Downgrade media::Pipeline CHECKs into NOTREACHED() + no-op.

Downgrade media::Pipeline CHECKs into NOTREACHED() + no-op.

r150930 set Pipeline::running_ to false when an error is encountered as opposed to keeping the Pipeline running. The end result is that an in-flight Seek() can hit CHECK(running_). We want a no-op in that situation in release mode, not a process crash.

BUG=141923
TBR=fischman

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=151662

[scherkus (not reviewing), 2012-08-15 04:58:21]

admittedly a bit of a bandaid but I'm working on a layout test to cover this
situation

[Ami GONE FROM CHROMIUM, 2012-08-15 05:07:31]

lgtm

[Ami GONE FROM CHROMIUM, 2012-08-15 05:08:11]

https://chromiumcodereview.appspot.com/10855167/diff/1/media/base/pipeline.cc
File media/base/pipeline.cc (right):

https://chromiumcodereview.appspot.com/10855167/diff/1/media/base/pipeline.cc...
media/base/pipeline.cc:113: if (running_) {
wait, what?  How can this happen?

[scherkus (not reviewing), 2012-08-15 05:08:59]

https://chromiumcodereview.appspot.com/10855167/diff/1/media/base/pipeline.cc
File media/base/pipeline.cc (right):

https://chromiumcodereview.appspot.com/10855167/diff/1/media/base/pipeline.cc...
media/base/pipeline.cc:113: if (running_) {
On 2012/08/15 05:08:11, Ami Fischman wrote:
> wait, what?  How can this happen?

being silly (i.e., calling Start() twice)

[Ami GONE FROM CHROMIUM, 2012-08-15 05:09:49]

>  wait, what?  How can this happen?
>
> being silly (i.e., calling Start() twice)
>
Is it possible to do this?

[scherkus (not reviewing), 2012-08-15 05:13:46]

On 2012/08/15 05:09:49, Ami Fischman wrote:
> >  wait, what?  How can this happen?
> >
> > being silly (i.e., calling Start() twice)
> >
> Is it possible to do this?

None of our existing code does this today.

[Ami GONE FROM CHROMIUM, 2012-08-15 05:21:48]

On Tue, Aug 14, 2012 at 10:13 PM, <scherkus@chromium.org> wrote:

> None of our existing code does this today.
>

Yeah, well, a followup CL can take it out, then :)