1892432 - BBB icons displayed incorrectly

Status: RESOLVED DUPLICATE of bug 1892262

(Core :: Layout: Text and Fonts, defect)

Description

[Paul Menzel]

Attachment: Screenshot of problematic icon. The microphone should be centered.

User Agent: Mozilla/5.0 (X11; Linux x86_64; rv:127.0) Gecko/20100101 Firefox/127.0

Steps to reproduce:

(This is a regression and works with Firefox 125.0.1.)

Join old BigBlueButton 2.2 (1157) instance.

Actual results:

The icons are rendered incorrectly.

Expected results:

Render the items correctly.

The problem is not there with newer BigBlueButton versions.

Comment 1

[BugBot [:suhaib / :marco/ :calixte]]

The Bugbug bot thinks this bug should belong to the 'Core::Widget: Gtk' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 2

[Martin Stránský [:stransky] (ni? me)]

Can you please use mozregression tool to find broken commit?
https://fedoraproject.org/wiki/How_to_debug_Firefox_problems#Use_Mozregression_tool
Thanks.

Comment 3

[Paul Menzel]

I never got it to work in the past. I am going to try again over the weekend.

Comment 4

[Paul Menzel]

Attachment: Minimal example

Comment 5

[Paul Menzel]

Still haven’t had time, but the regression is still there with Nightly 128.0a1, 20240514095049.

Comment 6

[Paul Menzel]

$ mozregression --good 2024-03-01
[…]
 7:06.65 INFO: Downloading build from: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/AzPPld4hRvmTiXPX27bc4w/runs/0/artifacts/public%2Fbuild%2Ftarget.tar.bz2
===== Downloaded 100% =====
 7:14.87 INFO: Running autoland build built on 2024-04-15 16:53:28.918000, revision 48974604
 7:16.39 WARNING: Skipping build c9e37a7248ac: Unable to find build info using the taskcluster route 'gecko.v2.autoland.shippable.revision.c9e37a7248acdd23ac32aef0dd2b49910cdf8049.firefox.linux64-opt'
 7:17.17 WARNING: Skipping build 98d8a81045e1: Unable to find build info using the taskcluster route 'gecko.v2.autoland.shippable.revision.98d8a81045e10519bd7621fb4d556f17ad0180db.firefox.linux64-opt'
 7:33.46 INFO: Launching /tmp/tmpq2rwu4bp/firefox/firefox
 7:33.46 INFO: Application command: /tmp/tmpq2rwu4bp/firefox/firefox -profile /tmp/tmppgtobtq8.mozrunner
 7:33.47 INFO: application_buildid: 20240415145812
 7:33.47 INFO: application_changeset: 4897460458d13272e6016fd9e4c184c3c41cbd74
 7:33.47 INFO: application_name: Firefox
 7:33.47 INFO: application_repository: https://hg.mozilla.org/integration/autoland
 7:33.47 INFO: application_version: 126.0a1
Was this integration build good, bad, or broken? (type 'good', 'bad', 'skip', 'retry', 'back' or 'exit' and press Enter): bad
 7:41.27 INFO: Narrowed integration regression window from [02fb4b67, 61472b53] (6 builds) to [02fb4b67, 48974604] (3 builds) (~1 steps left)
 7:41.27 INFO: Pushlog:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=02fb4b67c7e1d1db8d6c77992572c91e93b511ad&tochange=4897460458d13272e6016fd9e4c184c3c41cbd74

 7:41.27 INFO: Using local file: /dev/shm/.mozilla/mozregression/persist/8d4ab18c9e32-shippable--autoland--target.tar.bz2 (downloaded in background)
 7:41.27 INFO: Running autoland build built on 2024-04-15 14:35:17.409000, revision 8d4ab18c
 7:56.78 INFO: Launching /tmp/tmp8vsgtazd/firefox/firefox
 7:56.78 INFO: Application command: /tmp/tmp8vsgtazd/firefox/firefox -profile /tmp/tmpzxpf590s.mozrunner
 7:56.79 INFO: application_buildid: 20240415135744
 7:56.79 INFO: application_changeset: 8d4ab18c9e32b2b43ec1325206df74b0a971ba1b
 7:56.79 INFO: application_name: Firefox
 7:56.79 INFO: application_repository: https://hg.mozilla.org/integration/autoland
 7:56.79 INFO: application_version: 126.0a1
Was this integration build good, bad, or broken? (type 'good', 'bad', 'skip', 'retry', 'back' or 'exit' and press Enter): bad
 8:02.60 INFO: Narrowed integration regression window from [02fb4b67, 48974604] (3 builds) to [02fb4b67, 8d4ab18c] (2 builds) (~1 steps left)
 8:02.60 INFO: No more integration revisions, bisection finished.
 8:02.60 INFO: Last good revision: 02fb4b67c7e1d1db8d6c77992572c91e93b511ad
 8:02.60 INFO: First bad revision: 8d4ab18c9e32b2b43ec1325206df74b0a971ba1b
 8:02.60 INFO: Pushlog:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=02fb4b67c7e1d1db8d6c77992572c91e93b511ad&tochange=8d4ab18c9e32b2b43ec1325206df74b0a971ba1b

Comment 7

[Paul Menzel]

Probably https://hg.mozilla.org/integration/autoland/rev/7acc4ef7877fa6b0118604a5aa3f2bf70e6178c0 (Jonathan Kew — Bug 1891446 - Experimentally enable symmetrical letter-spacing in Nightly builds. r=layout-reviewers,emilio).

Comment 8

[Jonathan Kew [:jfkthame]]

Right, the issue is that the page is using letter-spacing: -.65rem (found in the .avatar--Z2lyL8K::after, .avatar--Z2lyL8K::before rules) to alter the effective width of those icons, and is assuming that letter-spacing simply alters the right sidebearing of the glyph. With the symmetrical letter-spacing model, this is no longer true, and the positioning therefore shifts.

Setting layout.css.letter-spacing.model to 0 will restore the old behavior, but it's unwise for sites to depend on this; it's an implementation detail and does not actually match what the CSS Text spec says, so shouldn't be relied on for important aspects of layout.

Comment 9

[Jonathan Kew [:jfkthame]]

Ah, I noticed in comment 0

The problem is not there with newer BigBlueButton versions.

which suggests they have changed this (fragile) implementation already in some way (though I'm not familiar with BigBlueButton at all...)

Comment 10

[Paul Menzel]

Thank you for looking into this and the explanation.

Does Firefox have a no-regressions policy like the Linux kernel?

Comment 11

[Jonathan Kew [:jfkthame]]

Not as such, afaik. Regressions are of course taken seriously, and ideally should be fixed promptly, but the trade-off between making a necessary or desirable change vs not disrupting any existing content is sometimes a tricky call.

In this case, the changed behavior (symmetrical letter-spacing) is currently enabled only for Nightly builds, precisely so that we can begin to discover cases like this where it may negatively affect sites. No decision has been taken regarding whether it will ship on release channels, and there's an open discussion around the feature in the CSS Working Group.

Comment 12

[Paul Menzel]

Thank you. Then I’ll leave everything to you with this issue.

PS: Do you have an URL to the discussion of the CSS Working Group?

Comment 13

[Jonathan Kew [:jfkthame]]

The most recent issue is https://github.com/w3c/csswg-drafts/issues/10193, though there have been related discussions in years past, too.

Comment 14

[Jonathan Kew [:jfkthame]]

Dup'ing to bug 1892262, where there's a patch that will resolve this.