1905611 - [REGRESSION] DarkReader's Filter+ mode is extremely low res with gfx.webrender.svg-filter-effects enabled (Nightly only)

Status: RESOLVED FIXED

(Core :: Graphics: WebRender, defect, P1)

Description

[Artem S. Tashkinov]

Attachment: slashdot.jxl

Steps to reproduce:

Using: Firefox Nightly for Android
129.0a1 (Build #2016029799), hg-0c9e7d00b984+
GV: 129.0a1-20240630091518
AS: 129.20240629050339

Enabling Static Filter in the DarkReader extension makes websites unreadable, specially https://slashdot.org/

Actual results:

See the attached screenshot

Comment 1

[Artem S. Tashkinov]

Attachment: dark-reader-settings.jxl

Comment 2

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

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

Comment 3

[Artem S. Tashkinov]

This is quite a recent regression and Firefox Nightly did not have it a week ago or so (I update weekly).

Comment 4

[Chris Peterson [:cpeterson]]

Artem, thanks for reporting this bug. Knowing this is a recent regression in Firefox Nightly is a big help, too.

QA, do you have a device setup to use mozregression to bisect the first bad build with this regression?

Comment 5

[Chris Peterson [:cpeterson]]

Attachment: slashdot.png

Here is a PNG version of the slashdot.jxl JPEG-XL image. The page text is blurry.

The most recent update of the Dark Reader extension itself was version 4.9.86 on 2024-05-30. If this bug just started about a week ago, then it's surely a regression in Firefox, not Dark Reader.

https://addons.mozilla.org/en-US/firefox/addon/darkreader/versions/

Comment 6

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

The bug has a release status flag that shows some version of Firefox is affected, thus it will be considered confirmed.

Comment 7

[Delia Pop (Android QA)]

I was able to reproduce this issue with Samsung Galaxy A53 5G (Android 14) and used mozregression to Narrowed nightly regression window from [2024-06-23, 2024-06-25] (2 days) to [2024-06-24, 2024-06-25] (1 days) (~0 steps left).
The regression was introduced in Nightly 129.0a1 from 06/25.

Comment 8

[William Durand [:willdurand]]

(In reply to Chris Peterson [:cpeterson] from comment #5)

Created attachment 9410761 [details]
slashdot.png

Here is a PNG version of the slashdot.jxl JPEG-XL image. The page text is blurry.

The most recent update of the Dark Reader extension itself was version 4.9.86 on 2024-05-30. If this bug just started about a week ago, then it's surely a regression in Firefox, not Dark Reader.

https://addons.mozilla.org/en-US/firefox/addon/darkreader/versions/

I am seeing version 4.9.87 released Jun 27, 2024 on that page.

Comment 9

[William Durand [:willdurand]]

I can reproduce on desktop so if this bug is a Firefox bug, it isn't specific to Android. I am going to move it to our component for now.

Comment 10

[Artem S. Tashkinov]

(In reply to William Durand [:willdurand] from comment #8)

I am seeing version 4.9.87 released Jun 27, 2024 on that page.

This is reproducible both in DarkReader 4.9.86 and 4.9.87 which has been pushed earlier today.

Comment 11

[William Durand [:willdurand]]

According to mozregression, Bug 1896503 is the regressor: https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=2147e0051e231c4a0e6081bf3fd72e775290a607&tochange=beec388d4b0e6a50ac243ed86c0e3f1f48cbe744

:ahale could you please take a look?

Comment 12

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

Set release status flags based on info from the regressing bug 1896503

Comment 13

[Ashley Hale [:ahale]]

This may have been fixed by https://phabricator.services.mozilla.com/D214912 already, but probably not. I will attempt repro later today and figure out the failure mode involved - it looks like it is rendering at low resolution which is not expected.

Comment 14

[Ashley Hale [:ahale]]

I have confirmed this bug with pref gfx.webrender.svg-filter-effects enabled when viewing slashdot.org with Dark Reader set to Filter+ mode (More -> Theme Generation Mode = Filter+).

A workaround is to disable gfx.webrender.svg-filter-effects in about:config for now, I'll be looking into how this bug is happening in the coming days.

Moving this to the Graphics: WebRender component as the affected code is in WebRender.

