1282649 - Shutdown assertion failure in mozilla::dom::BlobParent::RemoteBlobImpl::Destroy with SharedWorker

Status: RESOLVED DUPLICATE of bug 1353629

(Core :: DOM: Workers, defect)

Description

[Jesse Ruderman]

Attachment: testcase

During shutdown:

###!!! ASSERTION: Failed Dispatch after xpcom-shutdown-threads: 'false', file xpcom/threads/nsThread.cpp, line 726

Assertion failure: (((bool)(__builtin_expect(!!(!NS_FAILED_impl(mActorTarget->Dispatch(destroyRunnable, nsIEventTarget::DISPATCH_NORMAL))), 1)))), at dom/ipc/Blob.cpp:2575

That ugly expression comes from a macro expansion of:

-> 2574     MOZ_ALWAYS_SUCCEEDS(mActorTarget->Dispatch(destroyRunnable,
   2575                                                NS_DISPATCH_NORMAL));

Comment 1

[Jesse Ruderman]

Attachment: stack

Comment 2

[Andrew Overholt [:overholt]]

baku, can you take a look?

Comment 3

[Andrea Marchesini [:baku]]

Attachment: crash.patch

Comment 4

[Andrea Marchesini [:baku]]

This patch doesn't remove the warning. But we should not crash if the dispatching of a runnable fails.

Comment 5

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

Dispatching to the main thread should never be able to fail.  Really I think the bug is that a blob is alive this late in shutdown ...

Comment 6

[Andrea Marchesini [:baku]]

Right. It makes sense. At this point I really would like to reproduce this issue locally. Jesse, how complex is it to do it?
I tried with your test, but I cannot reproduce it.

Comment 7

[Kyle Huey (Exited; not receiving bugmail, old account, do not use)]

Comment on attachment 8766695 [details] [diff] [review]
crash.patch

Anyways, I don't think this is fixing the root cause.

Comment 8

[Ryan VanderMeulen [:RyanVM]]

I can't reproduce this locally on Ubuntu or Win10 either. Should we just call this WFM?

Comment 9

[Andrea Marchesini [:baku]]

Blob handling has been completely rewritten. Let's mark as duplicate.