Open Bug 1991800 Opened 15 days ago Updated 3 days ago

mail.google.com - Gmail stuck on loading screen, can't log in (URL redirects to chat)

Categories

(Core :: DOM: Core & HTML, defect)

Product:
Core
Component:
DOM: Core & HTML
Version:
Firefox 145
Platform:
x86_64
Linux
Type:
defect

Tracking

()

Status:
ASSIGNED
Tracking Flags:
Tracking Status
firefox-esr115 --- unaffected
firefox-esr140 --- unaffected
firefox143 --- unaffected
firefox144 --- unaffected
firefox145 + fixed

People

(Reporter: bj, Assigned: hsivonen)

References

(Regression,
URL
)

Details

(Keywords: nightly-community, regression, Whiteboard: [webcompat:sightline][webcompat:japan])

Attachments

(3 files)

Image ten minutes after trying to load the page
44.65 KB, image/png
Details
url.mp4
301.92 KB, video/mp4
Details
Mac_Crash_Report.txt
6.69 KB, text/plain
Details

Attempts to log into Gmail fail to complete, the load screen stays on.

Mozregression returned this:

0:33.08 INFO: Narrowed integration regression window from [fc66668a, 90a55b14] (18 builds) to [3e8cefc7, 90a55b14] (9 builds) (~3 steps left)
10:33.08 INFO: Pushlog:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=3e8cefc774718cc1afd637f37e5f516d6351f4d6&tochange=90a55b14824b4a4856737989385a889ec0bb4902

10:33.08 INFO: Downloading build from: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/Mfe2VHeHTOi6jABetvoLUw/runs/0/artifacts/public%2Fbuild%2Ftarget.tar.xz
===== Downloaded 100% =====
10:39.51 INFO: Running autoland build built on 2025-09-30 11:37:15.256000, revision eb670068
10:41.67 WARNING: Skipping build 9c6b04c2756b: Unable to find build info using the taskcluster route 'gecko.v2.autoland.shippable.revision.9c6b04c2756b71599c77921514aa72b0ec8ad56c.firefox.linux64-opt'
10:42.45 WARNING: Skipping build 0ddc5e1f9408: Unable to find build info using the taskcluster route 'gecko.v2.autoland.shippable.revision.0ddc5e1f940863db347f07071885ee3e9367eb22.firefox.linux64-opt'
10:43.28 WARNING: Skipping build f135d5933977: Unable to find build info using the taskcluster route 'gecko.v2.autoland.shippable.revision.f135d59339772805a3b0a2e21ff1f67453041d3b.firefox.linux64-opt'
10:43.29 INFO: Launching /tmp/tmp90gbfbth/firefox/firefox
10:43.29 INFO: Application command: /tmp/tmp90gbfbth/firefox/firefox -profile /tmp/tmprfyr_riw.mozrunner
10:43.29 INFO: application_buildid: 20250930112234
10:43.29 INFO: application_changeset: eb67006865895c703f5b3edeaef8db719c084712
10:43.29 INFO: application_name: Firefox
10:43.29 INFO: application_repository: https://hg.mozilla.org/integration/autoland
10:43.29 INFO: application_version: 145.0a1
Was this integration build good, bad, or broken? (type 'good', 'bad', 'skip', 'retry', 'back' or 'exit' and press Enter): b
11:33.67 INFO: Narrowed integration regression window from [3e8cefc7, 90a55b14] (6 builds) to [3e8cefc7, eb670068] (2 builds) (~1 steps left)
11:33.67 INFO: No more integration revisions, bisection finished.
11:33.67 INFO: Last good revision: 3e8cefc774718cc1afd637f37e5f516d6351f4d6
11:33.67 INFO: First bad revision: eb67006865895c703f5b3edeaef8db719c084712
11:33.67 INFO: Pushlog:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=3e8cefc774718cc1afd637f37e5f516d6351f4d6&tochange=eb67006865895c703f5b3edeaef8db719c084712

I am able to log into Google and use several other services. Only Gmail has had a problem.