Comment 15

[Ashley Hale [:ahale]]

A sampling of other websites I have confirmed are rendering at reduced resolution when using Dark Reader with Filter+ mode:
https://www.apple.com/
https://store.google.com/
https://www.youtube.com/
https://www.amazon.com/
https://en.wikipedia.org/wiki/Carrot (the main page looks fine but the actual wiki pages are low-res, probably due to their height)

Examples of websites I tried that are not affected by the low resolution issue (but seem to have other issues with Filter+ mode):
https://news.ycombinator.com/
https://www.facebook.com/
https://www.twitch.tv/

Note that wikipedia.org in particular has its own dark theme option (in the Beta features section on account settings), which works fine unlike the Dark Reader Filter+ mode, so I don't think this issue is specific to dark themes per se, but specific filter rendering behaviors.

I'm fairly certain that in WebRender the get_surface_rects code is computing a clipped rect that is too large to render in one texture, so it is reducing resolution, so I should be able to fix this by changing how get_surface_rects calculates the rects.

Comment 16

[Artem S. Tashkinov]

(In reply to Ashley Hale [:ahale] from comment #14)

I have confirmed this bug with pref gfx.webrender.svg-filter-effects enabled when viewing slashdot.org with Dark Reader set to Filter+ mode (More -> Theme Generation Mode = Filter+).

A workaround is to disable gfx.webrender.svg-filter-effects in about:config for now, I'll be looking into how this bug is happening in the coming days.

Moving this to the Graphics: WebRender component as the affected code is in WebRender.

gfx.webrender.svg-filter-effects -> false does not help on Android.

Comment 17

[Ashley Hale [:ahale]]

It would not help on a build older than the fix for that pref in https://phabricator.services.mozilla.com/D214912 mentioned in comment #13, but it does work here on desktop with nightly as of yesterday or possibly earlier.

Comment 18

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

Set release status flags based on info from the regressing bug 1896503

Comment 19

[Ashley Hale [:ahale]]

A bit of context:

• Given that this breaks an accessibility extension and is a regression, I am inclined to consider this S2 severity.
• This feature is only enabled in Nightly, and has a pref that acts as a kill-switch (gfx.webrender.svg-filter-effects in about:config), so it is easy to turn off as a mitigation.
• It is possible (though it seems to be rare) for regular web pages to do filtered stacking contexts that are larger than the view like this, which will hit this issue regardless of the extension.

Comment 20

[Ashley Hale [:ahale]]

Attachment: Bug 1905611 - apply clipping to PictureCompositeMode::SVGFEGraph r?#gfx-reviewers

Comment 21

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/bd4c6fe88753
apply clipping to PictureCompositeMode::SVGFEGraph rendering r=gfx-reviewers,gw

Comment 22

[tszentpeteri]

Backed out for causing mass failures.

• Backout link
• Push with failures
• Failure Log for Wr1
• Failure Log for R1
• Failure Log for R2
• Failure Log for R5
• Failure Log for R7
• Failure Log for R8
• Failure Log for C4
• Failure Log for reftest crash:

