Play WebM live stream

Play WebM live stream

BUG=44891
TEST=Can play live stream http://195.10.10.75:8800/live.webm

This patch allows a live WebM stream to be played.

After demuxing by FFmpeg, the duration was set to an invalid value. This
makes the video pipeline to set duration to 0, and so WebKit never allows
us to play because currentTime == duration all the time.

This patch sets duration to an infinite value when duration is reported as
invalid in FFmpeg.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=54983

[fgalligan1, 2010-08-03 14:17:38]

http://codereview.chromium.org/3031036/diff/1/3
File media/filters/ffmpeg_demuxer.cc (right):

http://codereview.chromium.org/3031036/diff/1/3#newcode479
media/filters/ffmpeg_demuxer.cc:479: 
Would it be better to check if max_duration == AV_NOPTS_VALUE instead of the
else on the format_context duration. At one point I think ogg files had a
duration on the streams but not on the format_context.

[fgalligan1, 2010-08-04 13:08:27]

LGTM

I tested on Windows and Linux overnight and the live streams were still playing
in sync in the morning.

[Alpha Left Google, 2010-08-04 18:03:12]

On 2010/08/03 14:17:38, fgalligan1 wrote:
> http://codereview.chromium.org/3031036/diff/1/3
> File media/filters/ffmpeg_demuxer.cc (right):
> 
> http://codereview.chromium.org/3031036/diff/1/3#newcode479
> media/filters/ffmpeg_demuxer.cc:479: 
> Would it be better to check if max_duration == AV_NOPTS_VALUE instead of the
> else on the format_context duration. At one point I think ogg files had a
> duration on the streams but not on the format_context.

Sorry I missed this comment, will try.

[fgalligan, 2010-08-05 15:59:32]

Will this make it into 6?

On Wed, Aug 4, 2010 at 2:03 PM, <hclam@chromium.org> wrote:

> On 2010/08/03 14:17:38, fgalligan1 wrote:
>
>> http://codereview.chromium.org/3031036/diff/1/3
>> File media/filters/ffmpeg_demuxer.cc (right):
>>
>
>  http://codereview.chromium.org/3031036/diff/1/3#newcode479
>> media/filters/ffmpeg_demuxer.cc:479:
>> Would it be better to check if max_duration == AV_NOPTS_VALUE instead of
>> the
>> else on the format_context duration. At one point I think ogg files had a
>> duration on the streams but not on the format_context.
>>
>
> Sorry I missed this comment, will try.
>
>
> http://codereview.chromium.org/3031036/show
>

[hclam, 2010-08-05 17:53:17]

How important do you think this is? The change itself is pretty small, but
the fact that it assumes infinity for invalid duration might be a little
nasty and I would like to see how it works in the wild.

Alpha

2010年8月5日8:59 Frank Galligan <fgalligan@google.com>:

> Will this make it into 6?
>
>
> On Wed, Aug 4, 2010 at 2:03 PM, <hclam@chromium.org> wrote:
>
>> On 2010/08/03 14:17:38, fgalligan1 wrote:
>>
>>> http://codereview.chromium.org/3031036/diff/1/3
>>> File media/filters/ffmpeg_demuxer.cc (right):
>>>
>>
>>  http://codereview.chromium.org/3031036/diff/1/3#newcode479
>>> media/filters/ffmpeg_demuxer.cc:479:
>>> Would it be better to check if max_duration == AV_NOPTS_VALUE instead of
>>> the
>>> else on the format_context duration. At one point I think ogg files had a
>>> duration on the streams but not on the format_context.
>>>
>>
>> Sorry I missed this comment, will try.
>>
>>
>> http://codereview.chromium.org/3031036/show
>>
>
>

[fgalligan, 2010-08-05 18:01:36]

My guess is it is fairly important. I was planning to add support
for Chrome but you beat me to it by a week or so.