Now gmail.com is redirecting me to https://mail.google.com/chat/u/0/#mail/u/0/ , or maybe it was before and I didn't notice.
This time mozregression gave a different result:

11:50.90 INFO: Narrowed integration regression window from [fc66668a, 76d615f2] (6 builds) to [fc66668a, e98932aa] (4 builds) (~2 steps left)
11:50.90 INFO: Pushlog:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=fc66668a8737d8f41d6f8ef83b740e7bfdc34f29&tochange=e98932aa84ff5395a02700c4e8908824b982fdca

11:50.90 INFO: Using local file: /home/bj/.mozilla/mozregression/persist/c5b047ab1e56-pgo--autoland--target.tar.xz (downloaded in background)
11:50.90 INFO: Running autoland build built on 2025-09-30 10:54:07.029000, revision c5b047ab
11:51.78 WARNING: Skipping build 1bea44284849: Unable to find build info using the taskcluster route 'gecko.v2.autoland.shippable.revision.1bea442848498086c687adb0cf33ac828cb2c34d.firefox.linux64-opt'
11:55.06 INFO: Launching /tmp/tmpvnf6o9aw/firefox/firefox
11:55.06 INFO: Application command: /tmp/tmpvnf6o9aw/firefox/firefox -profile /tmp/tmp6y16wl02.mozrunner
11:55.07 INFO: application_buildid: 20250930102835
11:55.07 INFO: application_changeset: c5b047ab1e5698b5006b6a9c456b7c086b470aaf
11:55.07 INFO: application_name: Firefox
11:55.07 INFO: application_repository: https://hg.mozilla.org/integration/autoland
11:55.07 INFO: application_version: 145.0a1
Was this integration build good, bad, or broken? (type 'good', 'bad', 'skip', 'retry', 'back' or 'exit' and press Enter): b
12:38.97 INFO: Narrowed integration regression window from [fc66668a, e98932aa] (3 builds) to [fc66668a, c5b047ab] (2 builds) (~1 steps left)
12:38.97 INFO: No more integration revisions, bisection finished.
12:38.97 INFO: Last good revision: fc66668a8737d8f41d6f8ef83b740e7bfdc34f29
12:38.97 INFO: First bad revision: c5b047ab1e5698b5006b6a9c456b7c086b470aaf
12:38.97 INFO: Pushlog:
https://hg.mozilla.org/integration/autoland/pushloghtml?fromchange=fc66668a8737d8f41d6f8ef83b740e7bfdc34f29&tochange=c5b047ab1e5698b5006b6a9c456b7c086b470aaf

The Bugbug bot thinks this bug should belong to the 'Firefox Build System::Task Configuration' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Component: Untriaged → Task Configuration
Product: Firefox → Firefox Build System
Component: Task Configuration → Untriaged
Product: Firefox Build System → Firefox
Duplicate of this bug: 1991810
URL: https://mail.google.com
Component: Untriaged → Site Reports
Product: Firefox → Web Compatibility
Summary: Can't log into Gmail with Nightly → mail.google.com - Can't log into Gmail with Nightly
Whiteboard: [webcompat:sightline][webcompat:japan]

for me mozregression said it come from

2025-10-01T10:42:23.276000: DEBUG : Found commit message:
Bug 1968903 - Avoid creating script and CSS loaders for documents loaded as data. r=firefox-style-system-reviewers,emilio,layout-reviewers

Differential Revision: https://phabricator.services.mozilla.com/D264779
Duplicate of this bug: 1991821

[Tracking Requested - why for this release]: issue with a major website.

status-firefox143: --- → unaffected
status-firefox144: --- → unaffected
status-firefox-esr115: --- → unaffected
status-firefox-esr140: --- → unaffected
tracking-firefox145: --- → ?
Summary: mail.google.com - Can't log into Gmail with Nightly → mail.google.com - Gmail stuck on loading screen, can't log in (URL redirects to chat)
Component: Site Reports → DOM: Core & HTML
Keywords: regression
Product: Web Compatibility → Core
Regressed by: 1968903

:hsivonen, since you are the author of the regressor, bug 1968903, could you take a look? Also, could you set the severity field?

For more information, please visit BugBot documentation.

Flags: needinfo?(hsivonen)
Duplicate of this bug: 1991793

Now that the regressor has been backed out, let's handle the rest of this as S3.

