Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536
https://codereview.chromium.org/2467563002/

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2461333003/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=

Committed: https://crrev.com/1222a443b7eea9676c3102902c5fc3d9f90eba89
Cr-Commit-Position: refs/heads/master@{#429274}

[danzh1, 2016-10-31 19:45:56]

Description was changed from

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616

public relnote: Add comment of quic async signing.

Merge internal change: 137415929

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335

n/a (style-only change)

Merge internal change: 137066933

emove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096

R=rch@chromium.org
BUG=
==========

to

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2467563002/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=
==========

[danzh1, 2016-10-31 19:49:34]

Description was changed from

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2467563002/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=
==========

to

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536
https://codereview.chromium.org/2467563002/

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2461333003/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=
==========

[danzh1, 2016-10-31 19:52:20]

The CQ bit was checked by danzh@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-10-31 19:52:46]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Ryan Hamilton, 2016-10-31 19:54:01]

lgtm

you're a merge machine today!

[commit-bot: I haz the power, 2016-10-31 20:30:23]

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

[commit-bot: I haz the power, 2016-10-31 20:30:27]

Dry run: Try jobs failed on following builders:
  ios-simulator on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-simulator/bui...)

[danzh1, 2016-11-01 14:28:24]

The CQ bit was checked by danzh@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-01 14:28:32]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-01 14:47:34]

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

[commit-bot: I haz the power, 2016-11-01 14:47:35]

Dry run: Try jobs failed on following builders:
  ios-device on master.tryserver.chromium.mac (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.mac/builders/ios-device/builds...)

[danzh1, 2016-11-01 18:20:54]

The CQ bit was checked by danzh@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-01 18:21:19]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-01 19:16:24]

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

[commit-bot: I haz the power, 2016-11-01 19:16:25]

Dry run: This issue passed the CQ dry run.

[danzh1, 2016-11-02 13:46:03]

On 2016/10/31 19:54:01, Ryan Hamilton wrote:
> lgtm
> 
> you're a merge machine today!

PTAL!

[Ryan Hamilton, 2016-11-02 14:23:42]

The CQ bit was checked by rch@chromium.org

[Ryan Hamilton, 2016-11-02 14:23:42]

lgtm

[commit-bot: I haz the power, 2016-11-02 14:23:55]

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

[commit-bot: I haz the power, 2016-11-02 14:29:26]

Description was changed from

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536
https://codereview.chromium.org/2467563002/

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2461333003/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=
==========

to

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536
https://codereview.chromium.org/2467563002/

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2461333003/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=
==========

[commit-bot: I haz the power, 2016-11-02 14:29:26]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-11-02 14:33:19]

Description was changed from

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536
https://codereview.chromium.org/2467563002/

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2461333003/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=
==========

to

==========
Landing Recent QUIC changes until Sat Oct 29 14:59:35.

Remove unused QuicTags.  No functional change.

Merge internal change: 137601502
https://codereview.chromium.org/2470473002/

Add ENDPOINT to a QuicSession DLOG output.

n/a (debug log)

Merge internal change: 137547744
https://codereview.chromium.org/2468663002/

Make QuicTestClient::response_headers() return a SPDY header block, not
BalsaHeaders.

n/a - test only

Merge internal change: 137425616
https://codereview.chromium.org/2466493003/

public relnote: Add comment of quic async signing.

Merge internal change: 137415929
https://codereview.chromium.org/2464983002/

Add explicit != nullptr check in quic_session. No behavior change.

Merge internal change: 137192395
https://codereview.chromium.org/2459293003/

Add missing plumbing of ProofSource::Details in some QUIC codepaths

This CL adds a bit of code for moving ProofSource::Details objects
through the codepaths for evaluating CHLOs.  Such an object is produced
by ProofSource::GetProof calls in three places:
QuicCryptoServerConfig::{ValidateClientHello, ProcessClientHello,
BuildServerConfigUpdateMessage}.  When a CHLO is evaluated, both
ValidateClientHello and ProcessClientHello are called, but the code is
structured so that only one of these will make the call to GetProof.
But it is possible for the first call not to call GetProof, and for the
second to call it instead.  In this case, the ProofSource::Details will
get dropped on the floor.

This issue was not detected in testing because in practice, apparently
the first call always invokes GetProof.  This change is therefore mostly
for completeness and future-proofing.

Additional plumbing for stats propagation in QUIC code.  Mostly
protected by existing flag --enable_async_get_proof.

Merge internal change: 137148536
https://codereview.chromium.org/2467563002/

Add connection_options argument to ProofSource::GetProof

This CL adds a new argument to both overloads of ProofSource::GetProof,
permitting connection_options to be passed in.  Currently the
ProofSource ignores these arguments, but they will be employed shortly
for passing information about Mentat experiments.

Added connection_options arguments to ProofSource::GetProof, currently
unused.  No functional change intended, not flag-protected.

Merge internal change: 137083335
https://codereview.chromium.org/2461333003/

n/a (style-only change)

Merge internal change: 137066933
https://codereview.chromium.org/2465933002/

Remove unused argument from CryptoTestUtils::SetupCryptoServerConfigForTest

n/a test changes only

Merge internal change: 137044096
https://codereview.chromium.org/2459283003/

R=rch@chromium.org
BUG=

Committed: https://crrev.com/1222a443b7eea9676c3102902c5fc3d9f90eba89
Cr-Commit-Position: refs/heads/master@{#429274}
==========

[commit-bot: I haz the power, 2016-11-02 14:33:20]

Patchset 3 (id:??) landed as
https://crrev.com/1222a443b7eea9676c3102902c5fc3d9f90eba89
Cr-Commit-Position: refs/heads/master@{#429274}