Use empty host for file URL's - #260

rmisev · 2017-02-24T13:09:27Z

Summary of changes:

Fixes Empty or null host when scheme is "file:"? #258
Introduces empty host term for file and non-special URL's
Changes definition: opaque host is non-empty ASCII string

Preview | Diff

Fixes whatwg#258.

base URL's host also must be checked it is empty/not empty string

annevk

Looks good overall, but many nits.

annevk · 2017-02-24T13:15:40Z

+ <dd><p>a <a>path-absolute-URL string</a> if <a>base URL</a>'s <a for=url>host</a> is either null or the empty string
 <dd><p>a <a>path-absolute-non-Windows-file-URL string</a> if <a>base URL</a>'s <a for=url>host</a>
- is non-null
+ is non-empty string


It's not always a string though. It could be an IPv4 address.

This should probably just say not an empty host.

annevk · 2017-02-24T13:16:04Z


 <p>A <dfn export id=concept-host>host</dfn> is a <a>domain</a>, an
-<a>IPv4 address</a>, an <a>IPv6 address</a>, or an <a>opaque host</a>. Typically a <a for=/>host</a>
+<a>IPv4 address</a>, an <a>IPv6 address</a>, an <a>opaque host</a>, or an <a>empty host</a>. Typically a <a for=/>host</a>


This doesn't meet the 100 column wrapping requirements.

annevk · 2017-02-24T13:16:16Z

 <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 an <a>ASCII string</a> holding data that can be used for
+<p>An <dfn export>opaque host</dfn> is a non-empty <a>ASCII string</a> holding data that can be used for


Same here, I think.

annevk · 2017-02-24T13:16:57Z

 followed by "<code>]</code>".

- <li><p>Otherwise, <var>host</var> is a <a>domain</a> or <a>opaque host</a>, return <var>host</var>.
+ <li><p>Otherwise, <var>host</var> is a <a>domain</a>, an <a>opaque host</a>, or an <a>empty host</a>, return <var>host</var>.


annevk · 2017-02-24T13:18:27Z

 <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 null
+ <dd><p>a <a>path-absolute-URL string</a> if <a>base URL</a>'s <a for=url>host</a> is either null or the empty string


This should say "an empty host" I think. A base URL is a URL record after all. But also, I thought the idea was that for file URLs the host could never be null so then we should leave that out too.

annevk · 2017-02-24T13:19:26Z


 <p>An <dfn export>opaque-host-and-port string</dfn> must be either an empty
-<a>valid opaque-host string</a> or: a non-empty <a>valid opaque-host string</a>, optionally followed
+string or: a <a>valid opaque-host string</a>, optionally followed


It's "the empty string" (and an empty host).

annevk · 2017-02-24T13:20:00Z


          <ol>
-           <li><p>If <var>url</var>'s <a for=url>host</a> is non-null,
+           <li><p>If <var>url</var>'s <a for=url>host</a> is not the empty string,


Here again we're talking about a URL record so we should say "not an empty host". Same below.

Both an empty host and an opaque host are strings by definition. So maybe it's fine to set host to the empty string in the parser? For example the opaque host parser returns string, but not an opaque host record.

Fair, I guess some level of implicit casting is probably okay.

annevk

The other thing I'm wondering about is if we should add some kind of explanatory text. That file URLs can have empty or non-opaque hosts, non-file-special URLs can have non-empty-non-opaque hosts, and non-special URLs can have empty, opaque, or null hosts. Such a note should probably go where we define the host field of URLs.

annevk · 2017-02-27T12:47:37Z

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

-<p>An <dfn export>valid opaque-host string</dfn> must be zero or more <a>URL units</a> or:
+<p>An <dfn export>valid opaque-host string</dfn> must be one or more <a>URL units</a> or:


Can you change "An" to "A" here while we're changing this? I can also do it before merging.

annevk · 2017-02-27T12:49:09Z


- <li><p>Otherwise, <var>host</var> is a <a>domain</a> or <a>opaque host</a>, return <var>host</var>.
+ <li><p>Otherwise, <var>host</var> is a <a>domain</a>, an <a>opaque host</a>, or an <a>empty
+ host</a>, return <var>host</var>.


I think it reads more naturally if you leave out the "an". So "A domain, opaque host, or empty host".

annevk · 2017-02-27T12:50:00Z

 <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 null
+ <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>


No newlines inside phrasing/inline elements.

annevk · 2017-02-27T12:51:10Z

           <a>validation error</a>.