Shippable-config builds that only turn off the script loader but not the CSS loader are coming up for testing in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=155965

Severity: -- → S3
Flags: needinfo?(hsivonen)

Meanwhile:

I can't repro this myself.

It would be useful if someone who has both a Google account that failed with yesterday's Nightly and a Google account that didn't fail could see how the JS files that Google serves in the two cases differ (if they differ) when loading a world-visible GDoc like https://docs.google.com/document/d/16ZrFP7yj4YIcFWvBMNQzjVw1d6p2iyVRhLi5janYT-s/edit?tab=t.0 .

(To avoid posting secrets here, it's safer to test a world-visible GDoc than Gmail. Also, to avoid posting secrets, diffing the JS at your end is safer than posting a larger capture here.)

My current hypothesis is that a) Google is serving different versions of their code for accounts that saw the failure vs. accounts that didn't see the failure and b) when the failure happens, Google is asking for style computation results on a "loadedAsData" document.

The regressing patch turned off styling for loadedAsData documents created from multiple scenarios, but not from cloning a document. Turning styling off for DOMParser seems to matter the most for performance, so if the upcoming test builds are OK, the next step would be narrowing down the problem scenario in the hope that it's not the DOMParser one.

(In reply to Henri Sivonen (:hsivonen) from comment #11)

It would be useful if someone who has both a Google account that failed with yesterday's Nightly and a Google account that didn't fail could see how the JS files that Google serves in the two cases differ (if they differ) when loading a world-visible GDoc like https://docs.google.com/document/d/16ZrFP7yj4YIcFWvBMNQzjVw1d6p2iyVRhLi5janYT-s/edit?tab=t.0 .

This loads fine for me, even with the test profile with just the problematic account logged in. Only Gmail doesn't load, and tries to redirect to Chat somehow.

Attached video url.mp4 —

See the change in the URL to https://mail.google.com/chat/u/0/#mail/u/0/.

If I close the tab and reopen it, it loads https://mail.google.com/chat/u/0/#chat/home.

(In reply to Henri Sivonen (:hsivonen) from comment #10)

Now that the regressor has been backed out, let's handle the rest of this as S3.

Shippable-config builds that only turn off the script loader but not the CSS loader are coming up for testing in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=155965

in linux. this build work for me. gmail works normally.

(In reply to Francesco Lodolo [:flod] from comment #13)

(In reply to Henri Sivonen (:hsivonen) from comment #11)

It would be useful if someone who has both a Google account that failed with yesterday's Nightly and a Google account that didn't fail could see how the JS files that Google serves in the two cases differ (if they differ) when loading a world-visible GDoc like https://docs.google.com/document/d/16ZrFP7yj4YIcFWvBMNQzjVw1d6p2iyVRhLi5janYT-s/edit?tab=t.0 .

This loads fine for me, even with the test profile with just the problematic account logged in. Only Gmail doesn't load, and tries to redirect to Chat somehow.

Thanks. Sadly, then it seems that folks who saw the problem with Gmail can't use a public GDoc for testing with fewer secrets to redact.

(In reply to mix5003 from comment #15)

(In reply to Henri Sivonen (:hsivonen) from comment #10)

Now that the regressor has been backed out, let's handle the rest of this as S3.

Shippable-config builds that only turn off the script loader but not the CSS loader are coming up for testing in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=155965

in linux. this build work for me. gmail works normally.

Just to confirm: And Gmail for the same Google account was broken with Nightly?

(In reply to mix5003 from comment #15)

(In reply to Henri Sivonen (:hsivonen) from comment #10)

Now that the regressor has been backed out, let's handle the rest of this as S3.

Shippable-config builds that only turn off the script loader but not the CSS loader are coming up for testing in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=155965

in linux. this build work for me. gmail works normally.

Which build are you using? For me, the Linux build does not show up as ready, yet.

(In reply to Henri Sivonen (:hsivonen) from comment #16)

(In reply to mix5003 from comment #15)

(In reply to Henri Sivonen (:hsivonen) from comment #10)

Now that the regressor has been backed out, let's handle the rest of this as S3.

