I saw it fixed on the manual audio tests only. Since it's still
broken on the Mobile spec on my end, I'm thinking there's more broken
than I initially thought. Unfortunately, I'm debugging this in
airports on the way to JSConf today, and MSP doesn't seem to have
awesome desks like SFO and YVR. I'm squatting a food table with
outlets at the moment.
I suspect that a change in cordova-js or in the config broke the media
API. The weird thing is that I still see the calls going through.
On Mon, May 27, 2013 at 2:07 PM, Max Woghiren <m...@chromium.org> wrote:
> I'm seeing it in the automatic tests. If you saw it fixed in the automatic
> tests then I probably just need to figure out what's wrong on my end.
>
>
> On Mon, May 27, 2013 at 2:58 PM, Joe Bowser <bows...@gmail.com> wrote:
>
>> I meant to test the manual tests. If you're still seeing NPObject,
>> something is still broken.
>> On May 27, 2013 1:47 PM, "Max Woghiren" <m...@chromium.org> wrote:
>>
>> > Joe, was your four-line commit intended to fix the "Error calling method
>> on
>> > NPObject" errors in the automatic Media tests? I'm still getting them on
>> > twelve of those tests. Just want to make sure it's something on my end.
>> >
>> > On Fri, May 24, 2013 at 2:44 PM, Joe Bowser <bows...@gmail.com> wrote:
>> >
>> > > I've already fixed this bug! We should be all good now!
>> > >
>> > > On Fri, May 24, 2013 at 11:40 AM, Ian Clelland <iclell...@chromium.org
>> >
>> > > wrote:
>> > > > Definitely, Joe --
>> > > >
>> > > > It looks like Shravan was a little over-eager to apply the
>> DataResource
>> > > API
>> > > > to replace FileHelper, and this is definitely a bug.
>> > > >
>> > > > I'm going to revert the changes that he made to AudioHandler.java,
>> and
>> > we
>> > > > can try the media tests again.
>> > > >
>> > > >
>> > > >
>> > > >
>> > > > On Fri, May 24, 2013 at 1:59 PM, Joe Bowser <bows...@gmail.com>
>> wrote:
>> > > >
>> > > >> OK, I was wrong. It turns out the DataResource code that was added
>> in
>> > > >> doesn't have support for streaming audio. We need to get this fixed
>> > > >> ASAP. Seriously, we have these tests for a reason. Every time we
>> > > >> start changing how an API works, we should run the Mobile-Spec tests
>> > > >> to make sure that it's covered. If it's not covered by mobile-spec,
>> > > >> we should create a ticket to write a test for the Android Test
>> Suite.
>> > > >>
>> > > >> The culprit appears to be on line 60 of AudioHandler.java:
>> > > >>
>> > > >> public String getFilePath(String url, String source){
>> > > >> DataResource dataResource =
>> > > >> DataResource.initiateNewDataRequestForUri(url,
>> > > >> this.webView.pluginManager, cordova, source);
>> > > >> return dataResource.getRealFile().getPath();
>> > > >> }
>> > > >>
>> > > >> If it's an web URI, we shouldn't screw with it. Seriously!
>> > > >>
>> > > >> Joe
>> > > >>
>> > > >> On Fri, May 24, 2013 at 10:36 AM, Joe Bowser <bows...@gmail.com>
>> > wrote:
>> > > >> > Hey
>> > > >> >
>> > > >> > After looking into the bug further, for some reason URIs such as
>> > > >> > http://cordova.apache.org/downloads/BlueZedEx.mp3 are losing
>> > slashes
>> > > >> > after the protocol. (i.e.
>> > > >> > http:/cordova.apache.org/downloads/BlueZedEx.mp3). This is
>> causing
>> > > >> > Android to epic fail the media tests and I suspect that it would
>> > cause
>> > > >> > the same problem in other platforms. Did anything change that
>> would
>> > > >> > cause the URIs to be malformed? This is a pretty important thing
>> to
>> > > >> > get fixed for 2.8.0rc1, and I don't want to tag until this is
>> > > >> > resolved.
>> > > >> >
>> > > >> > Joe
>> > > >>
>> > >
>> >
>>
