL("/🍣🍺")
} catch(e) {
  // that happened
}</code></pre>

 <p>For those cases a <a>base URL</a> is necessary:

 <pre><code class="lang-javascript">
var input = "/🍣🍺",
    url = new URL(input, document.baseURI)
url.href // "https://url.spec.whatwg.org/%F0%9F%8D%A3%F0%9F%8D%BA"</code></pre>

 <p>A {{URL}} object can be used as a <a>base URL</a> (as the IDL requires a string as argument, a
 {{URL}} object stringifies to its {{URL/href}} getter return value):</p>

 <pre><code class="lang-javascript">
var url = new URL("🏳️‍🌈", new URL("https://pride.example/hello-world"))
url.pathname // "/%F0%9F%8F%B3%EF%B8%8F%E2%80%8D%F0%9F%8C%88"</code></pre>
</div>

<hr>

<div algorithm>
<p>The static <dfn method for=URL><code>parse(<var>url</var>, <var>base</var>)</code></dfn> method
steps are:

<ol>
 <li><p>Let <var>parsedURL</var> be the result of running the <a>API URL parser</a> on
 <var>url</var> with <var>base</var>, if given.

 <li><p>If <var>parsedURL</var> is failure, then return null.

 <li><p>Let <var>url</var> be a new {{URL}} object.

 <li><p><a for=URL>Initialize</a> <var>url</var> with <var>parsedURL</var>.

 <li><p>Return <var>url</var>.
</ol>
</div>

<div algorithm>
<p>The static <dfn method for=URL><code>canParse(<var>url</var>, <var>base</var>)</code></dfn>
method steps are:

<ol>
 <li><p>Let <var>parsedURL</var> be the result of running the <a>API URL parser</a> on
 <var>url</var> with <var>base</var>, if given.

 <li><p>If <var>parsedURL</var> is failure, then return false.

 <li><p>Return true.
</ol>
</div>

<hr id=urlutils-members>

<div algorithm="href getter">
<p>The <dfn attribute for=URL><code>href</code></dfn> getter steps and the
<dfn method for=URL><code>toJSON()</code></dfn> method steps are to return the
<a lt="URL serializer">serialization</a> of <a>this</a>'s <a for=URL>URL</a>.
</div>

<div algorithm="href setter">
<p>The <code><a attribute for=URL>href</a></code> setter steps are:

<ol>
 <li><p>Let <var>parsedURL</var> be the result of running the <a>basic URL parser</a> on the given
 value.

 <li><p>If <var>parsedURL</var> is failure, then <a>throw</a> a {{TypeError}}.

 <li><p>Set <a>this</a>'s <a for=URL>URL</a> to <var>parsedURL</var>.

 <li><p>Empty <a>this</a>'s <a for=URL>query object</a>'s <a for=URLSearchParams>list</a>.

 <li><p>Let <var>query</var> be <a>this</a>'s <a for=URL>URL</a>'s <a for=url>query</a>.

 <li><p>If <var>query</var> is non-null, then set <a>this</a>'s
 <a for=URL>query object</a>'s <a for=URLSearchParams>list</a> to the result of
 <a lt="urlencoded string parser">parsing</a> <var>query</var>.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>origin</code></dfn> getter steps are to return the
<a lt="serialization of an origin">serialization</a> of <a>this</a>'s <a for=URL>URL</a>'s
<a for=url>origin</a>. [[!HTML]]
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>protocol</code></dfn> getter steps are to return <a>this</a>'s
<a for=URL>URL</a>'s <a for=url>scheme</a>, followed by U+003A (:).
</div>

<div algorithm="protocol setter">
<p>The <code><a attribute for=URL>protocol</a></code> setter steps are to
<a lt="basic URL parser">basic URL parse</a> the given value, followed by U+003A (:), with
<a>this</a>'s <a for=URL>URL</a> as <a for="basic URL parser"><i>url</i></a> and
<a>scheme start state</a> as <a for="basic URL parser"><i>state override</i></a>.
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>username</code></dfn> getter steps are to return <a>this</a>'s
<a for=URL>URL</a>'s <a for=url>username</a>.
</div>

<div algorithm="username setter">
<p>The <code><a attribute for=URL>username</a></code> setter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a> <a>cannot have a username/password/port</a>, then
 return.