Shippable-config builds that only turn off the script loader but not the CSS loader are coming up for testing in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=155965

in linux. this build work for me. gmail works normally.

Just to confirm: And Gmail for the same Google account was broken with Nightly?

yes same account. actually same profile folder

(In reply to Henri Sivonen (:hsivonen) from comment #17)

(In reply to mix5003 from comment #15)

(In reply to Henri Sivonen (:hsivonen) from comment #10)

Now that the regressor has been backed out, let's handle the rest of this as S3.

Shippable-config builds that only turn off the script loader but not the CSS loader are coming up for testing in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=155965

in linux. this build work for me. gmail works normally.

Which build are you using? For me, the Linux build does not show up as ready, yet.

i not sure is my step works or not. i go to that link click instr at Linux Shippable > Click Artifacts and Debugging Tools at bottom panel and download target.tar.xz

(In reply to Henri Sivonen (:hsivonen) from comment #16)

Thanks. Sadly, then it seems that folks who saw the problem with Gmail can't use a public GDoc for testing with fewer secrets to redact.

Would be nice to pinpoint the affected mozilla internal gdoc, copy it, and strip it down to achieve a minimal gdoc repro.

Or I'm confused with Hsin-Yi having weirdness in GDocs yesterday. I wonder she had the same issue in gmail.

Even the first 3 loaded JS (m=b, m=a, and m=m,m_i,...) have big size differences :

  • working account : content-length 499290, 147941, 656048
  • account with the issue : content-length 541510, 162554, 683657

All 6 scripts have the same "last-modified Sat, 27 Sep 2025 02:05:31 GMT".

The account with the issue works in Chrome, with the same JS sizes (at least for the first 3 requests).

So it shows the difference, interesting. Do you have the URLs for those scripts?

Gmail loads fine in the broken profile with the macOS Shippable build.

(In reply to Kagami Rosylight [:saschanaz] (they/them) from comment #22)

So it shows the difference, interesting. Do you have the URLs for those scripts?

Working :

Broken :

The next scripts are the same, but there are many others and I don't know how to compare them easily.

I can confirm the try build from comment 10 to also fix the gmail issue for me on Mac arm64 using the same profile as before.

The broken build also made the google mail, google calendar and other tabs spin for a long time.
The other tabs I had to close and load the URL in a new tab where it then worked.

The bug is marked as tracked for firefox145 (nightly). We have limited time to fix this, the soft freeze is in 8 days. However, the bug still isn't assigned and has low severity.

:hsinyi, could you please find an assignee and increase the severity for this tracked bug? Given that it is a regression and we know the cause, we could also simply backout the regressor. If you disagree with the tracking decision, please talk with the release managers.

For more information, please visit BugBot documentation.

Flags: needinfo?(htsai)

(In reply to BugBot [:suhaib / :marco/ :calixte] from comment #26)

The bug is marked as tracked for firefox145 (nightly). We have limited time to fix this, the soft freeze is in 8 days. However, the bug still isn't assigned and has low severity.

:hsinyi, could you please find an assignee and increase the severity for this tracked bug? Given that it is a regression and we know the cause, we could also simply backout the regressor. If you disagree with the tracking decision, please talk with the release managers.

For more information, please visit BugBot documentation.

The regressor patches were backed out.

status-firefox145: affected → fixed
Flags: needinfo?(htsai)

Thank you for the testing so far.

Shippable builds that restore stylability to DOMParser docs only coming up in:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156046

(The point being that if these don't work, we lose the most important case of omitting stylability.)

Attached file Mac_Crash_Report.txt —
> Shippable builds that restore stylability to DOMParser docs only coming up in: > https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156046 That new build crashes on start for me.

Shippable builds that restore stylability to DOMParser docs only coming up in:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156046

That new build crashes on start for my Arm64 Mac.
See attachment 9517493 [details] for Mac Crash Report data

(In reply to Henri Sivonen (:hsivonen) from comment #28)

Shippable builds that restore stylability to DOMParser docs only coming up in:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156046

for me this build not fixed this issue. it stuck at gmail loading screen again. (same as before i use instr version)

The correct version to test as the "B" items, and for Mac, the "macOS Shippable" row. "B" for that row isn't ready, yet.

