Skip to content

Use the URLPattern specification instead of custom URL pattern parsing #1106

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
whimboo wants to merge 5 commits into
base: main
Choose a base branch
from
Open

Use the URLPattern specification instead of custom URL pattern parsing #1106

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
dc72064
Use the URLPattern specification instead of custom URL pattern parsing
whimboo Mar 17, 2026
ba17e47
Fix the exception handling and run it for both URLPattern branches se…
whimboo Jun 16, 2026
4d30e8d
Add note that UrlPatternPattern and UrlPatternString now support patt…
whimboo Jun 16, 2026
3485767
Add issue for follow-up work around additional fields and the ignoreC…
whimboo Jun 16, 2026
ebf35c8
Fix invocation of create algorithm and removed example for escaping.
whimboo Jun 16, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
319 changes: 39 additions & 280 deletions index.bs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7841,7 +7841,7 @@ To <dfn>get the network intercepts</dfn> given |session|, |event|, |request|, an

1. For each |url pattern| in |url patterns|:

1. If [=match URL pattern=] with |url pattern| and |url|:
1. If [=URL pattern/match|URL pattern match=] with |url pattern| and |url|:

1. [=list/Append=] |intercept id| to |intercepts|.

Expand Down Expand Up @@ -8859,312 +8859,71 @@ network.UrlPattern = (

network.UrlPatternPattern = {
type: "pattern",
?protocol: text,
?hostname: text,
?port: text,
?pathname: text,
?search: text,
? protocol: text,
? hostname: text,
? port: text,
? pathname: text,
? search: text,
}


network.UrlPatternString = {
type: "string",
pattern: text,
}

</pre>

A <code>network.UrlPattern</code> represents a pattern used for matching request
URLs for [=network intercepts=].

When URLs are matched against a <code>network.UrlPattern</code> the URL is
parsed, and each component is compared for equality with the corresponding field
in the pattern, if present. Missing fields from the pattern always match.

Note: This syntax is designed with future extensibility in mind. In particular
the syntax forbids characters that are treated specially in the [[URLPattern]]
specification. These can be escaped by prefixing them with a U+005C (\) character.

<div algorithm>
To <dfn>unescape URL pattern</dfn> given |pattern|

1. Let |forbidden characters| be the [=/set=] of codepoints «U+0028 ((), U+0029
()), U+002A (*), U+007B ({), U+007D (})»

1. Let |result| be the empty string.

1. Let |is escaped character| be false.

1. For each |codepoint| in |pattern|:

1. If |is escaped character| is false:

1. If |forbidden characters| [=set/contains=] |codepoint|, return [=error=]
with [=error code=] [=invalid argument=].

1. If |codepoint| is U+005C (\):

1. Set |is escaped character| to true.

1. Continue.

1. Append |codepoint| to result.
Issue(1109): Add support for additional {{URLPatternInit}} fields (e.g.

@jgraham jgraham Jun 19, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Although this is "cheap" in that we're just passing in extra options, I'm not quite convinced that they all make sense for WebDriver use cases.

<code>baseURL</code>, <code>username</code>, <code>password</code>,
<code>hash</code>) and parse options (e.g. <code>ignoreCase</code>).

1. Set |is escaped character| to false.

1. Return [=success=] with data |result|.
A <code>network.UrlPattern</code> represents a pattern used for matching request
URLs for [=network intercepts=]. URL patterns use the syntax and semantics
defined by [[!URLPattern]].

</div>
Note: The component fields of <code>network.UrlPatternPattern</code> and the
<code>pattern</code> field of <code>network.UrlPatternString</code> are parsed
using [[!URLPattern]] syntax. Characters such as <code>*</code>,
<code>+</code>, <code>?</code>, <code>:</code>, and
<code>{</code>...<code>}</code> have special meaning as wildcards, modifiers,
named groups, and grouping constructs respectively. To match these characters
literally, prefix them with a backslash.