[task 2024-07-16T22:48:17.570Z] 22:48:17     INFO - REFTEST TEST-START | layout/svg/crashtests/441368-1.svg
[task 2024-07-16T22:48:17.571Z] 22:48:17     INFO - REFTEST TEST-LOAD | file:///builds/worker/workspace/build/tests/reftest/tests/layout/svg/crashtests/441368-1.svg | 3489 / 4046 (86%)
[task 2024-07-16T22:48:17.598Z] 22:48:17     INFO - ExceptionHandler::GenerateDump attempting to generate:</tmp/tmpr_dsfa8z.mozrunner/minidumps/435f59ec-208f-af6b-b22b-0e4302fa5a42.dmp
[task 2024-07-16T22:48:17.605Z] 22:48:17     INFO - ExceptionHandler::GenerateDump cloned child 6125
[task 2024-07-16T22:48:17.606Z] 22:48:17     INFO - ExceptionHandler::SendContinueSignalToChild sent continue signal to child
[task 2024-07-16T22:48:17.607Z] 22:48:17     INFO - ExceptionHandler::WaitForContinueSignal waiting for continue signal...
[task 2024-07-16T22:48:17.746Z] 22:48:17     INFO - ExceptionHandler::GenerateDump minidump generation <succeeded
[task 2024-07-16T22:48:17.777Z] 22:48:17     INFO - <Exiting due to channel error.
[task 2024-07-16T22:48:17.777Z] 22:48:17     INFO - Exiting due to channel error.
[task 2024-07-16T22:48:17.780Z] 22:48:17     INFO - Exiting due to channel error.
[task 2024-07-16T22:48:17.785Z] 22:48:17     INFO - Exiting due to channel error.
[task 2024-07-16T22:48:17.786Z] 22:48:17     INFO - Exiting due to channel error.
[task 2024-07-16T22:48:17.787Z] 22:48:17     INFO - Exiting due to channel error.
[task 2024-07-16T22:48:17.829Z] 22:48:17     INFO - REFTEST INFO | Downloading symbols from: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/X0NZvvEuSFiA0gAXSNePGg/artifacts/public/build/target.crashreporter-symbols.zip
[task 2024-07-16T22:48:33.844Z] 22:48:33     INFO - REFTEST INFO | Copy/paste: /builds/worker/fetches/minidump-stackwalk/minidump-stackwalk --symbols-url=https://symbols.mozilla.org/ --cyborg=/tmp/tmptq0z432v/435f59ec-208f-af6b-b22b-0e4302fa5a42.trace /tmp/tmpr_dsfa8z.mozrunner/minidumps/435f59ec-208f-af6b-b22b-0e4302fa5a42.dmp /tmp/tmpd12qdesr
[task 2024-07-16T22:48:37.712Z] 22:48:37     INFO - REFTEST INFO | Saved minidump as /builds/worker/workspace/build/blobber_upload_dir/435f59ec-208f-af6b-b22b-0e4302fa5a42.dmp
[task 2024-07-16T22:48:37.712Z] 22:48:37     INFO - REFTEST INFO | Saved app info as /builds/worker/workspace/build/blobber_upload_dir/435f59ec-208f-af6b-b22b-0e4302fa5a42.extra
[task 2024-07-16T22:48:38.016Z] 22:48:38     INFO - REFTEST PROCESS-CRASH | called `Option::unwrap()` on a `None` value [@ core::option::unwrap_failed] | layout/svg/crashtests/441368-1.svg 
[task 2024-07-16T22:48:38.016Z] 22:48:38     INFO - Process type: main
[task 2024-07-16T22:48:38.016Z] 22:48:38     INFO - Process pid: 1347
[task 2024-07-16T22:48:38.016Z] 22:48:38     INFO - Mozilla crash reason: called `Option::unwrap()` on a `None` value
[task 2024-07-16T22:48:38.016Z] 22:48:38     INFO - Crash dump filename: /tmp/tmpr_dsfa8z.mozrunner/minidumps/435f59ec-208f-af6b-b22b-0e4302fa5a42.dmp
[task 2024-07-16T22:48:38.017Z] 22:48:38     INFO - Operating system: Linux
[task 2024-07-16T22:48:38.017Z] 22:48:38     INFO -                   4.4.0-1014-aws #14taskcluster1-Ubuntu SMP Tue Apr 3 10:27:00 UTC 2018
[task 2024-07-16T22:48:38.017Z] 22:48:38     INFO - CPU: amd64
[task 2024-07-16T22:48:38.017Z] 22:48:38     INFO -      family 6 model 85 stepping 7
[task 2024-07-16T22:48:38.017Z] 22:48:38     INFO -      2 CPUs
[task 2024-07-16T22:48:38.017Z] 22:48:38     INFO - Linux Ubuntu 18.04 - bionic (Ubuntu 18.04.6 LTS)