Starting a new batch of shippable-config builds that restore styling to DOMParser with also an added null check:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156083

(In reply to Henri Sivonen (:hsivonen) from comment #33)

Starting a new batch of shippable-config builds that restore styling to DOMParser with also an added null check:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156083

this version is not fixed this issue. (test on Linux Shippable B version)

can i ask something? i guess instr is vanila build version and B is optimized build version by Profile-guided optimization.
if this case, this is not performance relate. is it safe to use instr version for test?
if it safe it can provide faster feedback for this bug. and if not i will always test with B next times.

(In reply to mix5003 from comment #35)

(In reply to Henri Sivonen (:hsivonen) from comment #33)

Starting a new batch of shippable-config builds that restore styling to DOMParser with also an added null check:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156083

this version is not fixed this issue. (test on Linux Shippable B version)

Thanks. To have more data points: Can someone else who saw the problem with the September 30 Nightly and didn't see the problem with the builds from comment 10 check if the problem occurs with a "B" build from comment 33?

can i ask something? i guess instr is vanila build version and B is optimized build version by Profile-guided optimization.
if this case, this is not performance relate. is it safe to use instr version for test?
if it safe it can provide faster feedback for this bug. and if not i will always test with B next times.

instr is not vanilla but instrumented to gather a profile. An instr build could write a .profraw file somewhere. In theory, if you are really unlucky, it could overwrite a .profraw file you wanted to keep. Whether or not the instr build is valid for testing depends on whether the bug that's being tested has a timing component. The B builds will most closely match the timing etc. characteristics of Nightly / release.

(In reply to mix5003 from comment #35)

(In reply to Henri Sivonen (:hsivonen) from comment #33)

Starting a new batch of shippable-config builds that restore styling to DOMParser with also an added null check:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156083

Gmail is broken for me with this build (macOS). Same behavior as comment 14.

(In reply to Henri Sivonen (:hsivonen) from comment #36)

(In reply to mix5003 from comment #35)

(In reply to Henri Sivonen (:hsivonen) from comment #33)

Starting a new batch of shippable-config builds that restore styling to DOMParser with also an added null check:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156083

this version is not fixed this issue. (test on Linux Shippable B version)

Thanks. To have more data points: Can someone else who saw the problem with the September 30 Nightly and didn't see the problem with the builds from comment 10 check if the problem occurs with a "B" build from comment 33?

i retest with "B" version again and this is result for linux shippable
comment #10 https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=155965 fixed
comment #28 https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156046 not fixed (no crash like mac version )
comment #33 https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156083 not fixed

i am not sure about time zone. but for me i found this issue at 1 October. (but i think i use 30 September version at you(or builder) timezone).
so i download from archive and re-test again
https://ftp.mozilla.org/pub/firefox/nightly/2025/09/2025-09-30-04-30-57-mozilla-central/ < No issue here
https://ftp.mozilla.org/pub/firefox/nightly/2025/09/2025-09-30-16-29-01-mozilla-central/ < Issue start here
https://ftp.mozilla.org/pub/firefox/nightly/2025/10/2025-10-01-07-16-00-mozilla-central/ < No Issue

Same here on Mac AArch64, not fixed with B build from comment 33

Thank you for testing!

New round of builds coming up in:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156262

Assignee: nobody → hsivonen
Status: NEW → ASSIGNED

(In reply to Henri Sivonen (:hsivonen) from comment #40)

Thank you for testing!

New round of builds coming up in:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156262

build failed

(In reply to mix5003 from comment #41)

(In reply to Henri Sivonen (:hsivonen) from comment #40)

Thank you for testing!

New round of builds coming up in:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156262

build failed

This was an infrastructure problem. There are now green "B" items available if you expand the group that shows the failure.

Could folks who have been able to observe the Gmail/GDoc failure, please, try a build from comment 40 and see if Gmail/GDoc works?

(In reply to Henri Sivonen (:hsivonen) from comment #42)

Could folks who have been able to observe the Gmail/GDoc failure, please, try a build from comment 40 and see if Gmail/GDoc works?

