809671 - "Assertion failure: count.numLocks >= count.numHidden" switching tabs in full-screen mode

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Jesse Ruderman]

Attachment: stack trace

In a debug build of Firefox for Mac:

1. Full screen a page (Cmd+Shift+F)
2. Open a new tab (Cmd+T)
3. Close that tab (Cmd+W) or switch back to the first tab

Assertion failure: count.numLocks >= count.numHidden, at hal/HalWakeLock.cpp:118

Based on when I started seeing this, I'm guessing it's a regression from bug 805017.

Comment 1

[Kan-Ru Chen [:kanru] (UTC+9)]

Attachment: Don't change wake lock state if visibility doesn't change. v1

We assumed that when visibility change event occurs, the visibility state is different. I should have asserted that. I don't know how could this happen though.

Comment 2

[Justin Lebar (not reading bugmail)]

Comment on attachment 679538 [details] [diff] [review]
Don't change wake lock state if visibility doesn't change. v1

Okay!

Comment 4

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/191155ed17b0

Comment 5

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/191155ed17b0

Comment 7

[Kan-Ru Chen [:kanru] (UTC+9)]

Nominated because this will cause wrong wakelock count which will affect power usage.

Comment 8

[Gene Lian (I already quit Mozilla)]

+1. Power consumption is a big issue. We need to prevent our phones from exhausting the battery quickly.

Comment 9

[Justin Lebar (not reading bugmail)]

I can confirm that this missed the b2g18 train.

I doubt we can take this on tef now.  That's a bummer, but missing bugs like this is the risk we knew we were taking when we adopted a many-branches model.

Comment 10

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-b2g18/rev/f136570bd27c

Comment 11

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/releases/mozilla-b2g18_v1_1_0_hd/rev/f136570bd27c