 <li><p><a for=url>Set the username</a> given <a>this</a>'s <a for=URL>URL</a> and the given value.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>password</code></dfn> getter steps are to return <a>this</a>'s
<a for=URL>URL</a>'s <a for=url>password</a>.
</div>

<div algorithm="password setter">
<p>The <code><a attribute for=URL>password</a></code> setter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a> <a>cannot have a username/password/port</a>, then
 return.

 <li><p><a for=url>Set the password</a> given <a>this</a>'s <a for=URL>URL</a> and the given value.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>host</code></dfn> getter steps are:

<ol>
 <li><p>Let <var>url</var> be <a>this</a>'s <a for=URL>URL</a>.

 <li><p>If <var>url</var>'s <a for=url>host</a> is null, then return the empty string.

 <li><p>If <var>url</var>'s <a for=url>port</a> is null, return <var>url</var>'s
 <a for=url>host</a>, <a lt="host serializer">serialized</a>.

 <li><p>Return <var>url</var>'s <a for=url>host</a>, <a lt="host serializer">serialized</a>,
 followed by U+003A (:) and <var>url</var>'s <a for=url>port</a>,
 <a lt="serialize an integer">serialized</a>.
</ol>
</div>

<div algorithm="host setter">
<p>The <code><a attribute for=URL>host</a></code> setter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a> has an <a for=url>opaque path</a>, then return.

 <li><p><a lt="basic URL parser">Basic URL parse</a> the given value with <a>this</a>'s
 <a for=URL>URL</a> as <a for="basic URL parser"><i>url</i></a> and <a>host state</a> as
 <a for="basic URL parser"><i>state override</i></a>.
</ol>

<p class=note>If the given value for the <code><a attribute for=URL>host</a></code> setter lacks a
<a lt="URL-port string">port</a>, <a>this</a>'s <a for=URL>URL</a>'s <a for=url>port</a> will not
change. This can be unexpected as <code>host</code> getter does return a <a>URL-port string</a> so
one might have assumed the setter to always "reset" both.
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>hostname</code></dfn> getter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a>'s <a for=url>host</a> is null, then return the empty
 string.

 <li><p>Return <a>this</a>'s <a for=URL>URL</a>'s <a for=url>host</a>,
 <a lt="host serializer">serialized</a>.
</ol>
</div>

<div algorithm="hostname setter">
<p>The <code><a attribute for=URL>hostname</a></code> setter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a> has an <a for=url>opaque path</a>, then return.

 <li><p><a lt="basic URL parser">Basic URL parse</a> the given value with <a>this</a>'s
 <a for=URL>URL</a> as <a for="basic URL parser"><i>url</i></a> and <a>hostname state</a> as
 <a for="basic URL parser"><i>state override</i></a>.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>port</code></dfn> getter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a>'s <a for=url>port</a> is null, then return the empty
 string.

 <li><p>Return <a>this</a>'s <a for=URL>URL</a>'s <a for=url>port</a>,
 <a lt="serialize an integer">serialized</a>.
</ol>
</div>

<div algorithm="port setter">
<p>The <code><a attribute for=URL>port</a></code> setter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a> <a>cannot have a username/password/port</a>, then
 return.

 <li><p>If the given value is the empty string, then set <a>this</a>'s <a for=URL>URL</a>'s
 <a for=url>port</a> to null.</p></li>

 <li><p>Otherwise, <a lt="basic URL parser">basic URL parse</a> the given value with
 <a>this</a>'s <a for=URL>URL</a> as <a for="basic URL parser"><i>url</i></a> and
 <a>port state</a> as <a for="basic URL parser"><i>state override</i></a>.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>pathname</code></dfn> getter steps are to return the result of
<a>URL path serializing</a> <a>this</a>'s <a for=URL>URL</a>.
</div>

<div algorithm="pathname setter">
<p>The <code><a attribute for=URL>pathname</a></code> setter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a> has an <a for=url>opaque path</a>, then return.

 <li><p><a for=list>Empty</a> <a>this</a>'s <a for=URL>URL</a>'s <a for=url>path</a>.

 <li><p><a lt="basic URL parser">Basic URL parse</a> the given value with <a>this</a>'s
 <a for=URL>URL</a> as <a for="basic URL parser"><i>url</i></a> and <a>path start state</a> as
 <a for="basic URL parser"><i>state override</i></a>.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>search</code></dfn> getter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a>'s <a for=url>query</a> is either null or the empty
 string, then return the empty string.