Still broken for me.

(In reply to Henri Sivonen (:hsivonen) from comment #42)

Could folks who have been able to observe the Gmail/GDoc failure, please, try a build from comment 40 and see if Gmail/GDoc works?

https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156262 < did not fix issue

i am not sure is this has any code change. compare to other link it will has 1 commit for code change and 1 commit for "Try Chooser Enhanced"
but this link only has "Try Chooser Enhanced". i am not sure about this tool, but i just want to let you known just in case

Could folks who have been able to observe the Gmail/GDoc failure, please, try a build from comment 40 and see if Gmail/GDoc works?

Also broken for me on Mac AArch64.

Thanks. The cause in neither of the more obvious cases. I guess I'll have to generate test builds for every possible cause.

Builds coming up for restoring style computation to template contents owner, which I suppose I should have counted as the third obvious candidate:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156569

Builds coming up for restoring style computation to XHR:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156572

Builds coming up for restoring style computation to JS Document constructor:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156573

createHTMLDocument:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156574

createDocument (should have bundled with previous, but oh well):
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156576

XSLT:
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156577

(In reply to Henri Sivonen (:hsivonen) from comment #47)

createDocument (should have bundled with previous, but oh well):
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156576

for me in Linux, Only this build fixed this issue

(In reply to mix5003 from comment #48)

(In reply to Henri Sivonen (:hsivonen) from comment #47)

createDocument (should have bundled with previous, but oh well):
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156576

for me in Linux, Only this build fixed this issue

Same on macOS.

Is there a more efficient way to debug this, e.g. adding extra logging instead of having to test multiple builds?

(In reply to mix5003 from comment #48)

(In reply to Henri Sivonen (:hsivonen) from comment #47)

createDocument (should have bundled with previous, but oh well):
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=156576

for me in Linux, Only this build fixed this issue

Thank you!

(In reply to Francesco Lodolo [:flod] (OOO Oct 3–10) from comment #49)

Same on macOS.

Thank you!

Is there a more efficient way to debug this, e.g. adding extra logging instead of having to test multiple builds?

Unfortunately, I don't have ideas for a better way than this, since I have a thing to tune per document that makes them different but I have no idea what specific mechanism Google uses to observe the difference.

So I could repro this and debug it. I think this is the place where the behavior changes:

Native stack:

