Closed Bug 1917470 Opened 25 days ago Closed 22 days ago

Firefox started using apple-touch-icon for bookmarks?

Categories

(Firefox :: Bookmarks & History, defect, P1)

Firefox 132
defect

Tracking

()

RESOLVED FIXED
132 Branch
Tracking Status
firefox-esr115 --- unaffected
firefox-esr128 --- unaffected
firefox130 --- unaffected
firefox131 --- unaffected
firefox132 --- fixed

People

(Reporter: eight04, Assigned: yazan)

References

(Regression)

Details

(Keywords: regression, Whiteboard: [sng])

Attachments

(2 files)

Attached video out.mp4

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:132.0) Gecko/20100101 Firefox/132.0

Steps to reproduce:

Recently, I noticed that lots of bookmark icons changed after visiting the site. For example:

https://schedule.hololive.tv/
https://holodex.net/
https://www.gamer.com.tw/

Actual results:

See the attachment.

It seems that Firefox started using apple-touch-icon for bookmark icons.

Expected results:

I think the old favicon works better in this case. Also Firefox still uses favicon for tab icons.

Probably relate?
https://bugzilla.mozilla.org/show_bug.cgi?id=1772264

The Bugbug bot thinks this bug should belong to the 'Firefox::Bookmarks & History' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Component: Untriaged → Bookmarks & History

(In reply to eight04 from comment #0)

It seems that Firefox started using apple-touch-icon for bookmark icons.

It always did, though the bug you linked to should have done exactly the opposite (prioritize non rich icons).

Looking at the database the rich flag is correctly set, it's likely we inadvertently inverted the condition.
This is luckily non destructive, a fix should be immediately effective.

Yazan, could you please check this?

Severity: -- → S4
Status: UNCONFIRMED → NEW
Ever confirmed: true
Flags: needinfo?(yalmacki)
Keywords: regression
Priority: -- → P1
Regressed by: 1772264

The query seems correct, the problem is likely in the while loop after it, it stops once width < aPreferredWidth assuming width was decreasing across results, that is not true anymore.
I think for now we could change the first if to say lastWidth <= width, we can go fancier in the future in bug 1494016.
This also means the test likely needs improvement (rich icon should have larger width than icon)

Assignee: nobody → yalmacki
Status: NEW → ASSIGNED
Whiteboard: [sng]

Yes indeed, the problem came from the fact that the loop assumed it would find smaller better-fitting icons after processing the smallest non-rich icon (in the case that this icon is larger than the preferred size, and that the preferred size is lower than 64px in width). I submitted a patch to fix this.

Flags: needinfo?(yalmacki)
Pushed by mak77@bonardo.net: https://hg.mozilla.org/integration/autoland/rev/80c9ebaa3dc0 Modify icon retrieval loop and test case to account for non width decreasing nature of new icon results. r=mak,places-reviewers
Status: ASSIGNED → RESOLVED
Closed: 22 days ago
Resolution: --- → FIXED
Target Milestone: --- → 132 Branch

I can see that Nightly is now using the appropriate icon.

You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: