1792797 - regression break Asana / Slack floating button

Status: NEW

(Core :: Panning and Zooming, defect, P3)

Description

[Irvin (MozTW)]

Attachment: button normally shows on current beta.png

On nightly, I got a regression issue for asana to not show a floating button on their "Inbox" page, which makes it not usable. (I'm not sure how to reproduce it without an Asana account.)

According to mozregression the regression window is [9695f66a, 7d85a597] (2 builds)

Bug 1778120 - Add the animation id to the webrender stacking context. r=botond
Differential Revision: https://phabricator.services.mozilla.com/D158035

Comment 1

[Irvin (MozTW)]

Attachment: on broken build the floating button is not showing.png

Comment 2

[Irvin (MozTW)]

Attachment: button not shows when close asana sidebar.mov

Comment 3

[Irvin (MozTW)]

Attachment: the html of the button -1

Comment 4

[Irvin (MozTW)]

Attachment: the html of the button -2

i don't find any special codes on these flex button

Comment 5

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

Set release status flags based on info from the regressing bug 1778120

:dlrobertson, since you are the author of the regressor, bug 1778120, could you take a look? Also, could you set the severity field?

For more information, please visit auto_nag documentation.

Comment 6

[Irvin (MozTW)]

Attachment: slack "mark as read" button on "Unreads" is not showing

found that the same regression break Slack's floating button on "Unreads" tab as well.

Comment 7

[Irvin (MozTW)]

Attachment: normal slack button at beta

Comment 8

[Irvin (MozTW)]

found that the same regression caused Slack's floating button to also not work (similar behavior).

Comment 9

[Irvin (MozTW)]

mozregression gui gave the same regression window

2022-09-29T20:04:46.965000: INFO : Narrowed integration regression window from [9695f66a, aa1973a8] (3 builds) to [9695f66a, 7d85a597] (2 builds) (~1 steps left)
2022-09-29T20:04:46.989000: DEBUG : Starting merge handling...
2022-09-29T20:04:46.990000: DEBUG : Using url: https://hg.mozilla.org/integration/autoland/json-pushes?changeset=7d85a597c9aec41cbde40d01172539cf90351156&full=1
2022-09-29T20:04:46.990000: DEBUG : redo: attempt 1/3
2022-09-29T20:04:46.990000: DEBUG : redo: retry: calling _default_get with args: ('https://hg.mozilla.org/integration/autoland/json-pushes?changeset=7d85a597c9aec41cbde40d01172539cf90351156&full=1',), kwargs: {}, attempt #1
2022-09-29T20:04:47.194000: DEBUG : urllib3.connectionpool: Resetting dropped connection: hg.mozilla.org
2022-09-29T20:04:48.747000: DEBUG : urllib3.connectionpool: https://hg.mozilla.org:443 "GET /integration/autoland/json-pushes?changeset=7d85a597c9aec41cbde40d01172539cf90351156&full=1 HTTP/1.1" 200 None
2022-09-29T20:04:48.762000: DEBUG : Found commit message:
Bug 1778120 - Add the animation id to the webrender stacking context. r=botond

On desktop add the animation id to the webrender stacking context.

Differential Revision: https://phabricator.services.mozilla.com/D158035

2022-09-29T20:04:48.762000: DEBUG : Did not find a branch, checking all integration branches
2022-09-29T20:04:48.771000: INFO : The bisection is done.
2022-09-29T20:04:48.779000: INFO : Stopped

Comment 10

[Dan Robertson (:dlrobertson)]

(In reply to Irvin (MozTW) from comment #9)

mozregression gui gave the same regression window

Thanks so much for the bug report! This might be similar to bug 1792830.

Comment 11

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

The bug is marked as tracked for firefox107 (nightly). We have limited time to fix this, the soft freeze is in 13 days. However, the bug still isn't assigned.

:fgriffith, could you please find an assignee 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 auto_nag documentation.

Comment 12

[Botond Ballo [:botond]]

Let's assign to Dan for now, we can back out the regressing patch if we can't get a fix in by the soft freeze.

Comment 14

[Ted Campbell [:tcampbell]]

It looks like the assert is getting us a number of concrete examples to investigate now. Should we back out the original patch while this is looked at to avoid hiding other issues in Nightly? Slack is still pretty consistently broken for me (missing buttons).

Comment 15

[Dan Robertson (:dlrobertson)]

(In reply to Ted Campbell [:tcampbell] from comment #14)

It looks like the assert is getting us a number of concrete examples to investigate now. Should we back out the original patch while this is looked at to avoid hiding other issues in Nightly? Slack is still pretty consistently broken for me (missing buttons).

Yeah, that works for me. The change that caused this regression doesn't introduce any new features, so we're not losing anything by backing it out. Some of the regressions introduced by the change may still exist on Android after the backout, so I'll still continue to work on debugging this.

Comment 16

[Donal Meehan [:dmeehan]]

Setting 107 as disabled, regressor Bug 1778120 was backed out of central

Comment 17

[Dan Robertson (:dlrobertson)]

From Layout chat (pasting here to make sure it doesn't get lost):

emilio: https://searchfox.org/mozilla-central/rev/ffa4d00965c5281def6d3ddcbcdf6259d38c9b9a/gfx/layers/wr/ClipManager.cpp#172-187 / https://searchfox.org/mozilla-central/rev/ffa4d00965c5281def6d3ddcbcdf6259d38c9b9a/gfx/layers/wr/ClipManager.cpp#206-215 are the code I was thinking of btw.
emilio: that's what maps from Gecko display items / clips / scroll roots to WR stacking context clip chains

These look like the causes of us optimizing out the sticky position content in some cases.

Comment 18

[Dan Robertson (:dlrobertson)]

Bumping to S3, since the regressing patch was backed out.