 <li><p>Return U+003F (?), followed by <a>this</a>'s <a for=URL>URL</a>'s <a for=url>query</a>.
</ol>
</div>

<div algorithm="search setter">
<p>The <code><a attribute for=URL>search</a></code> setter steps are:

<ol>
 <li><p>Let <var>url</var> be <a>this</a>'s <a for=URL>URL</a>.

 <li><p>If the given value is the empty string, then set <var>url</var>'s <a for=url>query</a> to
 null, <a for=list>empty</a> <a>this</a>'s <a for=URL>query object</a>'s
 <a for=URLSearchParams>list</a>, and return.

 <li><p>Let <var>input</var> be the given value with a single leading U+003F (?) removed, if any.

 <li><p>Set <var>url</var>'s <a for=url>query</a> to the empty string.

 <li><p><a lt="basic URL parser">Basic URL parse</a> <var>input</var> with <var>url</var> as
 <a for="basic URL parser"><i>url</i></a> and <a>query state</a> as
 <a for="basic URL parser"><i>state override</i></a>.

 <li><p>Set <a>this</a>'s <a for=URL>query object</a>'s <a for=URLSearchParams>list</a> to the
 result of <a lt="urlencoded string parser">parsing</a> <var>input</var>.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>searchParams</code></dfn> getter steps are to return
<a>this</a>'s <a for=URL>query object</a>.
</div>

<div algorithm>
<p>The <dfn attribute for=URL><code>hash</code></dfn> getter steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URL>URL</a>'s  <a for=url>fragment</a> is either null or the empty
 string, then return the empty string.

 <li><p>Return U+0023 (#), followed by <a>this</a>'s <a for=URL>URL</a>'s <a for=url>fragment</a>.
</ol>
</div>

<div algorithm="hash setter">
<p>The <code><a attribute for=URL>hash</a></code> setter steps are:

<ol>
 <li><p>If the given value is the empty string, then set <a>this</a>'s <a for=URL>URL</a>'s
 <a for=url>fragment</a> to null and return.

 <li><p>Let <var>input</var> be the given value with a single leading U+0023 (#) removed, if any.

 <li><p>Set <a>this</a>'s <a for=URL>URL</a>'s <a for=url>fragment</a> to the empty string.

 <li><p><a lt="basic URL parser">Basic URL parse</a> <var>input</var> with <a>this</a>'s
 <a for=URL>URL</a> as <a for="basic URL parser"><i>url</i></a> and <a>fragment state</a> as
 <a for="basic URL parser"><i>state override</i></a>.
</ol>
</div>


<h3 id=interface-urlsearchparams>URLSearchParams class</h3>

<pre class=idl>
[Exposed=*]
interface URLSearchParams {
  constructor(optional (sequence&lt;sequence&lt;USVString>> or record&lt;USVString, USVString> or USVString) init = "");

  readonly attribute unsigned long size;

  undefined append(USVString name, USVString value);
  undefined delete(USVString name, optional USVString value);
  USVString? get(USVString name);
  sequence&lt;USVString> getAll(USVString name);
  boolean has(USVString name, optional USVString value);
  undefined set(USVString name, USVString value);

  undefined sort();

  iterable&lt;USVString, USVString>;
  stringifier;
};
</pre>

<div class=example id=example-constructing-urlsearchparams>
 <p>Constructing and stringifying a {{URLSearchParams}} object is fairly straightforward:

 <pre><code class="lang-javascript">
let params = new URLSearchParams({key: "730d67"})
params.toString() // "key=730d67"</code></pre>
</div>

<div class=note>
 <p>As a {{URLSearchParams}} object uses the <a><code>application/x-www-form-urlencoded</code></a>
 format underneath there are some difference with how it encodes certain code points compared to a
 {{URL}} object (including {{URL/href}} and {{URL/search}}). This can be especially surprising when
 using {{URL/searchParams}} to operate on a <a for=/>URL</a>'s <a for=url>query</a>.

 <pre><code class="lang-javascript">
const url = new URL('https://example.com/?a=b ~');
console.log(url.href);   // "https://example.com/?a=b%20~"
url.searchParams.sort();
console.log(url.href);   // "https://example.com/?a=b+%7E"</code></pre>

 <pre><code class="lang-javascript">
const url = new URL('https://example.com/?a=~&amp;b=%7E');
console.log(url.search);                // "?a=~&amp;b=%7E"
console.log(url.searchParams.get('a')); // "~"
console.log(url.searchParams.get('b')); // "~"</code></pre>

 <p>{{URLSearchParams}} objects will percent-encode anything in the
 <a><code>application/x-www-form-urlencoded</code> percent-encode set</a>, and will encode
 U+0020 SPACE as U+002B (+).

