1931491 - Incorrect flexbox layout in virtualbox.org

Status: RESOLVED DUPLICATE of bug 1926015

(Core :: Layout: Flexbox, defect)

Description

[Tom25519]

This bug not only happens on latest Firefox nightly for Windows (134.0a1), but also on Android (Fenix 134.0a1).

Comment 1

[Tom25519]

Attachment: Firefox.png

Comment 2

[Tom25519]

Attachment: Edge.jpeg

Comment 3

[Tom25519]

Attachment: Fenix.jpg

Comment 4

[Tom25519]

Attachment: Edge for Android.jpg

Comment 5

[Alice0775 White]

Regression window:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=f68eda3d65dd80db08c49ffabb424ac78bcc4332&tochange=d0336fee53a3e4eaee3e4db7c46245b9363084df

Comment 6

[BugBot [:suhaib / :marco]]

Set release status flags based on info from the regressing bug 1909761

:TYLin, since you are the author of the regressor, bug 1909761, could you take a look? Also, could you set the severity field?

For more information, please visit BugBot documentation.

Comment 7

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Testcase 1 (Partially reduced from virtualbox.org)

I got a partially reduced testcase.

Comment 8

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Testcase 1 (Partially reduced from virtualbox.org)

This is the partially reduced testcase.

Comment 9

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Testcase 1 (Partially reduced from virtualbox.org)

This is a reduced testcase from virtualbox.org.

Comment 10

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

In the testcase, there are nested flex containers, and the image descendants have aspect-ratio. This bug look very similar to bug 1926015 or bug 1910290.

I got the expected behavior by adding min-width: 0 to .box style to skip the content size measurement.

Comment 11

[Christoph Kerschbaumer [:ckerschb]]

Daniel, I am the REO for Fx133, this bug is still marked affecting Fx133 - I think we are considering it for a dot-release ride along - what's your take - how important is that to get fixed? I know it's rated as S3, however I still wanted to ask explicitly - thank you!

Comment 12

[Daniel Holbert [:dholbert]]

(In reply to Christoph Kerschbaumer [:ckerschb] from comment #11)

Daniel, I am the REO for Fx133, this bug is still marked affecting Fx133 - I think we are considering it for a dot-release ride along - what's your take - how important is that to get fixed? I know it's rated as S3, however I still wanted to ask explicitly - thank you!

TYLin is prepping a fix which we think should address this, re-landing a fix that previously landed -- it was backed out for perf reasons, but in retrospect (in light of bugs like this one) the correctness is more important than the perf cost (and we can iterate on clawing back the perf regression).

We're pretty confident in the correctness of that fix, so I think it'd be suitable for a 133 dot-release ride-along (assuming the perf hit is acceptable).

Comment 13

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Testcase 2 (minimum reproducible testcase)

Testcase 2 is reduced from Testcase 1, and it still fails with emilio's patch in https://phabricator.services.mozilla.com/D229621

Comment 14

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Attachment: Testcase 2 (minimum reproducible testcase)

I attached a wrong testcase in Comment 13. This is the correct reduced testcase.

Comment 15

[Daniel Holbert [:dholbert]]

(In reply to Daniel Holbert [:dholbert] from comment #12)

TYLin is prepping a fix [...]
We're pretty confident in the correctness of that fix, so I think it'd be suitable for a 133 dot-release ride-along (assuming the perf hit is acceptable).

(Actually: I'm a bit less sure about the dot-release ride-along now -- we added an additional patch in bug 1926015 to improve our caching and to address the perf issues, and that makes the fix as-a-whole a bit more risky. So: given that we shipped this in 132, I think we can live with shipping it in the about-to-go-out 133 release as well, and let the fix arrive as part of 134.)

Comment 16

[Daniel Holbert [:dholbert]]

From some brief testing, I think this is fixed in latest Nightly and can perhaps be closed as a dupe of bug 1926015 -- TYLin, could you confirm?

Comment 17

[Ting-Yu Lin [:TYLin] (PST, UTC-7)]

Yes, this bug is fixed by 1926015. I test it on Nightly 134.0a1 (2024-11-22).

The testcase in comment 14 has been captured as a wpt in https://wpt.fyi/results/css/css-sizing/intrinsic-percent-replaced-027.html