-           <li><p>Set <var>url</var>'s <a for=url>host</a> to null and replace the second
+           <li><p>Set <var>url</var>'s <a for=url>host</a> to the empty string and replace the second


This is >100 columns. "second" needs to go to the next line.

rmisev · 2017-02-27T15:26:06Z

Maybe explanatory text is good idea, I don't know. But I would like to note, what file URL still can have null host (two invalid URL's cases: file:path and file:/path).

Also I think about such note below the empty host definition: "An empty host is used by non-special and file URLs."

annevk · 2017-02-27T15:51:59Z

I guess there's no real point trying to make those invalid cases consistent. And you're right about where the note should go. Although maybe we should have it in both places eventually, once we're certain this design can be widely implemented (i.e., once two browsers ship it).

rmisev · 2017-03-02T15:57:41Z

I think the such table is clearer than a long explanatory text #260 (review), #260 (comment):

Scheme	domain	IPv4 address	IPv6 address	opaque host	empty host	null
non-file special	+	+	+
"file"	+	+	+		+	+
non-special			+	+	+	+

Then maybe no need for additional notes. What do you think?

annevk · 2017-03-03T11:55:40Z

And have that below the definition of URL's host? I think that's reasonable.

annevk

Couple nits left. I'll also push a commit to adjust the formatting of the table a bit.

annevk · 2017-03-06T10:22:51Z

+for further processing.

 <p class="note no-backref">An <a>opaque host</a> is only used by <a lt="is special">non-special</a>
 <a for=/>URLs</a>.


I think either we should have a note like this for empty hosts too or we should drop this one, no?

annevk · 2017-03-06T10:23:34Z

+      <th><a>opaque host</a>
+      <th><a>empty host</a>
+      <th>null
+  <tr><td>non-file <a lt="special scheme">special</a><td>✅<td>✅<td>✅<td><td><td>


non-"<code>file</code>"

annevk · 2017-03-06T10:29:28Z

I ended up fixing the nits, let me know what you think.

if host is null, then no validation error and don't set host to ""

rmisev · 2017-03-06T14:45:55Z

I think your changes are OK.

Yet I have changed the "Windows drive letter quirk" in the path state, to be more logical (no validation error if host null) and consistent (null host is left the null):

For example, now: file:/C:/p --> host === null as for file:/a/p also host === null, but file://example.net/C:/ --> host === ""

Now LGTM.

annevk · 2017-03-06T14:47:31Z

Looks okay, blocked on speced/bikeshed#941 now.

domenic · 2017-03-07T17:42:27Z

To confirm, this only impacts validation, not parsing/serializing?

annevk · 2017-03-07T17:45:58Z

Yeah, this should not have changed either of those (other than logging validation errors).

rmisev · 2017-03-07T18:35:33Z

Yes, fixes validation bug. Parsing: file://host/C:/ and file:///C:/ now parses to empty host, previously to null. Serializing for file URL is the same host is null or empty - so no impact.

Follows whatwg/url#260. This does not impact parsing, only validation and the URL record representation.

domenic · 2017-03-14T17:44:50Z

Note that there is no test coverage for the case introduced here: in

        if (this.url.host !== "" && this.url.host !== null) {
          this.parseError = true;
          this.url.host = "";
        }

the condition never evaluates to true so those two lines are never reached. Maybe a coverage gap, maybe a spec bug.

rmisev · 2017-03-14T18:47:22Z

@domenic test case, which covers this code is in previous comment: #260 (comment):
file://host/C:/

It looks like WPT lacks this. I can prepare PR.

Tests windows drive letter quirk in the path state: https://url.spec.whatwg.org/#path-state Fixes whatwg/url#272 (whatwg/url#260 (comment))

- Update to spec - Add opaque hosts - File state did not correctly deal with lack of base URL - Cleanup API for file and non-special URLs - Allow % and IPv6 addresses in non-special URL hosts - Use specific names for percent-encode sets - Add empty host concept for file and non-special URLs - Clarify IPv6 serializer - Fix existing mistakes - Add missing ':' to forbidden host code point list. - Correct IPv4 parser empty label behavior - Maintain type equivalence in URLContext with spec - scheme, username, and password should always be strings - host, port, query, and fragment may be strings or null - Align scheme state more closely with the spec - Make sure the `special` variable is always synced with URL_FLAG_SPECIAL. PR-URL: nodejs#12523 Fixes: nodejs#10608 Fixes: nodejs#10634 Refs: whatwg/url#185 Refs: whatwg/url#225 Refs: whatwg/url#224 Refs: whatwg/url#218 Refs: whatwg/url#243 Refs: whatwg/url#260 Refs: whatwg/url#268 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>

