1540551 - Perma Windows 7 DevEdition open-features-non-integer-innerheight.html | features "innerheight=405.5" should set "height=405" starting with Gecko 68

Status: RESOLVED WONTFIX

(Firefox :: Theme, defect, P5)

Description

[Raul Gurzau (:RaulG)]

[Tracking Requested - why for this release]:

Central as beta simulation: https://treeherder.mozilla.org/#/jobs?repo=try&resultStatus=testfailed%2Cbusted%2Cexception&revision=7d697d08b5569edf179d3f1d700a5506ace0cd01&selectedJob=237220993&searchStr=windows%2C7%2Cdevedition%2Copt%2Cweb%2Cplatform%2Ctests%2Cwith%2Ce10s%2Ctest-windows7-32-devedition%2Fopt-web-platform-tests-e10s-9%2Cw-e10s%28wpt9%29

Log link: https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=237220993&repo=try&lineNumber=5538

Log snippet:

15:21:18 INFO - PID 1992 | JavaScript warning: chrome://browser/content/tabbrowser.js, line 4834: Array.unshift is deprecated; use Array.prototype.unshift instead
15:21:18 INFO - PID 1992 | JavaScript warning: chrome://global/content/bindings/tabbox.xml, line 162: Array.forEach is deprecated; use Array.prototype.forEach instead
15:21:18 INFO - PID 1992 | JavaScript warning: chrome://browser/content/browser-ctrlTab.js, line 568: Array.filter is deprecated; use Array.prototype.filter instead
15:21:18 INFO - PID 1992 | JavaScript warning: chrome://browser/content/tabbrowser.js, line 4834: Array.unshift is deprecated; use Array.prototype.unshift instead
15:21:33 INFO -
15:21:33 INFO - TEST-PASS | /html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-non-integer-innerheight.html | features "innerheight=/404" should NOT set "height=404"
15:21:33 INFO - TEST-PASS | /html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-non-integer-innerheight.html | features "innerheight=_404" should NOT set "height=404"
15:21:33 INFO - TEST-PASS | /html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-non-integer-innerheight.html | features "innerheight=L404" should NOT set "height=404"
15:21:33 INFO - TEST-UNEXPECTED-FAIL | /html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-non-integer-innerheight.html | features "innerheight=405.5" should set "height=405" - assert_equals: "top=0,left=0,width=401,innerheight=405.5 value after first non-digit will be ignored" expected 405 but got 400
15:21:33 INFO - runWindowTests/</</<@http://web-platform.test:8000/html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-non-integer-innerheight.html:68:9
15:21:33 INFO - Test.prototype.step@http://web-platform.test:8000/resources/testharness.js:1587:25
15:21:33 INFO - Test.prototype.step_func_done/<@http://web-platform.test:8000/resources/testharness.js:1627:32
15:21:33 INFO - PrefixedMessage.prototype.onMessage/<@http://web-platform.test:8000/common/PrefixedPostMessage.js:53:12
15:21:33 INFO -
15:21:33 INFO - TEST-UNEXPECTED-FAIL | /html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-non-integer-innerheight.html | features "innerheight=405.32" should set "height=405" - assert_equals: "top=0,left=0,width=401,innerheight=405.32 value after first non-digit will be ignored" expected 405 but got 400
15:21:33 INFO - runWindowTests/</</<@http://web-platform.test:8000/html/browsers/the-window-object/apis-for-creating-and-navigating-browsing-contexts-by-name/open-features-non-integer-innerheight.html:68:9
15:21:33 INFO - Test.prototype.step@http://web-platform.test:8000/resources/testharness.js:1587:25
15:21:33 INFO - Test.prototype.step_func_done/<@http://web-platform.test:8000/resources/testharness.js:1627:32
15:21:33 INFO - PrefixedMessage.prototype.onMessage/<@http://web-platform.test:8000/common/PrefixedPostMessage.js:53:12

Comment 1

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Is this also from bug 1525762? Changelog since the previous beta simulation: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=5f519306c8bdf77c3236abcb626eacfe44c5e1f0&tochange=c06dfc552c647a6ce96f35cd84c32a589dc85608
DevEdition makes it possible.

Comment 2

[Kris Maglione [:kmag]]

Plausible. Let's see what happens after bug 1540856 lands.

Comment 3

[Cristina Coroiu [:ccoroiu]]

Failed in today's beta sim, after bug 1540856 was merged to central.
https://treeherder.mozilla.org/#/jobs?repo=try&resultStatus=testfailed%2Cbusted%2Cexception%2Cretry%2Cusercancel%2Crunnable&revision=d71d7b6e68ea9dbc465c9da13913c331c22dad4c&selectedJob=238146053

Comment 4

[Hsin-Yi Tsai (she/her) [:hsinyi]]

As this is a new regression and causing perma failure, P1.
(Sorry for highering the priority, Alphan)

Comment 5

[Kris Maglione [:kmag]]

Same scenario as bug 1540566 comment 2. Web platform tests shouldn't care what browser theme is applied. If this test fails when the dev edition is applied, either the test is broken or our behavior is broken when the dev edition theme is applied. It's up to the owner of the test do decide which.

Comment 6

[Hsin-Yi Tsai (she/her) [:hsinyi]]

Hi Alphan, according to comment 5, could you please help investigate how this test is doing with dev edition theme and what we should do for 68 before the beta merge? Thank you very much.

Comment 7

[Alphan Chen [:alchen]]

The test is using window.open() with featureString to set "innerheight".
It should ignore the value after first non-digit.
In the failure testcase, we should set "405" rather than 405.5.
However, the value we get here is 400.

Trying to figure the difference....

Comment 8

[Alphan Chen [:alchen]]

There is a coincidence between bug 1540566 and this bug(bug 1540551).
The difference of target height and actual height is 5.

For bug 1540551,
expected 405 but got 400

For bug 1540566,
height (195) should be around twice what was requested (100)
--> It means expected 200(=100*2) but got 195

Comment 9

[Alphan Chen [:alchen]]

According to Comment 5 and Comment 8, there is difference between whether dev theme is applied or not.
The symptom is happened on Window 7.

From DOM side, I have no idea about the difference.
Could you input something about what will happen if we apply dev theme?
Or suggest someone who is familiar with dev theme?

Comment 10

[Patrick Brosset <:pbro>]

Looking at the test, I have no immediate ideas as to why changing the theme would impact it.
The theme that applies to the DevTools toolbox is different than the Firefox dark theme (which is, I think, a light-weight theme defined here). There is nothing in common between the 2 except that the Firefox dark theme originally was inspired by the devtools dark theme which, at some stage, was applied to the DevEdition release.
I think others than me would better be able to help. Sorry I don't know much about browser themes or WPT.
I'm also going to clear needinfo for Joe since he doesn't currently work on devtools.

Comment 11

[Kris Maglione [:kmag]]

This is about the Firefox dark theme, which is the default theme for dev edition builds, and therefore is applied for (plain and chrome) mochitests and web platform tests run in those builds.

Comment 12

[Henri Sivonen (:hsivonen)]

Works with a build from the simulation run when executing on Windows 10 1803 and running the test from w3c-test.org.

Comment 13

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

The issue reproduces here (OS: Windows 8.1, artifact build set to 32 bit) even before the bug got filed and not only with DevEdition:

assert_equals: "top=0,left=0,width=401,innerheight=405.5 value after first non-digit will be ignored" expected 405 but got 404

Comment 14

[Intermittent Failures Robot]

1 failures in 3025 pushes (0.0 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• try: 1

Platform breakdown:

• windows7-32-devedition: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1540551&startday=2019-04-08&endday=2019-04-14&tree=all

Comment 15

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

The first merge of central to beta is less than 2 weeks away. Please prioritize this.

Comment 16

[Hsin-Yi Tsai (she/her) [:hsinyi]]

Henri is helping with this. Do you have any findings so far?

Comment 17

[Henri Sivonen (:hsivonen)]

Not yet due to focusing on a couple of other priority items first. I'll give this issue a better look now.

Comment 18

[Henri Sivonen (:hsivonen)]

https://treeherder.mozilla.org/#/jobs?repo=try&revision=93b45a30cc3fc49877cade8c65fbe8382155e85f

Comment 19

[Henri Sivonen (:hsivonen)]

Attachment: Normal theme, no tab in JS-opened window

Comment 20

[Henri Sivonen (:hsivonen)]

Attachment: Dev edition theme shows a tab for JS-opened window

Comment 21

[Henri Sivonen (:hsivonen)]

I can't reproduce the problem when downloading the 32-bit try build and running it locally on 64-bit Windows 7.