#0  RefPtr<mozilla::css::Loader>::get (this=0x7ff4f1ccefc0) at /home/emilio/src/moz/firefox/obj-debug/dist/include/mozilla/RefPtr.h:314
#1  RefPtr<mozilla::css::Loader>::operator mozilla::css::Loader*() const & (this=0x7ff4f1ccefc0) at /home/emilio/src/moz/firefox/obj-debug/dist/include/mozilla/RefPtr.h:327
#2  mozilla::dom::Document::GetCSSLoader (this=0x7ff4f1ccee00) at /home/emilio/src/moz/firefox/obj-debug/dist/include/mozilla/dom/Document.h:1724
#3  mozilla::dom::LinkStyle::DoUpdateStyleSheet (this=0x7ff4fe10aec8, aOldDocument=aOldDocument@entry=0x0, aOldShadowRoot=aOldShadowRoot@entry=0x0, aObserver=aObserver@entry=0x0, aForceUpdate=aForceUpdate@entry=mozilla::dom::LinkStyle::ForceUpdate::No)
    at /home/emilio/src/moz/firefox/dom/base/LinkStyle.cpp:262
#4  0x00007ff50930252c in mozilla::dom::LinkStyle::UpdateStyleSheetInternal (this=0x7ff4f1ccee00, aOldDocument=0x0, aOldShadowRoot=0x0, aForceUpdate=mozilla::dom::LinkStyle::ForceUpdate::No) at /home/emilio/src/moz/firefox/dom/base/LinkStyle.cpp:193
#5  mozilla::dom::LinkStyle::UpdateStyleSheetInternal (this=0x7ff4f1ccee00) at /home/emilio/src/moz/firefox/obj-debug/dist/include/mozilla/dom/LinkStyle.h:127
#6  mozilla::dom::LinkStyle::BindToTree()::$_0::operator()() const (this=<optimized out>) at /home/emilio/src/moz/firefox/dom/base/LinkStyle.cpp:222
#7  mozilla::detail::RunnableFunction<mozilla::dom::LinkStyle::BindToTree()::$_0>::Run (this=<optimized out>) at /home/emilio/src/moz/firefox/obj-debug/dist/include/nsThreadUtils.h:549
#8  0x00007ff509086e5f in nsContentUtils::RemoveScriptBlocker () at /home/emilio/src/moz/firefox/dom/base/nsContentUtils.cpp:6888
#9  0x00007ff509249796 in mozilla::dom::Document::EndUpdate (this=0x7ff4f1ccee00) at /home/emilio/src/moz/firefox/dom/base/Document.cpp:8575
#10 0x00007ff509432425 in mozAutoDocUpdate::~mozAutoDocUpdate (this=<optimized out>) at /home/emilio/src/moz/firefox/dom/base/mozAutoDocUpdate.h:34
#11 nsINode::ReplaceOrInsertBefore (this=0x7ff4fe129820, aReplace=<optimized out>, aNewChild=<optimized out>, aRefChild=<optimized out>, aMutationEffectOnScript=MutationEffectOnScript::DropTrustWorthiness, aError=...)
    at /home/emilio/src/moz/firefox/dom/base/nsINode.cpp:3161
#12 0x00007ff509a99a7a in nsINode::InsertBeforeInternal (this=0x7ff4fe129820, aNode=..., aChild=0x0, aMutationEffectOnScript=MutationEffectOnScript::DropTrustWorthiness, aError=...) at /home/emilio/src/moz/firefox/dom/base/nsINode.h:2391
#13 nsINode::AppendChildInternal (this=0x7ff4fe129820, aNode=..., aMutationEffectOnScript=MutationEffectOnScript::DropTrustWorthiness, aError=...) at /home/emilio/src/moz/firefox/dom/base/nsINode.h:2405
#14 nsINode::AppendChild (this=0x7ff4fe129820, aNode=..., aError=...) at /home/emilio/src/moz/firefox/dom/base/nsINode.h:2399
#15 mozilla::dom::Node_Binding::appendChild

JS Stack:

