Preserve focus when entering the back-forward cache [chromium/src : main]

[Helmut Januschka (Gerrit)]

Attention needed from Joey Arhar

Helmut Januschka added 1 comment

Patchset-level comments

File-level comment, Patchset 2 (Latest):

Helmut Januschka . resolved

@jar...@chromium.org - PTAL, do you have an idea who would be a good reviewer?

thanks in advance, and no time pressure from my end, whenver you have a couple of minutes. thanks!

Open in Gerrit

Related details

Attention is currently required from:

• Joey Arhar

Submit Requirements:

• Code-Coverage
• Code-Owners
• Code-Review
• Review-Enforcement

Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings.

Gerrit

Gerrit-MessageType: comment

Gerrit-Project: chromium/src

Gerrit-Branch: main

Gerrit-Change-Id: Id42617b17da46656f35c855eea02a8df848543e0

Gerrit-Change-Number: 7829940

Gerrit-PatchSet: 2

Gerrit-Owner: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Joey Arhar <jar...@chromium.org>

Gerrit-CC: Chromium LUCI CQ <chromiu...@luci-project-accounts.iam.gserviceaccount.com>

Gerrit-CC: android-bu...@system.gserviceaccount.com <android-bu...@system.gserviceaccount.com>

Gerrit-CC: chromium...@chromium.org

Gerrit-Attention: Joey Arhar <jar...@chromium.org>

Gerrit-Comment-Date: Sun, 24 May 2026 23:14:14 +0000

Gerrit-HasComments: Yes

Gerrit-Has-Labels: No

[Fergal Daly (Gerrit)]

Attention needed from Fergal Daly, Helmut Januschka and Joey Arhar

Fergal Daly added 1 comment

File content/browser/back_forward_cache_browsertest.cc

Line 1879, Patchset 2 (Parent): // 5) Navigating back to |url_1|, we shouldn't restore the focus to the

Fergal Daly . unresolved

Do you know why we were testing for the opposite behaviour? Does the HTML spec have an opinion on this?

Open in Gerrit

Related details

Attention is currently required from:

• Fergal Daly
• Helmut Januschka
• Joey Arhar

Submit Requirements:

• Code-Coverage
• Code-Owners
• Code-Review

• No-Unresolved-Comments

• Review-Enforcement

Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings.

Gerrit

Gerrit-MessageType: comment

Gerrit-Project: chromium/src

Gerrit-Branch: main

Gerrit-Change-Id: Id42617b17da46656f35c855eea02a8df848543e0

Gerrit-Change-Number: 7829940

Gerrit-PatchSet: 2

Gerrit-Owner: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Fergal Daly <fer...@google.com>

Gerrit-Reviewer: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Joey Arhar <jar...@chromium.org>

Gerrit-CC: Chromium LUCI CQ <chromiu...@luci-project-accounts.iam.gserviceaccount.com>

Gerrit-CC: Fergal Daly <fer...@chromium.org>

Gerrit-CC: android-bu...@system.gserviceaccount.com <android-bu...@system.gserviceaccount.com>

Gerrit-CC: chromium...@chromium.org

Gerrit-Attention: Fergal Daly <fer...@google.com>

Gerrit-Attention: Helmut Januschka <hel...@januschka.com>

Gerrit-Attention: Joey Arhar <jar...@chromium.org>

Gerrit-Comment-Date: Sun, 24 May 2026 23:55:14 +0000

Gerrit-HasComments: Yes

Gerrit-Has-Labels: No

[Fergal Daly (Gerrit)]

Attention needed from Fergal Daly, Helmut Januschka and Joey Arhar

Fergal Daly added 1 comment

File third_party/blink/web_tests/http/tests/back-forward-cache/focus-preserved.html

File-level comment, Patchset 2 (Latest):

Fergal Daly . unresolved

https://wpt.fyi/results/html/browsers/browsing-the-web/back-forward-cache/focus.html?label=master&label=experimental&aligned&q=html%2Fbrowsers%2Fbrowsing-the-web%2Fback-forward-cache

already seems to test the same thing (at least by its description). It's also failing on Firefox, so I wonder if there is consensus on the behaviour.

Gerrit-Comment-Date: Mon, 25 May 2026 00:04:41 +0000

Gerrit-HasComments: Yes

Gerrit-Has-Labels: No

[Helmut Januschka (Gerrit)]

Attention needed from Fergal Daly, Fergal Daly and Joey Arhar

Helmut Januschka added 2 comments

File content/browser/back_forward_cache_browsertest.cc

Line 1879, Patchset 2 (Parent): // 5) Navigating back to |url_1|, we shouldn't restore the focus to the

Fergal Daly . resolved

Do you know why we were testing for the opposite behaviour? Does the HTML spec have an opinion on this?

Helmut Januschka

The opposite expectation (`blurCount == 1`) was added in 2020 by https://crrev.com/c/2362046, intentionally clearing focus + text-input state on BFCache entry so the IME / text-input plumbing matched a normal cross-document navigation. The test was just locking in that implementation choice; not a spec requirement.

File third_party/blink/web_tests/http/tests/back-forward-cache/focus-preserved.html

File-level comment, Patchset 2:

Fergal Daly . unresolved

https://wpt.fyi/results/html/browsers/browsing-the-web/back-forward-cache/focus.html?label=master&label=experimental&aligned&q=html%2Fbrowsers%2Fbrowsing-the-web%2Fback-forward-cache

already seems to test the same thing (at least by its description). It's also failing on Firefox, so I wonder if there is consensus on the behaviour.