In any case, this does not seem to be a DOM API issue. My hypothesis is that there is a one-pixel try-server-only layout glitch with the dev edition theme. I think it's not worthwhile to track down the layout glitch. Rather, I think the best avenue for fixing this would be making the dev edition theme not display a tab when opening a window with these options. See the attachments: The normal theme does not show a tab. The dev edition shows a tab, and unlike with normal browser windows, the tab is not painted into the title bar.

Comment 23

[Henri Sivonen (:hsivonen)]

layout glitch with the dev edition theme.

For clarity: When the new window opens, we spin the event loop and load the XUL UI. We then ask layout what size the XUL browser area ended up being. What gets laid out with the dev edition theme is substantially different from what gets laid out with the normal theme.

Comment 24

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Tested on a Windows 7 with Nightly (= Beta) and DevEdition from today's beta simulation and couldn't reproduce the issue (all tests passed). Even passed for DevEdition after changing the screen resolution from 1600x900 to 1280x768.

How shall this be fixed? Annotations of the failing tests as expected or disable them?

Comment 25

[Kris Maglione [:kmag]]

(In reply to Henri Sivonen (:hsivonen) from comment #23)

layout glitch with the dev edition theme.

For clarity: When the new window opens, we spin the event loop and load the XUL UI. We then ask layout what size the XUL browser area ended up being. What gets laid out with the dev edition theme is substantially different from what gets laid out with the normal theme.

That does sound like something that should be fixed...

Does the tab disappear after the window is painted, or does it stay visible?

Comment 26

[Henri Sivonen (:hsivonen)]

Attachment: Result of exact same window.open() call in Nightly and Dev Edition

(In reply to Kris Maglione [:kmag] from comment #25)

Does the tab disappear after the window is painted, or does it stay visible?

It appears right away (and has to appear when the window is still invisible during window creation, because otherwise the failures would be off by more than one pixel). The attached screenshot show the result of the exact same window.open() call in Nightly and in Dev Edition built from trunk.

It looks like the Dev Edition theme lacks special case for window.open()ed windows, but the native title bar is still window.open()-like. (I.e. window.open()ed windows have a native title bar in Dev Edition, too, unlike normal windows.)

Comment 28

[Intermittent Failures Robot]

4 failures in 3031 pushes (0.001 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• try: 4

Platform breakdown:

• windows7-32-devedition: 4

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1540551&startday=2019-04-22&endday=2019-04-28&tree=all

Comment 29

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Attachment: Bug 1540551 - Disable window height tests failing on Windows 7 DevEdition

Comment 30

[Intermittent Failures Robot]

2 failures in 4131 pushes (0.0 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• try: 2

Platform breakdown:

• windows7-32-devedition: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1540551&startday=2019-04-29&endday=2019-05-05&tree=all

Comment 31

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/f92e2f3f7fa9
Disable window height tests failing on Windows 7 DevEdition r=hsivonen

Comment 32

[Oana Pop-Rus]

https://hg.mozilla.org/mozilla-central/rev/f92e2f3f7fa9

Comment 33

[Andreea Pavel [:apavel]]

No longer occurs on recent beta sims

https://treeherder.mozilla.org/#/jobs?repo=try&resultStatus=success%2Ctestfailed%2Cbusted%2Cexception&revision=f93bb605459b4db524f5ecf50596ce01e85bea6e&searchStr=Windows%2C7%2CDevEdition%2Copt&group_state=expanded&selectedJob=245534804

Comment 34

[Intermittent Failures Robot]

1 failures in 4409 pushes (0.0 failures/push) were associated with this bug in the last 7 days.

Repository breakdown:

• mozilla-inbound: 1

Platform breakdown:

• macosx1014-64: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1540551&startday=2019-06-24&endday=2019-06-30&tree=all

Comment 35

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

Bugbug thinks this bug is a regression, but please revert this change in case of error.

Comment 36

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

The leave-open keyword is there and there is no activity for 6 months.
:dao, maybe it's time to close this bug?
For more information, please visit BugBot documentation.

Comment 37

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Support for Windows 7 has been dropped except for ESR 115.