Try creating sockets again if network change occurs after bind failed.

Try creating sockets again if network change occurs after bind failed.

If the network interface appears active, but binding the sockets fails,
then it won't produce any candidates even though it's never marked as
"network failed". So this was causing nothing to happen once a network
change event occurs and the interface becomes usable again.

So, this CL adds the condition that we only disable gathering of local
ports if we don't have them already.

See bug for more details.

BUG=webrtc:8256
Review-Url: https://codereview.webrtc.org/3015543002
Cr-Commit-Position: refs/heads/master@{#20007}
Committed: https://webrtc.googlesource.com/src/+/1c46a35c5e3ba0a4f73208f91305937384895c47

[Taylor Brandstetter, 2017-09-16 07:48:25]

deadbeef@webrtc.org changed reviewers:
+ pthatcher@webrtc.org, zhihuang@webrtc.org

[Taylor Brandstetter, 2017-09-16 07:48:25]

PTAL Peter and/or Zhi

[Zhi Huang, 2017-09-22 21:40:48]

On 2017/09/16 07:48:25, Taylor Brandstetter wrote:
> PTAL Peter and/or Zhi

lgtm.

[pthatcher1, 2017-09-26 20:44:39]

lgtm

[Taylor Brandstetter, 2017-09-27 12:40:48]

The CQ bit was checked by deadbeef@webrtc.org

[commit-bot: I haz the power, 2017-09-27 12:40:57]

CQ is trying da patch.

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

[commit-bot: I haz the power, 2017-09-27 12:48:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2017-09-27 12:48:07]

Try jobs failed on following builders:
  mac_asan on master.tryserver.webrtc (JOB_FAILED,
http://build.chromium.org/p/tryserver.webrtc/builders/mac_asan/builds/28408)

[Taylor Brandstetter, 2017-09-27 17:54:59]

I'm gonna go ahead and submit, but you might want to take a look again. I split
up "TestGatherLowCostNetworkOnly" because it was trying to test three things in
the same test, using the same port allocator. Which isn't good in the first
place, but in this case it was running into weird interactions with continual
gathering logic and regathering logic that the test wasn't meant to be
exercising in the first place.

So, I split it up. And also added a missing test that ensures that if costly
networks are disabled but cell is the only option, it ends up used anyway.

[Taylor Brandstetter, 2017-09-27 17:55:09]

The CQ bit was checked by deadbeef@webrtc.org

[Taylor Brandstetter, 2017-09-27 17:55:09]

The patchset sent to the CQ was uploaded after l-g-t-m from zhihuang@webrtc.org,
pthatcher@webrtc.org
Link to the patchset: https://codereview.webrtc.org/3015543002/#ps20001 (title:
"Split up TestGatherLowCostNetworkOnly")

[commit-bot: I haz the power, 2017-09-27 17:55:17]

CQ is trying da patch.

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

[commit-bot: I haz the power, 2017-09-27 18:24:07]

CQ is committing da patch.

Bot data: {"patchset_id": 20001, "attempt_start_ts": 1506534909088580,
"parent_rev": "df5bb65ce4795105ff1bd88d4672d1d5f5792849", "commit_rev":
"1c46a35c5e3ba0a4f73208f91305937384895c47"}

[commit-bot: I haz the power, 2017-09-27 18:24:12]

Description was changed from

==========
Try creating sockets again if network change occurs after bind failed.

If the network interface appears active, but binding the sockets fails,
then it won't produce any candidates even though it's never marked as
"network failed". So this was causing nothing to happen once a network
change event occurs and the interface becomes usable again.

So, this CL adds the condition that we only disable gathering of local
ports if we don't have them already.

See bug for more details.

BUG=webrtc:8256
==========

to

==========
Try creating sockets again if network change occurs after bind failed.

If the network interface appears active, but binding the sockets fails,
then it won't produce any candidates even though it's never marked as
"network failed". So this was causing nothing to happen once a network
change event occurs and the interface becomes usable again.

So, this CL adds the condition that we only disable gathering of local
ports if we don't have them already.

See bug for more details.

BUG=webrtc:8256

Review-Url: https://codereview.webrtc.org/3015543002
Cr-Commit-Position: refs/heads/master@{#20007}
Committed:
https://webrtc.googlesource.com/src/+/1c46a35c5e3ba0a4f73208f91305937384895c47
==========

[commit-bot: I haz the power, 2017-09-27 18:24:13]

Committed patchset #2 (id:20001) as
https://webrtc.googlesource.com/src/+/1c46a35c5e3ba0a4f73208f91305937384895c47