W3C home > Mailing lists > Public > public-audio@w3.org > April to June 2012

Re: Potential WebKit bug when using JavaScriptAudioNode

From: Jussi Kalliokoski <jussi.kalliokoski@gmail.com>
Date: Thu, 24 May 2012 00:32:56 +0300
Message-ID: <CAJhzemWd=YdU8+eX-zT2ddwin6FEwVvbdfnSJZeON-B5Z+dkFA@mail.gmail.com>
To: Colin Sullivan <colin@colin-sullivan.net>
Cc: public-audio@w3.org
I think this is because of garbage collection and it could be a bug either
way. You have no static references (e.g. global scope) to the AudioContext
or the nodes. That means that either the same buffer starting to repeat is
the intended behavior and the timeout just gives the nodes a bit more
lifetime, but the same behavior should occur after the timeout has been
reached. The other option is that the intended behavior is that connected
nodes aren't garbage collected. I'm not sure what is the current wanted
behavior. I also think this might be the same bug as this:
http://code.google.com/p/chromium/issues/detail?id=82795

Cheers,
Jussi

On Wed, May 23, 2012 at 8:14 PM, Colin Sullivan <colin@colin-sullivan.net>wrote:

> Hey all,
>
> I hope this list is an appropriate place for this discussion, hoping to
> get feedback on this issue and determine if it is actually a bug or if I am
> just interpreting the spec incorrectly.
>
> The bug I filed is here:
> https://code.google.com/p/chromium/issues/detail?id=129306
>
> The test page:
> https://ccrma.stanford.edu/~colinsul/audioOnCanvas/examples/jsaudionode_bug/
>
> My JavaScriptAudioNode's `onaudioprocess` method consistently ceases to
> fire, and it sounds like the previous buffer repeats forever.  This only
> happens for me if I don't wait a few seconds before retrieving and playing
> the audio, but I have not tested on many other systems/browsers.
> ----------
> Colin
>
>
Received on Wednesday, 23 May 2012 21:33:26 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 23 May 2012 21:33:26 GMT