[task 2024-07-16T22:48:38.018Z] 22:48:38     INFO - 
[task 2024-07-16T22:48:38.018Z] 22:48:38     INFO - Crash reason:  SIGSEGV / SEGV_MAPERR
[task 2024-07-16T22:48:38.018Z] 22:48:38     INFO - Crash address: 0x0000000000000000
[task 2024-07-16T22:48:38.018Z] 22:48:38     INFO - Crashing instruction: `mov dword [0x0], esi`
[task 2024-07-16T22:48:38.018Z] 22:48:38     INFO - Memory accessed by instruction:
[task 2024-07-16T22:48:38.018Z] 22:48:38     INFO -   0. Address: 0x0000000000000000
[task 2024-07-16T22:48:38.019Z] 22:48:38     INFO -      Size: 4
[task 2024-07-16T22:48:38.019Z] 22:48:38     INFO - Process uptime: not available
[task 2024-07-16T22:48:38.019Z] 22:48:38     INFO - 
[task 2024-07-16T22:48:38.019Z] 22:48:38     INFO - Thread 73 WRRende~ckend#3 (crashed)
[task 2024-07-16T22:48:38.019Z] 22:48:38     INFO -  0  libxul.so!MOZ_Crash(char const*, int, char const*) [Assertions.h:bd4c6fe88753c923a5a13e027e051659378a6979 : 317]
[task 2024-07-16T22:48:38.020Z] 22:48:38     INFO -     Found by: inlining
[task 2024-07-16T22:48:38.020Z] 22:48:38     INFO -  1  libxul.so!RustMozCrash [wrappers.cpp:bd4c6fe88753c923a5a13e027e051659378a6979 : 18 + 0xa]
[task 2024-07-16T22:48:38.020Z] 22:48:38     INFO -      rax = 0x000055c370d29db8    rdx = 0x00007f2f7c8f2334
[task 2024-07-16T22:48:38.021Z] 22:48:38     INFO -      rcx = 0x00007f2f7c8f1f3c    rbx = 0x00007f2fa13a3fe8
[task 2024-07-16T22:48:38.021Z] 22:48:38     INFO -      rsi = 0x0000000000000143    rdi = 0x00007f2f7c8f2130
[task 2024-07-16T22:48:38.021Z] 22:48:38     INFO -      rbp = 0x00007f2f7c8f2120    rsp = 0x00007f2f7c8f2120
[task 2024-07-16T22:48:38.021Z] 22:48:38     INFO -       r8 = 0x0000000000000014     r9 = 0x00007f2f7c8f2180
[task 2024-07-16T22:48:38.021Z] 22:48:38     INFO -      r10 = 0xc930fd8236c1054f    r11 = 0x00007f2f7c8f2314
[task 2024-07-16T22:48:38.022Z] 22:48:38     INFO -      r12 = 0x00007f2f7c8f2334    r13 = 0x00007f2f7c8f2130
[task 2024-07-16T22:48:38.022Z] 22:48:38     INFO -      r14 = 0x0000000000000042    r15 = 0x0000000000000143
[task 2024-07-16T22:48:38.022Z] 22:48:38     INFO -      rip = 0x00007f2fa80e5bfe
[task 2024-07-16T22:48:38.022Z] 22:48:38     INFO -     Found by: given as instruction pointer in context
[task 2024-07-16T22:48:38.022Z] 22:48:38     INFO -  2  libxul.so!mozglue_static::panic_hook [lib.rs:bd4c6fe88753c923a5a13e027e051659378a6979 : 98]
[task 2024-07-16T22:48:38.022Z] 22:48:38     INFO -     Found by: inlining
[task 2024-07-16T22:48:38.023Z] 22:48:38     INFO -  3  libxul.so!core::ops::function::Fn::call [function.rs:129f3b9964af4d4a709d1383930ade12dfe7c081 : 79 + 0xbb]
[task 2024-07-16T22:48:38.023Z] 22:48:38     INFO -      rbx = 0x00007f2fa13a3fe8    rbp = 0x00007f2f7c8f2560
[task 2024-07-16T22:48:38.023Z] 22:48:38     INFO -      rsp = 0x00007f2f7c8f2130    r12 = 0x00007f2f7c8f2334
[task 2024-07-16T22:48:38.023Z] 22:48:38     INFO -      r13 = 0x00007f2f7c8f2130    r14 = 0x0000000000000042
[task 2024-07-16T22:48:38.023Z] 22:48:38     INFO -      r15 = 0x0000000000000143    rip = 0x00007f2fa724ddc3
[task 2024-07-16T22:48:38.023Z] 22:48:38     INFO -     Found by: call frame info
[task 2024-07-16T22:48:38.024Z] 22:48:38     INFO -  4  libxul.so!<alloc::boxed::Box<F,A> as core::ops::function::Fn<Args>>::call [boxed.rs:129f3b9964af4d4a709d1383930ade12dfe7c081 : 2036]
[task 2024-07-16T22:48:38.024Z] 22:48:38     INFO -     Found by: inlining