- Update to spec - Add opaque hosts - File state did not correctly deal with lack of base URL - Cleanup API for file and non-special URLs - Allow % and IPv6 addresses in non-special URL hosts - Use specific names for percent-encode sets - Add empty host concept for file and non-special URLs - Clarify IPv6 serializer - Fix existing mistakes - Add missing ':' to forbidden host code point list. - Correct IPv4 parser empty label behavior - Maintain type equivalence in URLContext with spec - scheme, username, and password should always be strings - host, port, query, and fragment may be strings or null - Align scheme state more closely with the spec - Make sure the `special` variable is always synced with URL_FLAG_SPECIAL. PR-URL: #12523 Fixes: #10608 Fixes: #10634 Refs: whatwg/url#185 Refs: whatwg/url#225 Refs: whatwg/url#224 Refs: whatwg/url#218 Refs: whatwg/url#243 Refs: whatwg/url#260 Refs: whatwg/url#268 Reviewed-By: James M Snell <jasnell@gmail.com> Reviewed-By: Daijiro Wachi <daijiro.wachi@gmail.com> Reviewed-By: Joyee Cheung <joyeec9h3@gmail.com>

Fix of: whatwg/url#258 Follows: whatwg/url#260

rmisev added 4 commits February 24, 2017 10:57

uses empty host instead of null in path state

ff5ab18

Fixes whatwg#258.

fixes "relative-URL string" definition for "file"

ac38a3a

base URL's host also must be checked it is empty/not empty string

introduces "empty host" term

98ec637

opaque-host is a non-empty string

f91dc1b

annevk requested changes Feb 24, 2017

View reviewed changes

formatting nits and more "empty host" use

03d2424

annevk reviewed Feb 27, 2017

View reviewed changes

fixes some typos and wrapping

4335e76

Add a table with allowed URL’s scheme / host combinations

eb71fd1

annevk reviewed Mar 6, 2017

View reviewed changes

remove note, adjust table

19f5a62

Fix "Windows drive letter quirk" in the path state

3893222

if host is null, then no validation error and don't set host to ""

annevk approved these changes Mar 6, 2017

View reviewed changes

annevk merged commit 5807b28 into whatwg:master Mar 7, 2017

domenic added a commit to jsdom/whatwg-url that referenced this pull request Mar 14, 2017

Spec update: add the empty host concept

2403948

Follows whatwg/url#260. This does not impact parsing, only validation and the URL record representation.

annevk mentioned this pull request Mar 14, 2017

Empty host file URL fallout? #272

Closed

rmisev mentioned this pull request Mar 14, 2017

File URL's parsing tests for path state web-platform-tests/wpt#5146

Merged

domenic pushed a commit to web-platform-tests/wpt that referenced this pull request Mar 14, 2017

File URL's parsing tests for path state

634175d

Tests windows drive letter quirk in the path state: https://url.spec.whatwg.org/#path-state Fixes whatwg/url#272 (whatwg/url#260 (comment))

TimothyGu mentioned this pull request Apr 23, 2017

url: update WHATWG URL parser to latest spec nodejs/node#12523

Closed

4 tasks

rmisev added a commit to upa-url/upa that referenced this pull request May 24, 2020

path state: url_parser::parse_path(..): null host --> empty host

62f985c

Fix of: whatwg/url#258 Follows: whatwg/url#260

coderabbitai Bot mentioned this pull request Aug 21, 2026

Make pipeline e2e tests handle the MLflow integration section opendatahub-io/odh-dashboard#9443

Merged

5 tasks

Uh oh!

Conversation

rmisev commented Feb 24, 2017 • edited by pr-preview Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

annevk left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

annevk left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rmisev commented Feb 27, 2017

Uh oh!

annevk commented Feb 27, 2017

Uh oh!

rmisev commented Mar 2, 2017

Uh oh!

annevk commented Mar 3, 2017

Uh oh!

annevk left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

annevk commented Mar 6, 2017

Uh oh!

rmisev commented Mar 6, 2017

Uh oh!

annevk commented Mar 6, 2017

Uh oh!

domenic commented Mar 7, 2017

Uh oh!

annevk commented Mar 7, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

rmisev commented Mar 7, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

domenic commented Mar 14, 2017

Uh oh!

rmisev commented Mar 14, 2017

Uh oh!

Reviewers

Assignees

Labels

Milestone

Development

Uh oh!

3 participants

rmisev commented Feb 24, 2017 •

edited by pr-preview Bot

Loading

annevk commented Mar 7, 2017 •

edited

Loading

rmisev commented Mar 7, 2017 •

edited

Loading