<div algorithm>
To <dfn>parse URL pattern</dfn>, given |pattern|:

1. Let |has protocol| be true.

1. Let |has hostname| be true.

1. Let |has port| be true.

1. Let |has pathname| be true.

1. Let |has search| be true.

1. If |pattern| matches the <code>network.UrlPatternPattern</code> production:

1. Let |pattern url| be the empty string.

1. If |pattern| [=map/contains=] "<code>protocol</code>":

1. If |pattern|["<code>protocol</code>"] is the empty string, return [=error=]
with [=error code=] [=invalid argument=].

1. Let |protocol| be the result of [=trying=] to [=unescape
URL Pattern=] with |pattern|["<code>protocol</code>"].

1. For each |codepoint| in |protocol|:

1. If |codepoint| is not [=ASCII alphanumeric=] and «U+002B (+), U+002D
(-), U+002E (.)» does not [=set/contains|contain=] |codepoint|:

1. Return [=error=] with [=error code=] invalid argument.

1. Append |protocol| to |pattern url|.

1. Otherwise:

1. Set |has protocol| to false.

1. Append "<code>http</code>" to |pattern url|.

1. Let |scheme| be [=ASCII lowercase=] with |pattern url|.

1. Append "<code>:</code>" to |pattern url|.

1. If |scheme| [=is special=], append "<code>//</code>" to |pattern url|.

1. If |pattern| [=map/contains=] "<code>hostname</code>":

1. If |pattern|["<code>hostname</code>"] is the empty string, return [=error=]
with [=error code=] [=invalid argument=].

1. If |scheme| is "<code>file</code>" return [=error=] with [=error code=]
[=invalid argument=].

1. Let |hostname| be the result of [=trying=] to [=unescape
URL Pattern=] with |pattern|["<code>hostname</code>"].

1. Let |inside brackets| be false.

1. For each |codepoint| in |hostname|:

1. If «U+002F (/), U+003F (?), U+0023 (#)» [=set/contains=]
|codepoint|:

1. Return [=error=] with [=error code=] [=invalid argument=].

1. If |inside brackets| is false and |codepoint| is U+003A (:):

1. Return [=error=] with [=error code=] [=invalid argument=].

1. If |codepoint| is U+005B ([), set |inside brackets| to true.

1. If |codepoint| is U+005D (]), set |inside brackets| to false.

1. Append |hostname| to |pattern url|.

1. Otherwise:

1. If |scheme| is not "<code>file</code>", append "<code>placeholder</code>" to
|pattern url|.

1. Set |has hostname| to false.

1. If |pattern| [=map/contains=] "<code>port</code>":

1. If |pattern|["<code>port</code>"] is the empty string, return [=error=]
with [=error code=] [=invalid argument=].

1. Let |port| be the result of [=trying=] to [=unescape
URL Pattern=] with |pattern|["<code>port</code>"].

1. Append "<code>:</code>" to |pattern url|.

1. For each |codepoint| in |port|:

1. If |codepoint| is not an [=ASCII digit=]:
1. Let |init| be a new {{URLPatternInit}}.

@jgraham jgraham Jun 19, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Technically you can't really do this; you're trying to create a JS object outside a realm. The URLPattern specification accepts an infra Map object as URLPatternInit in cases where it's not running in a JS realm. I think you want to use the entry point at https://urlpattern.spec.whatwg.org/#build-a-url-pattern-from-an-infra-value.


1. Return [=error=] with [=error code=] [=invalid argument=].
1. If |pattern| [=map/contains=] "<code>protocol</code>", set
|init|["<code>protocol</code>"] to |pattern|["<code>protocol</code>"].

1. Append |port| to |pattern url|.
1. If |pattern| [=map/contains=] "<code>hostname</code>", set
|init|["<code>hostname</code>"] to |pattern|["<code>hostname</code>"].

1. Otherwise:

1. Set |has port| to false.

1. If |pattern| [=map/contains=] "<code>pathname</code>":

1. Let |pathname| be the result of [=trying=] to [=unescape URL Pattern=] with
|pattern|["<code>pathname</code>"].

1. If |pathname| does not [=string/starts with|start with=] U+002F (/), then
append "<code>/</code>" to |pattern url|.

1. For each |codepoint| in |pathname|:

1. If «U+003F (?), U+0023 (#)» [=set/contains=]
|codepoint|:

1. Return [=error=] with [=error code=] [=invalid argument=].

1. Append |pathname| to |pattern url|.

1. Otherwise:

1. Set |has pathname| to false.

1. If |pattern| [=map/contains=] "<code>search</code>":

1. Let |search| be the result of [=trying=] to [=unescape URL pattern=] with
|pattern|["<code>search</code>"].

1. If |search| does not [=string/starts with|start with=] U+003F (?), then
append "<code>?</code>" to |pattern url|.

1. For each |codepoint| in |search|:
1. If |pattern| [=map/contains=] "<code>port</code>", set
|init|["<code>port</code>"] to |pattern|["<code>port</code>"].