 <p>Ignoring encodings (use <a>UTF-8</a>), {{URL/search}} will percent-encode anything in the
 <a>query percent-encode set</a> or the <a>special-query percent-encode set</a> (depending on
 whether or not the <a for=/>URL</a> <a>is special</a>).
</div>

<p>A {{URLSearchParams}} object has an associated:

<ul class=brief>
 <li><dfn export for=URLSearchParams id=concept-urlsearchparams-list>list</dfn>: a <a for=/>list</a>
 of <a for=/>tuples</a> each consisting of a name and a value, initially empty.
 <li><dfn export for=URLSearchParams id=concept-urlsearchparams-url-object>URL object</dfn>: null or
 a {{URL}} object, initially null.
</ul>

<div algorithm>
<p>To <dfn for=URLSearchParams oldids=concept-urlsearchparams-new>initialize</dfn> a
{{URLSearchParams}} object <var>query</var> with <var>init</var>:

<ol>
 <li>
  <p>If <var>init</var> is a <a>sequence</a>, then <a for=list>for each</a> <var>innerSequence</var>
  of <var>init</var>:

  <ol>
   <li><p>If <var>innerSequence</var>'s <a for=list>size</a> is not 2, then <a>throw</a> a
   {{TypeError}}.

   <li><p><a for=list>Append</a> (<var>innerSequence</var>[0], <var>innerSequence</var>[1]) to
   <var>query</var>'s <a for=URLSearchParams>list</a>.
  </ol>

 <li><p>Otherwise, if <var>init</var> is a <a for=/>record</a>, then <a for=map>for each</a>
 <var>name</var> → <var>value</var> of <var>init</var>, <a for=list>append</a> (<var>name</var>,
 <var>value</var>) to <var>query</var>'s <a for=URLSearchParams>list</a>.

 <li>
  <p>Otherwise:

  <ol>
   <li><p>Assert: <var>init</var> is a string.

   <li><p>Set <var>query</var>'s <a for=URLSearchParams>list</a> to the result of
   <a lt="urlencoded string parser">parsing</a> <var>init</var>.
  </ol>
</ol>
</div>

<div algorithm>
<p>To <dfn for=URLSearchParams id=concept-urlsearchparams-update>update</dfn> a {{URLSearchParams}}
object <var>query</var>:

<ol>
 <li><p>If <var>query</var>'s <a for=URLSearchParams>URL object</a> is null, then return.

 <li><p>Let <var>serializedQuery</var> be the <a lt="urlencoded serializer">serialization</a> of
 <var>query</var>'s <a for=URLSearchParams>list</a>.

 <li><p>If <var>serializedQuery</var> is the empty string, then set <var>serializedQuery</var> to
 null.

 <li><p>Set <var>query</var>'s <a for=URLSearchParams>URL object</a>'s <a for=URL>URL</a>'s
 <a for=url>query</a> to <var>serializedQuery</var>.
</ol>
</div>

<div algorithm>
<p>The
<dfn constructor for=URLSearchParams lt="URLSearchParams(init)"><code>new URLSearchParams(<var>init</var>)</code></dfn>
constructor steps are:</p>

<ol>
 <li><p>If <var>init</var> is a string and starts with U+003F (?), then remove the first code point
 from <var>init</var>.

 <li><p><a for=URLSearchParams>Initialize</a> <a>this</a> with <var>init</var>.
</ol>
</div>

<div algorithm>
<p>The <dfn attribute for=URLSearchParams><code>size</code></dfn> getter steps are to return
<a>this</a>'s <a for=URLSearchParams>list</a>'s <a for=list>size</a>.
</div>

<div algorithm>
<p>The <dfn method for=URLSearchParams><code>append(<var>name</var>, <var>value</var>)</code></dfn>
method steps are:

<ol>
 <li><p><a for=list>Append</a> (<var>name</var>, <var>value</var>) to <a>this</a>'s
 <a for=URLSearchParams>list</a>.

 <li><p><a for=URLSearchParams>Update</a> <a>this</a>.
</ol>
</div>

<div algorithm>
<p>The <dfn method for=URLSearchParams><code>delete(<var>name</var>, <var>value</var>)</code></dfn>
method steps are:

<ol>
 <li><p>If <var>value</var> is given, then <a for=list>remove</a> all <a for=/>tuples</a> whose name
 is <var>name</var> and value is <var>value</var> from <a>this</a>'s
 <a for=URLSearchParams>list</a>.

