[Webkit-unassigned] [Bug 247209] LayoutTests follow system power saving settings by default, video tests will time out when the computer is in power saving mode

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=247209

--- Comment #1 from Michael Catanzaro <mcatanzaro at gnome.org> ---
(In reply to Alicia Boya García from comment #0)
> This is even written explicitly in Internals::resetToConsistentState(), who
> calls page.setLowPowerModeEnabledOverrideForTesting(std::nullopt); -- The
> nullopt means "follow the operating system". Why? I would argue that's not a
> sane default.

I agree. Let's just change it.

> Also, tangentially: as of writing, Internals::resetToConsistentState() is
> only called from InjectedBundlePage::resetAfterTest(). I could confirm this
> indeed means it only gets called before the second test in the same worker.
> This is risky, as any divergence between the values set in
> Internals::resetToConsistentState() and those WebKit starts with will apply
> differently to the first test in a given worker. This is a very risky
> footgun, especially considering the many regular random flakes we have in
> EWS. Why do we do this? Does it take so much CPU time to reset these values
> that is worth it having flakes that depend solely on (largely random) worker
> test dispatch?

That's surely a mistake. We should fix that too. Good catch.

-- 
You are receiving this mail because:
You are the assignee for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.webkit.org/pipermail/webkit-unassigned/attachments/20221029/0b951042/attachment-0001.htm>