Helmut Januschka

Acknowledged, dropped `third_party/blink/web_tests/http/tests/back-forward-cache/focus-preserved.html`.

On consensus: Safari already preserves focus across BFCache, so this CL aligns Chromium with WebKit. Firefox is the other outlier then. The HTML spec is silent about it (`reactivate a Document` says nothing about focus).

I'm happy to file a spec issue to make this explicit. Want me to file it as a follow-up to this CL, or block on it?

Open in Gerrit

Related details

Attention is currently required from:

• Fergal Daly
• Fergal Daly
• Joey Arhar

Submit Requirements:

• Code-Coverage
• Code-Owners
• Code-Review
• No-Unresolved-Comments
• Review-Enforcement

Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings.

Gerrit

Gerrit-MessageType: comment

Gerrit-Project: chromium/src

Gerrit-Branch: main

Gerrit-Change-Id: Id42617b17da46656f35c855eea02a8df848543e0

Gerrit-Change-Number: 7829940

Gerrit-PatchSet: 3

Gerrit-Owner: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Fergal Daly <fer...@google.com>

Gerrit-Reviewer: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Joey Arhar <jar...@chromium.org>

Gerrit-CC: Chromium LUCI CQ <chromiu...@luci-project-accounts.iam.gserviceaccount.com>

Gerrit-CC: Fergal Daly <fer...@chromium.org>

Gerrit-CC: Zhe Su <su...@chromium.org>

Gerrit-CC: android-bu...@system.gserviceaccount.com <android-bu...@system.gserviceaccount.com>

Gerrit-CC: chromium...@chromium.org

Gerrit-Attention: Fergal Daly <fer...@google.com>

Gerrit-Attention: Joey Arhar <jar...@chromium.org>

Gerrit-Attention: Fergal Daly <fer...@chromium.org>

Gerrit-Comment-Date: Mon, 25 May 2026 20:32:23 +0000

Gerrit-HasComments: Yes

Gerrit-Has-Labels: No

Comment-In-Reply-To: Fergal Daly <fer...@chromium.org>

[Fergal Daly (Gerrit)]

Attention needed from Fergal Daly, Helmut Januschka and Joey Arhar

Fergal Daly added 1 comment

File third_party/blink/web_tests/http/tests/back-forward-cache/focus-preserved.html

File-level comment, Patchset 2:

Fergal Daly . unresolved

https://wpt.fyi/results/html/browsers/browsing-the-web/back-forward-cache/focus.html?label=master&label=experimental&aligned&q=html%2Fbrowsers%2Fbrowsing-the-web%2Fback-forward-cache

already seems to test the same thing (at least by its description). It's also failing on Firefox, so I wonder if there is consensus on the behaviour.

Helmut Januschka

Acknowledged, dropped `third_party/blink/web_tests/http/tests/back-forward-cache/focus-preserved.html`.

On consensus: Safari already preserves focus across BFCache, so this CL aligns Chromium with WebKit. Firefox is the other outlier then. The HTML spec is silent about it (`reactivate a Document` says nothing about focus).

I'm happy to file a spec issue to make this explicit. Want me to file it as a follow-up to this CL, or block on it?

Fergal Daly

Unfortunately, I think we should block. The focus behaviour was added very intentionally.

https://crrev.com/c/2362046

and there's a real question as to what the correct behaviour should be. There's a tension with BFCache between making it seem exactly like a normal navigation just faster and the fact that web pages are stateful and we are restoring that state.

I'm also hesistant to change focus stuff. I know we already have a mess because input events are not on the same mojo pipe as RFH and other navigation events, including focus. I feel like this has been a source of flakiness, although it was quite a while ago.

Open in Gerrit

Related details

Attention is currently required from:

• Fergal Daly
• Helmut Januschka
• Joey Arhar

Submit Requirements:

• Code-Coverage
• Code-Owners
• Code-Review
• No-Unresolved-Comments
• Review-Enforcement

Inspect html for hidden footers to help with email filtering. To unsubscribe visit settings.

Gerrit

Gerrit-MessageType: comment

Gerrit-Project: chromium/src

Gerrit-Branch: main

Gerrit-Change-Id: Id42617b17da46656f35c855eea02a8df848543e0

Gerrit-Change-Number: 7829940

Gerrit-PatchSet: 3

Gerrit-Owner: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Fergal Daly <fer...@google.com>

Gerrit-Reviewer: Helmut Januschka <hel...@januschka.com>

Gerrit-Reviewer: Joey Arhar <jar...@chromium.org>

Gerrit-CC: Chromium LUCI CQ <chromiu...@luci-project-accounts.iam.gserviceaccount.com>

Gerrit-CC: Fergal Daly <fer...@chromium.org>

Gerrit-CC: Zhe Su <su...@chromium.org>

Gerrit-CC: android-bu...@system.gserviceaccount.com <android-bu...@system.gserviceaccount.com>

Gerrit-CC: chromium...@chromium.org

Gerrit-Attention: Helmut Januschka <hel...@januschka.com>

Gerrit-Attention: Fergal Daly <fer...@google.com>

Gerrit-Attention: Joey Arhar <jar...@chromium.org>

Gerrit-Comment-Date: Tue, 26 May 2026 00:59:49 +0000

Gerrit-HasComments: Yes

Gerrit-Has-Labels: No

Comment-In-Reply-To: Helmut Januschka <hel...@januschka.com>

Comment-In-Reply-To: Fergal Daly <fer...@chromium.org>