958254 - B2G: Assertion during child startup: mozilla::dom::TabChild::HandlePossibleViewportChange

Status: RESOLVED FIXED

(Core :: Panning and Zooming, defect)

Description

[Gregor Wagner [:gwagner]]

With --enable-debug on current trunk when starting up a downloaded app.

Program received signal SIGSEGV, Segmentation fault.
0xb53b6210 in mozilla::dom::TabChild::HandlePossibleViewportChange (
    this=0xb31026e0) at ../../../dom/ipc/TabChild.cpp:593
593	  MOZ_ASSERT(NS_SUCCEEDED(rv));
(gdb) bt
#0  0xb53b6210 in mozilla::dom::TabChild::HandlePossibleViewportChange (
    this=0xb31026e0) at ../../../dom/ipc/TabChild.cpp:593
#1  0xb53b6710 in mozilla::dom::TabChild::Observe (this=0xb31026e0, 
    aSubject=<optimized out>, aTopic=<optimized out>, aData=0x0)
    at ../../../dom/ipc/TabChild.cpp:365
#2  0xb4cdc204 in nsObserverList::NotifyObservers (this=<optimized out>, 
    aSubject=0xb2668800, aTopic=0xb6680363 "before-first-paint", someData=0x0)
    at ../../../xpcom/ds/nsObserverList.cpp:96
#3  0xb4cdc7ec in NotifyObservers (someData=0x0, 
    aTopic=0xb6680363 "before-first-paint", aSubject=0xb2668800, 
    this=0xb3e8f640) at ../../../xpcom/ds/nsObserverService.cpp:302
#4  nsObserverService::NotifyObservers (this=0xb3e8f640, aSubject=0xb2668800, 
    aTopic=0xb6680363 "before-first-paint", someData=0x0)
    at ../../../xpcom/ds/nsObserverService.cpp:290
#5  0xb5975476 in nsBeforeFirstPaintDispatcher::Run (this=0xb213ae90)
    at ../../../layout/base/nsPresShell.cpp:507
#6  0xb564cdd4 in nsContentUtils::AddScriptRunner (aRunnable=0xb213ae90)
    at ../../../../content/base/src/nsContentUtils.cpp:4743
#7  0xb5977fb2 in UnsuppressAndInvalidate (this=0xb31b8880)

Seems like we hit this code line now because of http://hg.mozilla.org/mozilla-central/annotate/effa95701c83/dom/ipc/TabChild.cpp#l509

Comment 1

[Gregor Wagner [:gwagner]]

I don't know why we assert for rv when a function can clearly return an error.

Comment 2

[Botond Ballo [:botond]]

Attachment: bug958254.patch

Looks like that assertion is there by mistake. The next line even checks for NS_FAILED(rv).

Comment 3

[Anthony Jones (:ajones, :kentuckyfriedtakahe, :k17e)]

Comment on attachment 8358003 [details] [diff] [review]
bug958254.patch

Review of attachment 8358003 [details] [diff] [review]:
-----------------------------------------------------------------

I don't really know what it means for the assertion to fail. The code below just makes it fail more gracefully and makes it compile with assertions off. However GetFirstPaint() does to fail so we should just remove the assertion.

Comment 5

[Botond Ballo [:botond]]

https://hg.mozilla.org/integration/b2g-inbound/rev/21715f6a3ba2

Comment 6

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/mozilla-central/rev/21715f6a3ba2