 <li><p>Otherwise, <a for=list>remove</a> all <a for=/>tuples</a> whose name is <var>name</var> from
 <a>this</a>'s <a for=URLSearchParams>list</a>.

 <li><p><a for=URLSearchParams>Update</a> <a>this</a>.
</ol>
</div>

<div algorithm>
<p>The <dfn method for=URLSearchParams><code>get(<var>name</var>)</code></dfn> method steps are to
return the value of the first <a for=/>tuple</a> whose name is <var>name</var> in <a>this</a>'s
<a for=URLSearchParams>list</a>, if there is such a <a for=/>tuple</a>; otherwise null.
</div>

<div algorithm>
<p>The <dfn method for=URLSearchParams><code>getAll(<var>name</var>)</code></dfn> method steps are
to return the values of all <a for=/>tuples</a> whose name is <var>name</var> in <a>this</a>'s
<a for=URLSearchParams>list</a>, in list order; otherwise the empty sequence.
</div>

<div algorithm>
<p>The <dfn method for=URLSearchParams><code>has(<var>name</var>, <var>value</var>)</code></dfn>
method steps are:

<ol>
 <li><p>If <var>value</var> is given and there is a <a for=/>tuple</a> whose name is <var>name</var>
 and value is <var>value</var> in <a>this</a>'s <a for=URLSearchParams>list</a>, then return true.

 <li><p>If <var>value</var> is not given and there is a <a for=/>tuple</a> whose name is
 <var>name</var> in <a>this</a>'s <a for=URLSearchParams>list</a>, then return true.

 <li><p>Return false.
</ol>
</div>

<div algorithm>
<p>The <dfn method for=URLSearchParams><code>set(<var>name</var>, <var>value</var>)</code></dfn>
method steps are:

<ol>
 <li><p>If <a>this</a>'s <a for=URLSearchParams>list</a> <a for=list>contains</a> any
 <a for=/>tuples</a> whose name is <var>name</var>, then set the value of the first such
 <a for=/>tuple</a> to <var>value</var> and <a for=list>remove</a> the others.

 <li><p>Otherwise, <a for=list>append</a> (<var>name</var>, <var>value</var>) to <a>this</a>'s
 <a for=URLSearchParams>list</a>.

 <li><p><a for=URLSearchParams>Update</a> <a>this</a>.
</ol>
</div>

<hr>

<div class=example id=example-searchparams-sort>
 <p>It can be useful to sort the name-value tuples in a {{URLSearchParams}} object, in particular to
 increase cache hits. This can be accomplished through invoking the
 {{URLSearchParams/sort()}} method:

 <pre><code class=lang-javascript>
const url = new URL("https://example.org/?q=🏳️‍🌈&amp;key=e1f7bc78");
url.searchParams.sort();
url.search; // "?key=e1f7bc78&amp;q=%F0%9F%8F%B3%EF%B8%8F%E2%80%8D%F0%9F%8C%88"</code></pre>

 <p>To avoid altering the original input, e.g., for comparison purposes, construct a new
 {{URLSearchParams}} object:

 <pre><code class=lang-javascript>
const sorted = new URLSearchParams(url.search)
sorted.sort()</code></pre>
</div>

<div algorithm>
<p>The <dfn method for=URLSearchParams><code>sort()</code></dfn> method steps are:

<ol>
 <li><p>Set <a>this</a>'s <a for=URLSearchParams>list</a> to the result of
 <a for=list>sorting in ascending order</a> <a>this</a>'s <a for=URLSearchParams>list</a>,
 with |a| being less than |b| if |a|'s name is <a>code unit less than</a> |b|'s name.

 <li><p><a for=URLSearchParams>Update</a> <a>this</a>.
</ol>
</div>

<hr>

<p>The <a>value pairs to iterate over</a> are <a>this</a>'s <a for=URLSearchParams>list</a>'s
<a for=/>tuples</a> with the key being the name and the value being the value.

<p>The <dfn for=URLSearchParams>stringification behavior</dfn> steps are to return the
<a lt="urlencoded serializer">serialization</a> of <a>this</a>'s <a for=URLSearchParams>list</a>.


<h3 id=url-apis-elsewhere>URL APIs elsewhere</h3>

<p>A standard that exposes <a for=/>URLs</a>, should expose the <a for=/>URL</a> as a string (by
<a lt="URL serializer">serializing</a> an internal <a for=/>URL</a>). A standard should not expose a
<a for=/>URL</a> using a {{URL}} object. {{URL}} objects are meant for <a for=/>URL</a>
manipulation. In IDL the USVString type should be used.