[task 2024-07-16T22:48:38.024Z] 22:48:38     INFO -  5  libxul.so!std::panicking::rust_panic_with_hook [panicking.rs:129f3b9964af4d4a709d1383930ade12dfe7c081 : 799 + 0x2]
[task 2024-07-16T22:48:38.024Z] 22:48:38     INFO -      rbx = 0x0000000000000001    rbp = 0x00007f2f7c8f2660
[task 2024-07-16T22:48:38.024Z] 22:48:38     INFO -      rsp = 0x00007f2f7c8f2570    r12 = 0x00007f2fa85a9268
[task 2024-07-16T22:48:38.025Z] 22:48:38     INFO -      r13 = 0x00007f2f7c8f2678    r14 = 0x00007f2f757c6800
[task 2024-07-16T22:48:38.025Z] 22:48:38     INFO -      r15 = 0x0000000000000000    rip = 0x00007f2fa759ea68
[task 2024-07-16T22:48:38.025Z] 22:48:38     INFO -     Found by: call frame info
[task 2024-07-16T22:48:38.025Z] 22:48:38     INFO -  6  libxul.so!std::panicking::begin_panic_handler::{{closure}} [panicking.rs:129f3b9964af4d4a709d1383930ade12dfe7c081 : 656 + 0xf]
[task 2024-07-16T22:48:38.025Z] 22:48:38     INFO -      rbx = 0x00007f2f71687848    rbp = 0x00007f2f7c8f26a0
[task 2024-07-16T22:48:38.025Z] 22:48:38     INFO -      rsp = 0x00007f2f7c8f2670    r12 = 0x0000000000000028
[task 2024-07-16T22:48:38.026Z] 22:48:38     INFO -      r13 = 0x0000000000000001    r14 = 0x00007f2fb56bbb00
[task 2024-07-16T22:48:38.026Z] 22:48:38     INFO -      r15 = 0x0000000000000000    rip = 0x00007f2fa759e75a
<...>
[task 2024-07-16T22:48:38.035Z] 22:48:38     INFO - 21  libxul.so!webrender::frame_builder::FrameBuilder::build [frame_builder.rs:bd4c6fe88753c923a5a13e027e051659378a6979 : 567 + 0x609c]
[task 2024-07-16T22:48:38.035Z] 22:48:38     INFO -      rbx = 0x00007f2f7c8f7b30    rbp = 0x00007f2f7c8f8700
[task 2024-07-16T22:48:38.036Z] 22:48:38     INFO -      rsp = 0x00007f2f7c8f73c0    r12 = 0x0000000000000920
[task 2024-07-16T22:48:38.036Z] 22:48:38     INFO -      r13 = 0x0000000000000018    r14 = 0x00007f2f7c8f8108
[task 2024-07-16T22:48:38.036Z] 22:48:38     INFO -      r15 = 0x0000000000000004    rip = 0x00007f2fa7b2291c
[task 2024-07-16T22:48:38.036Z] 22:48:38     INFO -     Found by: call frame info
[task 2024-07-16T22:48:38.036Z] 22:48:38     INFO - 22  libxul.so!webrender::render_backend::Document::build_frame [render_backend.rs:bd4c6fe88753c923a5a13e027e051659378a6979 : 528 + 0x48]
[task 2024-07-16T22:48:38.037Z] 22:48:38     INFO -      rbx = 0x00007f2f7c8f9d80    rbp = 0x00007f2f7c8f8a90
[task 2024-07-16T22:48:38.037Z] 22:48:38     INFO -      rsp = 0x00007f2f7c8f8710    r12 = 0x00007f2f71687b50
[task 2024-07-16T22:48:38.037Z] 22:48:38     INFO -      r13 = 0x00007f2f716879b8    r14 = 0x00007f2f71687538
[task 2024-07-16T22:48:38.037Z] 22:48:38     INFO -      r15 = 0x00007f2f71687ba0    rip = 0x00007f2fa7a62cad
[task 2024-07-16T22:48:38.037Z] 22:48:38     INFO -     Found by: call frame info
[task 2024-07-16T22:48:38.038Z] 22:48:38     INFO - 23  libxul.so!webrender::render_backend::RenderBackend::update_document [render_backend.rs:bd4c6fe88753c923a5a13e027e051659378a6979 : 1447 + 0x32]
[task 2024-07-16T22:48:38.038Z] 22:48:38     INFO -      rbx = 0x0000000000002328    rbp = 0x00007f2f7c8fa8c0
[task 2024-07-16T22:48:38.038Z] 22:48:38     INFO -      rsp = 0x00007f2f7c8f8aa0    r12 = 0x00007f2f7c8fca40
[task 2024-07-16T22:48:38.039Z] 22:48:38     INFO -      r13 = 0x00007f2f7c8fd340    r14 = 0x0000000000000000
[task 2024-07-16T22:48:38.039Z] 22:48:38     INFO -      r15 = 0x00007f2f71687fc8    rip = 0x00007f2fa7a37b6d
[task 2024-07-16T22:48:38.039Z] 22:48:38     INFO -     Found by: call frame info

