1232888 - Assertion failure: !window || window->type == NPWindowTypeWindow on Windows on mozilla-beta after the merge of Gecko 44

Status: RESOLVED FIXED

(Core Graveyard :: Plug-ins, defect)

Description

[Phil Ringnalda (:philor)]

Blocking reopening mozilla-beta, thus blocking the first beta, thus giving ritu ulcers.

For a nice clean assertion failure, https://treeherder.mozilla.org/logviewer.html#?job_id=685470&repo=mozilla-beta

17:45:26     INFO -  Assertion failure: !window || window->type == NPWindowTypeWindow, at c:/builds/moz2_slave/m-beta-w64-d-00000000000000000/build/layout/generic/nsPluginFrame.cpp:1564
17:45:40     INFO -  #01: mozilla::FrameLayerBuilder::PaintItems(nsTArray<mozilla::FrameLayerBuilder::ClippedDisplayItem> &,mozilla::gfx::IntRectTyped<mozilla::gfx::UnknownUnits> const &,gfxContext *,nsRenderingContext *,nsDisplayListBuilder *,nsPresContext *,mozilla::gfx::IntPointTyped<mozilla::gfx::UnknownUnits> const &,float,float,int) [layout/base/FrameLayerBuilder.cpp:5656]
17:45:40     INFO -  #02: mozilla::FrameLayerBuilder::DrawPaintedLayer(mozilla::layers::PaintedLayer *,gfxContext *,nsIntRegion const &,nsIntRegion const &,mozilla::layers::DrawRegionClip,nsIntRegion const &,void *) [layout/base/FrameLayerBuilder.cpp:5815]
17:45:40     INFO -  #03: mozilla::layers::ClientPaintedLayer::PaintThebes() [gfx/layers/client/ClientPaintedLayer.cpp:95]
17:45:40     INFO -  #04: mozilla::layers::ClientPaintedLayer::RenderLayerWithReadback(mozilla::layers::ReadbackProcessor *) [gfx/layers/client/ClientPaintedLayer.cpp:142]
17:45:40     INFO -  #05: mozilla::layers::ClientContainerLayer::RenderLayer() [gfx/layers/client/ClientContainerLayer.h:68]
17:45:40     INFO -  #06: mozilla::layers::ClientContainerLayer::RenderLayer() [gfx/layers/client/ClientContainerLayer.h:68]
17:45:40     INFO -  #07: mozilla::layers::ClientLayerManager::EndTransactionInternal(void (*)(mozilla::layers::PaintedLayer *,gfxContext *,nsIntRegion const &,nsIntRegion const &,mozilla::layers::DrawRegionClip,nsIntRegion const &,void *),void *,mozilla::layers::LayerManager::EndTransactionFlags) [gfx/layers/client/ClientLayerManager.cpp:282]
17:45:40     INFO -  #08: mozilla::layers::ClientLayerManager::EndTransaction(void (*)(mozilla::layers::PaintedLayer *,gfxContext *,nsIntRegion const &,nsIntRegion const &,mozilla::layers::DrawRegionClip,nsIntRegion const &,void *),void *,mozilla::layers::LayerManager::EndTransactionFlags) [gfx/layers/client/ClientLayerManager.cpp:325]
17:45:40     INFO -  #09: nsDisplayList::PaintRoot(nsDisplayListBuilder *,nsRenderingContext *,unsigned int) [layout/base/nsDisplayList.cpp:1693]
17:45:40     INFO -  #10: nsLayoutUtils::PaintFrame(nsRenderingContext *,nsIFrame *,nsRegion const &,unsigned int,unsigned int) [layout/base/nsLayoutUtils.cpp:3383]

which is the assertion dvander added in bug 1218688 while removing unused synchronous plugin drawing paths. There also seem to be some scarier assertion failures in crashtests, https://treeherder.mozilla.org/logviewer.html#?job_id=685349&repo=mozilla-beta

Possibly significant, the Win8 crashtest run, https://treeherder.mozilla.org/logviewer.html#?job_id=685465&repo=mozilla-beta, hit what I think is roughly the assertion "!mozilla::ipc::MessageChannel::IsPumpingMessages() || mozilla::ipc::SuppressedNeuteringRegion::IsNeuteringSuppressed()" for which last go round we disabled asyncInit in bug 1220010, though I've no idea whether we should do that again, or whether doing so will keep us away from hitting dvander's assertion.

Comment 1

[Phil Ringnalda (:philor)]

Indeed, the answers to all my questions were "yes" - we should disable asyncInit again, more than just crashtests were hitting that assertion, and when we did disable it, all the things that were hitting dvander's assertion stopped doing so.

No longer a release blocker for this release, and if I can manage to remember until (if) aurora reopens and land the disabling of asyncInit there and on the trunk, not a release blocker for future releases either.