839753 - Setting expando on AudioContext destination causes bad things

Status: RESOLVED FIXED

(Core :: Audio/Video, defect)

Description

[Jesse Ruderman]

Attachment: testcase

With
  user_pref("media.webaudio.enabled", true);


###!!! ASSERTION: Cycle collection participant didn't traverse to preserved wrapper! This will probably crash.: 'callback.mFound', file content/base/src/nsContentUtils.cpp, line 6128

###!!! ASSERTION: Cycle collection participant didn't trace preserved wrapper! This will probably crash.: 'callback.mFound', file content/base/src/nsContentUtils.cpp, line 6134

Assertion failure: !PreservingWrapper() (Clearing a preserved wrapper!), at ../../../dom/base/nsWrapperCache.h:112

Comment 1

[Andrew McCreight [:mccr8]]

The QI for AudioDestinationNode looks wrong to me.  I don't see how it is going to produce the CC participant.  I think this is another thing that would be caught by bug 614238.

This is a regression from the bug landed as bug 804837, but I'm not sure what the real bug number was.

Comment 2

[Andrew McCreight [:mccr8]]

I can fix this on Monday, anybody else should feel free to fix it earlier if they want to.

Comment 3

[Andrew McCreight [:mccr8]]

Attachment: untested WIP

Comment 4

[Andrew McCreight [:mccr8]]

Attachment: tested locally, added test

https://tbpl.mozilla.org/?tree=Try&rev=1c29d48852eb

Comment 5

[Olli Pettay [:smaug][[e-mail]]]

Comment on attachment 712941 [details] [diff] [review]
tested locally, added test

We can do this, but I wonder why AudioNode isn't wrapperCached.

Comment 6

[Andrew McCreight [:mccr8]]

It used to be, but the patch that broke this intentionally made it so that only destination is.  I'm not sure what the actual bug was where that happened, so I'm not sure why exactly.

Comment 7

[Robert O'Callahan (:roc) (email my personal email if necessary)]

That is intentional (and correct). We need to allow the JS binding to be GCed so we can determine that there's no longer any way the AudioNode can produce output in the future (then we can optimize it away).

Comment 8

[Ryan VanderMeulen [:RyanVM]]

https://hg.mozilla.org/integration/mozilla-inbound/rev/9814c03a6e2e

Comment 9

[Ed Morley [:emorley]]

https://hg.mozilla.org/mozilla-central/rev/9814c03a6e2e