Comment 23

[Ashley Hale [:ahale]]

These failures are confusing to me as I did a try run before landing that showed no issues with these reftests.

Comment 24

[Timothy Nikkel (:tnikkel)]

(In reply to Ashley Hale [:ahale] from comment #23)

These failures are confusing to me as I did a try run before landing that showed no issues with these reftests.

I looked at your try run. It seems the changeset you pushed is based on revisions that are from the mozilla-beta tree but do not exist in the mozilla-central (or autoland) tree.

The changeset you pushed

https://hg.mozilla.org/try/rev/e55d92521183a0de9fa89d6ebce66e18b999bc8a

has parent

https://hg.mozilla.org/try/rev/25cada7f1b5ca8d631637705ebc795dd8f739887

but there is no 25cada7f1b5ca8d631637705ebc795dd8f739887 rev in m-c or autoland, there is however

https://hg.mozilla.org/releases/mozilla-beta/rev/25cada7f1b5ca8d631637705ebc795dd8f739887

And then since that is mozilla-beta your new filter code is disabled, that would explain what you saw. I'm not sure how you pushed on top of a beta revision though.

Comment 25

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

There is an r+ patch which didn't land and no activity in this bug for 2 weeks.
:ahale, could you have a look please?
If you still have some work to do, you can add an action "Plan Changes" in Phabricator.
For more information, please visit BugBot documentation.

Comment 26

[Ashley Hale [:ahale]]

I'm working on this as time permits, I've identified that the reftest failures all involve feGaussianBlur or feDropShadow, I will debug what is going wrong with the subregion math or blur copy task rect math, and may make some wrench reftests in service of that debugging.

Comment 27

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/ebe6cedc1f12
apply clipping to PictureCompositeMode::SVGFEGraph r=gfx-reviewers,gw

Comment 28

[Ashley Hale [:ahale]]

Since this regression was the only issue found in Nightly over some months of testing, the above patch enables gfx.webrender.svg-filter-effects for all channels so it will ride the trains as normal, a couple weeks in Nightly before we go to Beta should be sufficient to confirm the clipping that fixes this regression doesn't cause any other regressions.

Comment 29

[Sandor Molnar[:smolnar]]

Backed out for causing reftest failures

Backout link: https://hg.mozilla.org/integration/autoland/rev/ebf0e33ba93e6be120a2c9d1b257b840b467c235

Push with failures

Failure log

Comment 30

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/bb0e07ccdb0f
apply clipping to PictureCompositeMode::SVGFEGraph r=gfx-reviewers,gw

Comment 31

[pstanciu]

https://hg.mozilla.org/mozilla-central/rev/bb0e07ccdb0f

Comment 32

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

Since nightly and release are affected, beta will likely be affected too.
For more information, please visit BugBot documentation.

Comment 33

[Ashley Hale [:ahale]]

So we're up to 3 pixel5 reftest failures, and an unexpected pass on a WPT reftest (which is probably just a metadata file I forgot to delete, it is expected to pass with this new code). I'll be looking into these.