Open Bug 1520284 Opened 3 years ago Updated 2 years ago

"WebDriver:MaximizeWindow" times out after 5000ms when running in headless mode

Categories

(Testing :: Marionette, defect, P3)

defect

Tracking

(firefox65 wontfix, firefox66 wontfix, firefox67 wontfix, firefox68 fix-optional, firefox69 fix-optional)

Tracking Status
firefox65 --- wontfix
firefox66 --- wontfix
firefox67 --- wontfix
firefox68 --- fix-optional
firefox69 --- fix-optional

People

(Reporter: whimboo, Unassigned)

References

(Depends on 1 open bug, Regression)

Details

(Keywords: regression)

I can see timeouts being reported in the "WebDriver:Maximize" command when running tests in headless mode. It might not happen the first time, but using the --run-until-failure argument clearly shows the following each time after a while:

1547583548188 Marionette DEBUG 10 -> [0,3,"WebDriver:MaximizeWindow",{}]
[..]
1547583552691 Marionette WARN TimedPromise timed out after 4500 ms: stacktrace:
bail@chrome://marionette/content/sync.js:224:64

I assume it's something similar with waiting for sizemodechange events.

It's a regression from bug 1492499, and sadly we don't raise an exception but hide it. But I will mark it blocking the meta bug 1515867, which tracks the regressions.

That is indeed because of waiting for the sizemodechange event, which is not happening in headless mode. Checking mochitests they all use resize events to wait for the window to be maximized.

Given that we have to get those changes in as soon as possible to be able to uplift to beta for the 65 release, I will do it as part of my patch on bug 1520302. It's actually blocked on this.

Depends on: 1520302

With bug 1521527 landed the timeout has been increased to 5s. Adjusting the summary to reflect reality.

Depends on: 1521527
Summary: "WebDriver:MaximizeWindow" times out after 4500ms when running in headless mode → "WebDriver:MaximizeWindow" times out after 5000ms when running in headless mode
Priority: -- → P3
No longer blocks: 1492499
Regressed by: 1492499

Since the team has already triaged this and it's P3, I don't think that the regression triage meeting needs to keep covering this.

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