Revert of Removed the legacy behavior of stopping playout when setting new receive codecs.

Revert of Removed the legacy behavior of stopping playout when setting new receive codecs. (patchset #1 id:1 of https://codereview.webrtc.org/2409483003/ )

Reason for revert:
Reverting because of the reasons given in comment #16:

"This change breaks a scenario that is unfortunately not covered by unit tests,
but can easily happen in a real call.

The scenario that is broken by the change is this:
1. A sends an offer to B, with a set of codecs C_a (which is a subset of C_b,
the codecs supported by B)
2. B responds with an answer, and sets the receive codecs to C_a.
3. At a later time, B generates a new offer which by default includes all codecs
in C_b.
4. B calls SetLocalDescription() with this offer, that adds new receive codecs.
5. Adding the new codecs fails, because of the check at
https://cs.chromium.org/chromium/src/third_party/webrtc/voice_engine/channel.....
This causes SetLocalDescription() itself to fail. The net effect is that B
cannot set a local description it just generated.

Before the CL mentioned above, we'd stop playout before changing the codecs, and
the operation would succeed."

Original issue's description:
> Removed the legacy behavior of stopping playout when setting new receive codecs.
>
> BUG=webrtc:4690
>
> Committed: https://crrev.com/917d4e1e7131f35764cff932a8793151585e8179
> Cr-Commit-Position: refs/heads/master@{#14610}

TBR=solenberg@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:4690
Committed: https://crrev.com/37b8b11661cce2691adff93ccc6bb6d1ada1bc3a
Cr-Commit-Position: refs/heads/master@{#14905}

[kwiberg-webrtc, 2016-11-03 09:17:19]

The CQ bit was checked by kwiberg@webrtc.org

[kwiberg-webrtc, 2016-11-03 09:17:19]

Created Revert of Removed the legacy behavior of stopping playout when setting
new receive codecs.

[commit-bot: I haz the power, 2016-11-03 09:17:33]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.webrtc.org/...

[kwiberg-webrtc, 2016-11-03 09:18:16]

Description was changed from

==========
Revert of Removed the legacy behavior of stopping playout when setting new
receive codecs. (patchset #1 id:1 of https://codereview.webrtc.org/2409483003/ )

Reason for revert:
Reverting because of the reasons given in comment #16:

"This change breaks a scenario that is unfortunately not covered by unit tests,
but can easily happen in a real call.

The scenario that is broken by the change is this:
1. A sends an offer to B, with a set of codecs C_a (which is a subset of C_b,
the codecs supported by B)
2. B responds with an answer, and sets the receive codecs to C_a.
3. At a later time, B generates a new offer which by default includes all codecs
in C_b.
4. B calls SetLocalDescription() with this offer, that adds new receive codecs.
5. Adding the new codecs fails, because of the check at
https://cs.chromium.org/chromium/src/third_party/webrtc/voice_engine/channel.....
This causes SetLocalDescription() itself to fail. The net effect is that B
cannot set a local description it just generated.

Before the CL mentioned above, we'd stop playout before changing the codecs, and
the operation would succeed."

Original issue's description:
> Removed the legacy behavior of stopping playout when setting new receive
codecs.
> 
> BUG=webrtc:4690
> 
> Committed: https://crrev.com/917d4e1e7131f35764cff932a8793151585e8179
> Cr-Commit-Position: refs/heads/master@{#14610}

TBR=solenberg@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:4690
==========

to

==========
Revert of Removed the legacy behavior of stopping playout when setting new
receive codecs. (patchset #1 id:1 of https://codereview.webrtc.org/2409483003/ )

Reason for revert:
Reverting because of the reasons given in comment #16:

"This change breaks a scenario that is unfortunately not covered by unit tests,
but can easily happen in a real call.

The scenario that is broken by the change is this:
1. A sends an offer to B, with a set of codecs C_a (which is a subset of C_b,
the codecs supported by B)
2. B responds with an answer, and sets the receive codecs to C_a.
3. At a later time, B generates a new offer which by default includes all codecs
in C_b.
4. B calls SetLocalDescription() with this offer, that adds new receive codecs.
5. Adding the new codecs fails, because of the check at
https://cs.chromium.org/chromium/src/third_party/webrtc/voice_engine/channel.....
This causes SetLocalDescription() itself to fail. The net effect is that B
cannot set a local description it just generated.

Before the CL mentioned above, we'd stop playout before changing the codecs, and
the operation would succeed."

Original issue's description:
> Removed the legacy behavior of stopping playout when setting new receive
codecs.
> 
> BUG=webrtc:4690
> 
> Committed: https://crrev.com/917d4e1e7131f35764cff932a8793151585e8179
> Cr-Commit-Position: refs/heads/master@{#14610}

TBR=solenberg@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:4690
==========

[kwiberg-webrtc, 2016-11-03 09:18:16]

kwiberg@webrtc.org changed reviewers:
+ stefan@webrtc.org

[commit-bot: I haz the power, 2016-11-03 09:46:56]

Description was changed from

==========
Revert of Removed the legacy behavior of stopping playout when setting new
receive codecs. (patchset #1 id:1 of https://codereview.webrtc.org/2409483003/ )

Reason for revert:
Reverting because of the reasons given in comment #16:

"This change breaks a scenario that is unfortunately not covered by unit tests,
but can easily happen in a real call.

