`emulation.setClientHintsOverride` -> `userAgentClientHints.setClientHintsOverride` by sadym-chromium · Pull Request #393 · WICG/ua-client-hints

sadym-chromium · 2026-01-21T11:03:41Z

Addressing #391 by moving command to a dedicated userAgentClientHints domain.

whimboo

The renaming looks fine to me but I have a couple more questions regarding the overall steps and internal data definitions. If those should be on its own PR feel free to move them out.

whimboo · 2026-01-21T12:20:18Z

For our BiDi emulation commands we have such maps on the BiDi session or remote end. Could we use the remote end definition from BiDi here?

For our BiDi emulation commands we have such maps on the BiDi session or remote end. Could we use the remote end definition from BiDi here?

This spec uses term "user agent" as a synonym to WebDriver's "remote end".

So what should we follow then? Has the term from the actual spec the precedence I assume?

CC @jgraham.

The spec gets to decide on its own internal definitions.

However.

The Infra definition of "user agent" is understood to be closer to WebDriver BiDi's "user context" than "remote end". Therefore, without any further investigation, if you are trying to use it for the latter you might be wrong.

(I think Infra made a mistake here; the argument is basically that web specs can't see outside their own storage partition anyway, so from their point of view defining a "user agent" as a single storage partition makes things easier because it's the biggest domain that a web spec might want to address. But of course that's not what anyone else means by "user agent" and WebDriver can see outside specific storage partitions, so it just means we need to avoid the term)

whimboo · 2026-01-21T12:20:58Z

@@ -1084,28 +1084,28 @@
 * [=struct/item=] named <dfn for="emulated client hints" export>emulated client hints per navigables</dfn>,
    which is a weak map between [=/navigables=] and [=user agent client hints=], initially empty.


Other emulation commands need only a single map. Why do we have to maintain three of them for this command?

Other emulation commands need only a single map. Why do we have to maintain three of them for this command?

This approach was suggested by extra headers and is used in several places in the main spec. I personally find it more readable and structured. Also it allows for more explicit global settings handling:

https://w3c.github.io/webdriver-bidi/#session-emulated-user-agent

https://w3c.github.io/webdriver-bidi/#session-emulated-maxtouchpoints

https://w3c.github.io/webdriver-bidi/#session-emulated-network-conditions

https://w3c.github.io/webdriver-bidi/#download-behavior

Oh I read this wrongly. It's a single struct with individual maps as properties. That's fine then.

whimboo · 2026-01-21T12:24:09Z


 1. If |command parameters| [=map/contains=] "<code>userContexts</code>"
    and |command parameters| [=map/contains=] "<code>contexts</code>",
    return [=error=] with [=error code=] [=invalid argument=].


It would be great if the order of the steps would be equal to those that we have in our BiDi spec, which would make it much easier to parse and implement. Maybe this can be fixed? Also I do not see a step that checks that at least contexts or userContexts are present.

Also I do not see a step that checks that at least contexts or userContexts are present.

It is allowed to skip both contexts or userContexts and it would mean global setting. w3c/webdriver-bidi#956

Ah ok, Not all of our existing emulation commands support that yet and I actually checked only those. :/ Thanks for pointing out.

whimboo

I think that this looks fine but I would like to have @jgraham taking the review as well from our side.

whimboo · 2026-01-21T14:11:08Z

So what should we follow then? Has the term from the actual spec the precedence I assume?

CC @jgraham.

whimboo · 2026-01-21T14:12:17Z

@@ -1084,28 +1084,28 @@ A [=User agent=] has <dfn>emulated client hints</dfn>, which is a [=struct=] wit
 * [=struct/item=] named <dfn for="emulated client hints" export>emulated client hints per navigables</dfn>,
    which is a weak map between [=/navigables=] and [=user agent client hints=], initially empty.


Oh I read this wrongly. It's a single struct with individual maps as properties. That's fine then.

jgraham · 2026-01-27T09:26:28Z

Now that it's using its own namespace I don't have further concerns.

whimboo · 2026-01-27T09:30:02Z

@miketaylr can you please review these changes? Thanks!