0 Nme(a = ".,, {\n        background-position: ,px ,px;\n      }", ""ita-icon-"", "0", "-14", "-17") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":2525:185]
1 _.Ume() ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":2526:349]
    this = [object Object]
2 _.Wme(a = ""es-t-i0-und"", b = "[object HTMLDocument]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":2526:102]
    this = [object Object]
3 _.Xme(a = ""es-t-i0-und"") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":2526:196]
    this = [object Object]
4 _.w.Wa() ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":2867:373]
    this = [object Object]
5 _.w.nn(a = "[object HTMLDivElement]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":1189:65]
    this = [object Object]
6 _.w.render(a = "[object HTMLDivElement]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":1188:393]
    this = [object Object]
7 Xoe(a = "[object Object]", b = "true", c = "[object HTMLDivElement]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":2614:260]
8 _.ste.prototype.Wa() ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":2884:126]
    this = [Anonymous]
9 _.mYb(a = "[Anonymous]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":1111:486]
    this = [object Object]
10 _.w.Fa() ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":1111:350]
    this = [Anonymous]
11 _.xJf.prototype.Wa() ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_1/m=a":1678:299]
    this = [Anonymous]
12 _.mYb(a = "[Anonymous]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":1111:486]
    this = [object Object]
13 _.w.Fa() ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":1111:350]
    this = [Anonymous]
14 WBf(a = "[Anonymous]", "[object Object]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_1/m=a":1451:205]
15 FBf(a = "[[Anonymous] type:mail:tl]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_1/m=a":1447:345]
16 ABf(a = "[[Anonymous] type:mail:tl]") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_1/m=a":1434:367]
17 _.IT(a = "[object Object]", b = "null", c = "0", d = "[object Object]", "0") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_1/m=a":1431:292]
    this = [[Anonymous] type:mail:tl]
18 WWi("[object Object]", "search@inbox", "0", [function], "0") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=a,b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_2/m=m,m_i,i20jfd,lKrWxc,hkjXJ,gYOl6d,HXLjIb,Ng2pOb,mlxMDe,fYVVAd,dwi4A,CGR0he,DL8jZe,xaQcye,J41knb,sLkgSb,eoq5Ke,oRmHt,E1P0kd,X3h7x,pE92lb,v2eEBc,lLhY8,UCS7w":5272:217]
    this = [[Anonymous] type:mail:tl]
19 XWi.prototype.hOa(a = "[object Object]", b = "search@inbox", c = "0", d = "[object Object]", e = "0") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=a,b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_2/m=m,m_i,i20jfd,lKrWxc,hkjXJ,gYOl6d,HXLjIb,Ng2pOb,mlxMDe,fYVVAd,dwi4A,CGR0he,DL8jZe,xaQcye,J41knb,sLkgSb,eoq5Ke,oRmHt,E1P0kd,X3h7x,pE92lb,v2eEBc,lLhY8,UCS7w":5272:435]
    this = [object Object]
20 BQi.prototype.hOa/</<("undefined") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/exm=a,b/ed=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_2/m=m,m_i,i20jfd,lKrWxc,hkjXJ,gYOl6d,HXLjIb,Ng2pOb,mlxMDe,fYVVAd,dwi4A,CGR0he,DL8jZe,xaQcye,J41knb,sLkgSb,eoq5Ke,oRmHt,E1P0kd,X3h7x,pE92lb,v2eEBc,lLhY8,UCS7w":5156:417]
21 Rmb/e.ha</e.Aa<(k = "undefined") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":559:479]
    this = null
22 $mb(a = "[object Object]", b = "2", c = "undefined") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":562:314]
23 Wmb(a = "[object Object]", b = "[object Object]", c = "2", d = "undefined") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":562:189]
24 _.dr.prototype.ftb() ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":562:57]
    this = [object Object]
25 I_a("undefined") ["https://mail.google.com/_/scs/mail-static/_/js/k=gmail.main.en_GB.3vBqyu-d_jc.es5.OL/am=qAEDDgAA8AHI_-QN_w4OGxAAgH8dH7_6wB-bcHDBEAADHiJgRKAMQB9QmSq8KTZ2ezXf_eAz2McfEgAQgCxHN-vBB0BuvUYAYAEAgGmMwDIMAAAAAAAAAAAAAAAAAAAAEI9GCAQ/d=1/im=1/dg=0/br=1/wt=1/rs=AHGWq9D2cpKaI1vgSIezuhRdOepFmGgkbg/cb=loaded_0/m=b":207:342]

So we refuse to create an inline stylesheet for this empty <style> element, and the JS code then proceeds to read the .sheet and assume it's non-null.

In fact I think I just found the problematic bit of JS:

Nme=function(a){for(var b=_.nq.apply(1,arguments),c=a[0],d=0;d<a.length-1;d++)c+=String(b[d]),c+=a[d+1];b=document.implementation.createHTMLDocument("");d=b.createElement(_.Qn);b.head.appendChild(d);b=d.sheet;b.insertRule(c,0);if(b.cssRules.length===1&&(c=b.cssRules[0],c instanceof CSSStyleRule))return _.bq(c.cssText.replace(/</g,"\\3C "))};

So it's I think doing some sort of CSS sanitization? _.Qn is style. So I guess reduced test-case would be:

test(function() {
  let doc = document.implementation.createHTMLDocument("");
  let style = doc.createElement("style");
  doc.head.appendChild(style)
  assert_true(!!style.sheet);
});

So, yeah, maybe we need to lazily initialize the CSS loader and co?

There's not a new round of shippable-config builds in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=158062

Do those work with Google Accounts that showed the failure earlier?

(In reply to Henri Sivonen (:hsivonen) from comment #52)

There's not a new round of shippable-config builds in
https://treeherder.mozilla.org/jobs?repo=try&landoCommitID=158062

Do those work with Google Accounts that showed the failure earlier?

This build fixed for me. (linux build)

This seems to work for me as well on macOS.

Thanks!

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

Attachment

General

Creator:
Created:
Updated:
Size: