1498282 - On Linux, resistfingerprinting content resize tests can get earlier resize event and get confused about final window sizing when not loading initial about:blank

Status: RESOLVED WORKSFORME

(Firefox :: General, defect, P2)

Description

[:Gijs (out for now; he/him)]

Spotted in https://treeherder.mozilla.org/#/jobs?repo=try&selectedJob=203135076&revision=52f0d7b35760ac92e7eb677bc99645b697a86b3b

TEST-UNEXPECTED-FAIL | browser/components/resistfingerprinting/test/browser/browser_roundedWindow_windowSetting_mid_inner.js | The screen.height has a correct rounded value - 100 == 600 -
TEST-UNEXPECTED-FAIL | browser/components/resistfingerprinting/test/browser/browser_roundedWindow_windowSetting_mid_inner.js | The window.innerHeight has a correct rounded value - 100 == 600 -
Return code: 1

Because we initially open the window at 100x100 and then set window.inner/outerWidth/Height with a content task, it's possible the resize handler triggers for the first size of the window (on Linux, where window sizing + focusing is async), and that causes the test to fail.

Similar to the changes made in bug 1489700, we can just wait for the correct resize event and that'll make the test pass.

Comment 1

[:Gijs (out for now; he/him)]

Attachment: Bug 1498282 - make resistfingerprinting window sizing tests wait for the 'right' resize event, r?tjr

Comment 2

[:Gijs (out for now; he/him)]

Not seeing issues here anymore atm, can reopen if this reappears.