The scenario that is broken by the change is this:
1. A sends an offer to B, with a set of codecs C_a (which is a subset of C_b,
the codecs supported by B)
2. B responds with an answer, and sets the receive codecs to C_a.
3. At a later time, B generates a new offer which by default includes all codecs
in C_b.
4. B calls SetLocalDescription() with this offer, that adds new receive codecs.
5. Adding the new codecs fails, because of the check at
https://cs.chromium.org/chromium/src/third_party/webrtc/voice_engine/channel.....
This causes SetLocalDescription() itself to fail. The net effect is that B
cannot set a local description it just generated.

Before the CL mentioned above, we'd stop playout before changing the codecs, and
the operation would succeed."

Original issue's description:
> Removed the legacy behavior of stopping playout when setting new receive
codecs.
> 
> BUG=webrtc:4690
> 
> Committed: https://crrev.com/917d4e1e7131f35764cff932a8793151585e8179
> Cr-Commit-Position: refs/heads/master@{#14610}

TBR=solenberg@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:4690
==========

to

==========
Revert of Removed the legacy behavior of stopping playout when setting new
receive codecs. (patchset #1 id:1 of https://codereview.webrtc.org/2409483003/ )

Reason for revert:
Reverting because of the reasons given in comment #16:

"This change breaks a scenario that is unfortunately not covered by unit tests,
but can easily happen in a real call.

The scenario that is broken by the change is this:
1. A sends an offer to B, with a set of codecs C_a (which is a subset of C_b,
the codecs supported by B)
2. B responds with an answer, and sets the receive codecs to C_a.
3. At a later time, B generates a new offer which by default includes all codecs
in C_b.
4. B calls SetLocalDescription() with this offer, that adds new receive codecs.
5. Adding the new codecs fails, because of the check at
https://cs.chromium.org/chromium/src/third_party/webrtc/voice_engine/channel.....
This causes SetLocalDescription() itself to fail. The net effect is that B
cannot set a local description it just generated.

Before the CL mentioned above, we'd stop playout before changing the codecs, and
the operation would succeed."

Original issue's description:
> Removed the legacy behavior of stopping playout when setting new receive
codecs.
> 
> BUG=webrtc:4690
> 
> Committed: https://crrev.com/917d4e1e7131f35764cff932a8793151585e8179
> Cr-Commit-Position: refs/heads/master@{#14610}

TBR=solenberg@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:4690
==========

[commit-bot: I haz the power, 2016-11-03 09:46:57]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-11-03 09:47:08]

Description was changed from

==========
Revert of Removed the legacy behavior of stopping playout when setting new
receive codecs. (patchset #1 id:1 of https://codereview.webrtc.org/2409483003/ )

Reason for revert:
Reverting because of the reasons given in comment #16:

"This change breaks a scenario that is unfortunately not covered by unit tests,
but can easily happen in a real call.

The scenario that is broken by the change is this:
1. A sends an offer to B, with a set of codecs C_a (which is a subset of C_b,
the codecs supported by B)
2. B responds with an answer, and sets the receive codecs to C_a.
3. At a later time, B generates a new offer which by default includes all codecs
in C_b.
4. B calls SetLocalDescription() with this offer, that adds new receive codecs.
5. Adding the new codecs fails, because of the check at
https://cs.chromium.org/chromium/src/third_party/webrtc/voice_engine/channel.....
This causes SetLocalDescription() itself to fail. The net effect is that B
cannot set a local description it just generated.

Before the CL mentioned above, we'd stop playout before changing the codecs, and
the operation would succeed."

Original issue's description:
> Removed the legacy behavior of stopping playout when setting new receive
codecs.
> 
> BUG=webrtc:4690
> 
> Committed: https://crrev.com/917d4e1e7131f35764cff932a8793151585e8179
> Cr-Commit-Position: refs/heads/master@{#14610}

TBR=solenberg@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:4690
==========

to

==========
Revert of Removed the legacy behavior of stopping playout when setting new
receive codecs. (patchset #1 id:1 of https://codereview.webrtc.org/2409483003/ )

Reason for revert:
Reverting because of the reasons given in comment #16:

"This change breaks a scenario that is unfortunately not covered by unit tests,
but can easily happen in a real call.

The scenario that is broken by the change is this:
1. A sends an offer to B, with a set of codecs C_a (which is a subset of C_b,
the codecs supported by B)
2. B responds with an answer, and sets the receive codecs to C_a.
3. At a later time, B generates a new offer which by default includes all codecs
in C_b.
4. B calls SetLocalDescription() with this offer, that adds new receive codecs.
5. Adding the new codecs fails, because of the check at
https://cs.chromium.org/chromium/src/third_party/webrtc/voice_engine/channel.....
This causes SetLocalDescription() itself to fail. The net effect is that B
cannot set a local description it just generated.

Before the CL mentioned above, we'd stop playout before changing the codecs, and
the operation would succeed."

Original issue's description:
> Removed the legacy behavior of stopping playout when setting new receive
codecs.
>
> BUG=webrtc:4690
>
> Committed: https://crrev.com/917d4e1e7131f35764cff932a8793151585e8179
> Cr-Commit-Position: refs/heads/master@{#14610}

TBR=solenberg@webrtc.org
# Not skipping CQ checks because original CL landed more than 1 days ago.
BUG=webrtc:4690

Committed: https://crrev.com/37b8b11661cce2691adff93ccc6bb6d1ada1bc3a
Cr-Commit-Position: refs/heads/master@{#14905}
==========

[commit-bot: I haz the power, 2016-11-03 09:47:09]

Patchset 1 (id:??) landed as
https://crrev.com/37b8b11661cce2691adff93ccc6bb6d1ada1bc3a
Cr-Commit-Position: refs/heads/master@{#14905}