1. If |codepoint| is U+0023 (#):
1. If |pattern| [=map/contains=] "<code>pathname</code>", set
|init|["<code>pathname</code>"] to |pattern|["<code>pathname</code>"].

1. Return [=error=] with [=error code=] [=invalid argument=].
1. If |pattern| [=map/contains=] "<code>search</code>", set
|init|["<code>search</code>"] to |pattern|["<code>search</code>"].

1. Append |search| to |pattern url|.

1. Otherwise:

1. Set |has search| to false.
1. Let |url pattern| be the result of
[=URL pattern/create|creating a URLPattern=] given |init|, null, and an
empty [=/map=]. If this throws an exception, return [=error=] with
[=error code=] [=invalid argument=].

1. Otherwise, if |pattern| matches the <code>network.UrlPatternString</code> production:

1. Let |pattern url| be the result of [=trying=] to [=unescape URL
pattern=] with |pattern|["<code>pattern</code>"].

1. Let |url| be the result of [=URL Parser|parsing=] |pattern url|.

1. If |url| is failure, return [=error=] with [=error code=] [=invalid argument=].

1. Let |parsed| be a struct with the following fields:
<dl>
<dt>protocol
<dd>|url|'s [=url/scheme=] if |has protocol| is true, or null otherwise.
<dt>hostname
<dd>|url|'s [=url/host=] if |has hostname| is true, or null otherwise.
<dt>port
<dd>
1. If |has port| is false:

1. null.

1. Otherwise:

1. If |url|'s [=url/scheme=] [=is special=] and |url|'s [=url/scheme=]'s
[=default port=] is not null, and |url|'s [=url/port=] is null or is
equal to [=url/scheme=]'s [=default port=]:

1. The empty string.

1. Otherwise, if |url|'s [=url/port=] is not null:

1. [=Serialize an integer=] with |url|'s [=url/port=].

1. Otherwise:
1. Let |url pattern| be the result of [=URL pattern/create|creating a URLPattern=]
Comment thread
whimboo marked this conversation as resolved.
given |pattern|["<code>pattern</code>"], null, and an empty [=/map=]. If
this throws an exception, return [=error=] with [=error code=]
[=invalid argument=].

1. null.

<dt>pathname
<dd>
1. If |has pathname| is false:

1. null.

1. Otherwise:

1. The result of running the [=URL path serializer=] with |url|, if
|url|'s [=url/path=] is not the empty string and is not [=list/empty=],
or null otherwise.

<dt>search
<dd>
1. If |has search| is false:

1. null.

1. Otherwise:

1. The empty string if |url|'s [=url/query=] is null, or |url|'s [=url/query=] otherwise.
</dl>

1. Return [=success=] with data |parsed|.

</div>

<div algorithm>
To <dfn>match URL pattern</dfn> given |url pattern| and |url string|:

1. Let |url| be the result of [=URL parser|parsing=] |url string|.

1. If |url pattern|'s protocol is not null and is not equal to |url|'s
[=url/scheme=], return false.

1. If |url pattern|'s hostname is not null and is not equal to |url|'s
[=url/host=], return false.

1. If |url pattern|'s port is not null:

1. Let |port| be null.

1. If |url|'s [=url/scheme=] [=is special=] and |url|'s
[=url/scheme=]'s [=default port=] is not null, and |url|'s [=url/port=] is
null or is equal to [=url/scheme=]'s [=default port=]:

1. Set |port| to the empty string.

1. Otherwise, if |url|'s [=url/port=], is not null:

1. Set |port| to [=serialize an integer=] with |url|'s [=url/port=].

1. If |url pattern|'s port is not equal to |port|, return false.

1. If |url pattern|'s pathname is not null and is not equal to the result of
running the [=URL path serializer=] with |url|, return false.

1. If |url pattern|'s search is not null:

1. Let |url query| be |url|'s [=url/query=].

1. If |url query| is null, set |url query| to the empty string.

1. If |url pattern|'s search is not equal to |url query|, return false.

1. Return true.
1. Return [=success=] with data |url pattern|.

</div>

Expand Down
Loading