Frank


On Thu, Aug 5, 2010 at 1:53 PM, Alpha (Hin-Chung) Lam <hclam@google.com>wrote:

> How important do you think this is? The change itself is pretty small, but
> the fact that it assumes infinity for invalid duration might be a little
> nasty and I would like to see how it works in the wild.
>
> Alpha
>
> 2010年8月5日8:59 Frank Galligan <fgalligan@google.com>:
>
> Will this make it into 6?
>>
>>
>> On Wed, Aug 4, 2010 at 2:03 PM, <hclam@chromium.org> wrote:
>>
>>> On 2010/08/03 14:17:38, fgalligan1 wrote:
>>>
>>>> http://codereview.chromium.org/3031036/diff/1/3
>>>> File media/filters/ffmpeg_demuxer.cc (right):
>>>>
>>>
>>>  http://codereview.chromium.org/3031036/diff/1/3#newcode479
>>>> media/filters/ffmpeg_demuxer.cc:479:
>>>> Would it be better to check if max_duration == AV_NOPTS_VALUE instead of
>>>> the
>>>> else on the format_context duration. At one point I think ogg files had
>>>> a
>>>> duration on the streams but not on the format_context.
>>>>
>>>
>>> Sorry I missed this comment, will try.
>>>
>>>
>>> http://codereview.chromium.org/3031036/show
>>>
>>
>>
>

[hclam, 2010-08-05 18:04:48]

I can change it's priority and merge it into chrome 6. Since andrew has been
doing the merging, what's the schedule for chrome 6?

Alpha

2010年8月5日11:01 Frank Galligan <fgalligan@google.com>:

> My guess is it is fairly important. I was planning to add support
> for Chrome but you beat me to it by a week or so.
>
> Frank
>
>
> On Thu, Aug 5, 2010 at 1:53 PM, Alpha (Hin-Chung) Lam <hclam@google.com>wrote:
>
>> How important do you think this is? The change itself is pretty small, but
>> the fact that it assumes infinity for invalid duration might be a little
>> nasty and I would like to see how it works in the wild.
>>
>> Alpha
>>
>> 2010年8月5日8:59 Frank Galligan <fgalligan@google.com>:
>>
>> Will this make it into 6?
>>>
>>>
>>> On Wed, Aug 4, 2010 at 2:03 PM, <hclam@chromium.org> wrote:
>>>
>>>> On 2010/08/03 14:17:38, fgalligan1 wrote:
>>>>
>>>>> http://codereview.chromium.org/3031036/diff/1/3
>>>>> File media/filters/ffmpeg_demuxer.cc (right):
>>>>>
>>>>
>>>>  http://codereview.chromium.org/3031036/diff/1/3#newcode479
>>>>> media/filters/ffmpeg_demuxer.cc:479:
>>>>> Would it be better to check if max_duration == AV_NOPTS_VALUE instead
>>>>> of the
>>>>> else on the format_context duration. At one point I think ogg files had
>>>>> a
>>>>> duration on the streams but not on the format_context.
>>>>>
>>>>
>>>> Sorry I missed this comment, will try.
>>>>
>>>>
>>>> http://codereview.chromium.org/3031036/show
>>>>
>>>
>>>
>>
>

[hclam, 2010-08-06 22:13:01]

I have merged it into branch 472. Are you building this branch for testing?

Alpha


2010年8月5日11:04 Alpha (Hin-Chung) Lam <hclam@google.com>:

> I can change it's priority and merge it into chrome 6. Since andrew has
> been doing the merging, what's the schedule for chrome 6?
>
> Alpha
>
> 2010年8月5日11:01 Frank Galligan <fgalligan@google.com>:
>
> My guess is it is fairly important. I was planning to add support
>> for Chrome but you beat me to it by a week or so.
>>
>> Frank
>>
>>
>> On Thu, Aug 5, 2010 at 1:53 PM, Alpha (Hin-Chung) Lam
<hclam@google.com>wrote:
>>
>>> How important do you think this is? The change itself is pretty small,
>>> but the fact that it assumes infinity for invalid duration might be a little
>>> nasty and I would like to see how it works in the wild.
>>>
>>> Alpha
>>>
>>> 2010年8月5日8:59 Frank Galligan <fgalligan@google.com>:
>>>
>>> Will this make it into 6?
>>>>
>>>>
>>>> On Wed, Aug 4, 2010 at 2:03 PM, <hclam@chromium.org> wrote:
>>>>
>>>>> On 2010/08/03 14:17:38, fgalligan1 wrote:
>>>>>
>>>>>> http://codereview.chromium.org/3031036/diff/1/3
>>>>>> File media/filters/ffmpeg_demuxer.cc (right):
>>>>>>
>>>>>
>>>>>  http://codereview.chromium.org/3031036/diff/1/3#newcode479
>>>>>> media/filters/ffmpeg_demuxer.cc:479:
>>>>>> Would it be better to check if max_duration == AV_NOPTS_VALUE instead
>>>>>> of the
>>>>>> else on the format_context duration. At one point I think ogg files
>>>>>> had a
>>>>>> duration on the streams but not on the format_context.
>>>>>>
>>>>>
>>>>> Sorry I missed this comment, will try.
>>>>>
>>>>>
>>>>> http://codereview.chromium.org/3031036/show
>>>>>
>>>>
>>>>
>>>
>>
>

[fgalligan, 2010-08-09 13:33:12]

No, But I'm going to check some changes in soon. So I'm reading up on it
right now.

Frank


On Fri, Aug 6, 2010 at 6:12 PM, Alpha (Hin-Chung) Lam <hclam@google.com>wrote:

> I have merged it into branch 472. Are you building this branch for testing?
>
> Alpha
>
>
> 2010年8月5日11:04 Alpha (Hin-Chung) Lam <hclam@google.com>:
>
> I can change it's priority and merge it into chrome 6. Since andrew has
>> been doing the merging, what's the schedule for chrome 6?
>>
>> Alpha
>>
>> 2010年8月5日11:01 Frank Galligan <fgalligan@google.com>:
>>
>> My guess is it is fairly important. I was planning to add support
>>> for Chrome but you beat me to it by a week or so.
>>>
>>> Frank
>>>
>>>
>>> On Thu, Aug 5, 2010 at 1:53 PM, Alpha (Hin-Chung) Lam
<hclam@google.com>wrote:
>>>
>>>> How important do you think this is? The change itself is pretty small,
>>>> but the fact that it assumes infinity for invalid duration might be a
little
>>>> nasty and I would like to see how it works in the wild.
>>>>
>>>> Alpha
>>>>
>>>> 2010年8月5日8:59 Frank Galligan <fgalligan@google.com>:
>>>>
>>>> Will this make it into 6?
>>>>>
>>>>>
>>>>> On Wed, Aug 4, 2010 at 2:03 PM, <hclam@chromium.org> wrote:
>>>>>
>>>>>> On 2010/08/03 14:17:38, fgalligan1 wrote:
>>>>>>
>>>>>>> http://codereview.chromium.org/3031036/diff/1/3
>>>>>>> File media/filters/ffmpeg_demuxer.cc (right):
>>>>>>>
>>>>>>
>>>>>>  http://codereview.chromium.org/3031036/diff/1/3#newcode479
>>>>>>> media/filters/ffmpeg_demuxer.cc:479:
>>>>>>> Would it be better to check if max_duration == AV_NOPTS_VALUE instead
>>>>>>> of the
>>>>>>> else on the format_context duration. At one point I think ogg files
>>>>>>> had a
>>>>>>> duration on the streams but not on the format_context.
>>>>>>>
>>>>>>
>>>>>> Sorry I missed this comment, will try.
>>>>>>
>>>>>>
>>>>>> http://codereview.chromium.org/3031036/show
>>>>>>
>>>>>
>>>>>
>>>>
>>>
>>
>