<p class=note>The higher-level notion here is that values are to be exposed as immutable data
structures.

<p>If a standard decides to use a variant of the name "URL" for a feature it defines, it should name
such a feature "url" (i.e., lowercase and with an "l" at the end). Names such as "URL", "URI", and
"IRI" should not be used. However, if the name is a compound, "URL" (i.e., uppercase) is preferred,
e.g., "newURL" and "oldURL".

<p class=note>The {{EventSource}} and {{HashChangeEvent}} interfaces in <cite>HTML</cite> are
examples of proper naming. [[HTML]]



<h2 id=acknowledgments class=no-num>Acknowledgments</h2>

<p>There have been a lot of people that have helped make <a for=/>URLs</a> more interoperable over
the years and thereby furthered the goals of this standard. Likewise many people have helped making
this standard what it is today.

<p>With that, many thanks to
100の人,<!-- https://twitter.com/esperecyan -->
Adam Barth,
Addison Phillips,
Adrián Chaves,<!-- Gallaecio; GitHub -->
Adrien Ricciardi,
Albert Wiersch,
Alex Christensen,
Alexis Hunt,<!-- alercah; GitHub -->
Alexandre Morgaut,
Alexis Hunt,
Alwin Blok,
Andrew Sullivan,
Arkadiusz Michalski,
Behnam Esfahbod,
Bobby Holley,
Boris Zbarsky,
Brad Hill,
Brandon Ross,
Cailyn Hansen,
Chris Dumez,
Chris Rebert,
Corey Farwell,
Dan Appelquist,
Daniel Bratell,
Daniel Stenberg,
David Burns,
David Håsäther,
David Sheets,
David Singer,
David Walp,
Domenic Denicola,
Emily Schechter,
Emily Stark,
Eric Lawrence,
Erik Arvidsson,
Gavin Carothers,
Geoff Richards,
Glenn Maynard,
Gordon P. Hemsley,
hemanth,<!-- GitHub -->
Henri Sivonen,
Ian Hickson,
Ilya Grigorik,
Italo A. Casas,
Jakub Gieryluk,
James Graham,
James Manger,
James Ross,
Jeff Hodges,
Jeffrey Posnick,
Jeffrey Yasskin,
Joe Duarte,
Joshua Bell,
Jxck,
Karl Wagner,
Kemal Zebari,
田村健人 (Kent TAMURA),
Kevin Grandon,
Kornel Lesiński,
Larry Masinter,
Leif Halvard Silli,
Mark Amery,
Mark Davis,
Marcos Cáceres,
Marijn Kruisselbrink,
Martin Dürst,
Mathias Bynens,
Matt Falkenhagen,
Matt Giuca,
Michael Peick,
Michael™ Smith,
Michal Bukovský,
Michel Suignard,
Mikaël Geljić,
Nikita Skovoroda,<!-- ChALkeR; GitHub -->
Noah Levitt,
Peter Occil,
Philip Jägenstedt,
Philippe Ombredanne,
Prayag Verma,
Rimas Misevičius,
Robert Kieffer,
Rodney Rehm,
Roy Fielding,
Ryan Sleevi,
Sam Ruby,
Sam Sneddon,
Santiago M. Mola,
Sebastian Mayr,
Shannon Booth,<!-- shannonbooth; GitHub -->
Simon Pieters,
Simon Sapin,
Steven Vachon,
Stuart Cook,
Sven Uhlig,
Tab Atkins,
吉野剛史 (Takeshi Yoshino),
Tantek Çelik,
Tiancheng "Timothy" Gu,
Tim Berners-Lee,
簡冠庭 (Tim Guan-tin Chien),
Titi_Alone,
Tomek Wytrębowicz,
Trevor Rowbotham,
Tristan Seligmann,
Valentin Gosu,
Vyacheslav Matva,
Wei Wang,
Wolf Lammen,
山岸和利 (Yamagishi Kazutoshi),
Yongsheng Zhang,
成瀬ゆい (Yui Naruse), and
zealousidealroll
for being awesome!

<p>This standard is written by <a lang=nl href=https://annevankesteren.nl/>Anne van Kesteren</a>
(<a href=https://www.apple.com/>Apple</a>, <a href=mailto:annevk@annevk.nl>annevk@annevk.nl</a>).