…HintsOverride`

sadym-chromium · 2026-02-09T09:54:45Z

@miketaylr PTAL

miketaylr · 2026-02-09T22:20:06Z

Apologies for the delay - have been traveling for summits. @arichiv, do you mind reviewing? It may be a few days before I can take a look.

arichiv

seems reasonable overall, not fully read-in on the motivation but if it makes web-driver side people happy the changes here seem mostly like a no-op (better defined, no changes in Chrome required).

sadym-chromium · 2026-02-10T14:25:20Z

@arichiv please consider merging

…s.setClientHintsOverride SHA: 6cf1025 Reason: push, by arichiv Co-authored-by: github-actions[bot] <41898282+github-actions[bot]@users.noreply.github.com>

…tHints.setClientHintsOverride` Follow-up after WICG/ua-client-hints#393

…tClientHintsOverride` (#4053) Follow-up after WICG/ua-client-hints#393 --------- Signed-off-by: Browser Automation Bot <browser-automation-bot@google.com> Co-authored-by: Browser Automation Bot <browser-automation-bot@google.com>

…tHints.setClientHintsOverride` (#57681) Follow-up after WICG/ua-client-hints#393

…tClientHintsOverride` (#4053) Follow-up after WICG/ua-client-hints#393 --------- Signed-off-by: Browser Automation Bot <browser-automation-bot@google.com> Co-authored-by: Browser Automation Bot <browser-automation-bot@google.com>

…sOverride` -> `userAgentClientHints.setClientHintsOverride`, a=testonly Automatic update from web-platform-tests [wdspec] rename `emulation.setClientHintsOverride` -> `userAgentClientHints.setClientHintsOverride` (#57681) Follow-up after WICG/ua-client-hints#393 -- wpt-commits: 014898d321c48cff1d98f5dfb03d116e7210c13f wpt-pr: 57681

sadym-chromium mentioned this pull request Jan 21, 2026

Automation section extends a module defined in WebDriver-BiDi #391

Closed

whimboo suggested changes Jan 21, 2026

View reviewed changes

sadym-chromium force-pushed the sadym/userAgentClientHints.setClientHintsOverride branch from 912bf7e to ca4a72b Compare January 21, 2026 13:34

sadym-chromium requested a review from whimboo January 21, 2026 13:35

whimboo approved these changes Jan 21, 2026

View reviewed changes

sadym-chromium added 3 commits February 9, 2026 10:51

emulation.setClientHintsOverride -> `userAgentClientHints.setClient…

c956229

…HintsOverride`

Addressing comments from PR

157f78d

use term remot end

e85e188

sadym-chromium force-pushed the sadym/userAgentClientHints.setClientHintsOverride branch from 3e5636e to e85e188 Compare February 9, 2026 09:53

arichiv approved these changes Feb 10, 2026

View reviewed changes

Comment thread index.bs

arichiv merged commit 6cf1025 into WICG:main Feb 10, 2026
2 checks passed

sadym-chromium mentioned this pull request Feb 10, 2026

chore: emulation.setClientHintsOverride -> userAgentClientHints.setClientHintsOverride GoogleChromeLabs/chromium-bidi#4053

Merged

sadym-chromium added a commit to web-platform-tests/wpt that referenced this pull request Feb 10, 2026

[wdspec] rename emulation.setClientHintsOverride -> `userAgentClien…

a5c47a3

…tHints.setClientHintsOverride` Follow-up after WICG/ua-client-hints#393

sadym-chromium mentioned this pull request Feb 10, 2026

[wdspec] rename emulation.setClientHintsOverride -> userAgentClientHints.setClientHintsOverride web-platform-tests/wpt#57681

Merged

foolip pushed a commit to web-platform-tests/wpt that referenced this pull request Feb 10, 2026

[wdspec] rename emulation.setClientHintsOverride -> `userAgentClien…

014898d

…tHints.setClientHintsOverride` (#57681) Follow-up after WICG/ua-client-hints#393

		@@ -1084,28 +1084,28 @@
		* [=struct/item=] named <dfn for="emulated client hints" export>emulated client hints per navigables</dfn>,
		which is a weak map between [=/navigables=] and [=user agent client hints=], initially empty.

		@@ -1084,28 +1084,28 @@ A [=User agent=] has <dfn>emulated client hints</dfn>, which is a [=struct=] wit
		* [=struct/item=] named <dfn for="emulated client hints" export>emulated client hints per navigables</dfn>,
		which is a weak map between [=/navigables=] and [=user agent client hints=], initially empty.

Conversation

sadym-chromium commented Jan 21, 2026 • edited by pr-preview Bot Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

whimboo left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

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!

sadym-chromium Jan 21, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

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!

whimboo 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!

jgraham commented Jan 27, 2026

Uh oh!

whimboo commented Jan 27, 2026

Uh oh!

sadym-chromium commented Feb 9, 2026

Uh oh!

miketaylr commented Feb 9, 2026

Uh oh!

arichiv left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

sadym-chromium commented Feb 10, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

sadym-chromium commented Jan 21, 2026 •

edited by pr-preview Bot

Loading

sadym-chromium Jan 21, 2026 •

edited

Loading