Add context-aware "time" library wrapper.

Add context-aware "time" library wrapper.

TEST=unittest
R=iannucci@chromium.org
BUG=

Committed: https://chromium.googlesource.com/infra/infra/+/a6bb752f1c7a93854d3bb17c31d608282d539a69

[dnj, 2015-06-02 03:10:42]

PTAL

[iannucci, 2015-06-02 06:15:34]

wdyt?

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/clo...
File go/src/infra/libs/clock/clock.go (right):

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/clo...
go/src/infra/libs/clock/clock.go:22: // Implementation of Clock that uses Go's
standard library.
implementations should go in subpackages I think. maybe:


clock/
  // Clock interface
  // Testable interface
  .SetFactory
  .Set
  .Get
  realclock/
    .Use(context) context  // adds a real clock implementation to context
  testclock/
    .Use(context) context  // adds a testing clock implementation to context
    // Also implements clock.Testable, so clock.Get(c).(clock.Testable) works.
    // or have Use return a clock.Testable for pulling the strings on the clock.

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/con...
File go/src/infra/libs/clock/context.go (right):

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/con...
go/src/infra/libs/clock/context.go:20: func SetClockFactory(ctx context.Context,
f func(context.Context) Clock) context.Context {
  type ClockFactory func(context.Context) Clock

  func SetClockFactory(ctx context.Context, f ClockFactory) context.Context
{...}

mebbeh?

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/dur...
File go/src/infra/libs/clock/duration.go (right):

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/dur...
go/src/infra/libs/clock/duration.go:5: package clock
maybe this could go in

  clock/clockflag

?

Then `clockflag.Duration`, `clockflag.Time`, etc.

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/tes...
File go/src/infra/libs/clock/testclock.go (right):

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/tes...
go/src/infra/libs/clock/testclock.go:21: 
Add a line here

  var _ Clock = (*TestClock)(nil)

Then you don't need the `X implements Y` comments.

[dnj (Google), 2015-06-02 17:01:13]

Patchset #2 (id:20001) has been deleted

[dnj, 2015-06-02 17:03:12]

The one thing we lose for sure is the ability to have the context default to
systemClock if no clock is set, as that would create an import cycle in the
refactored world (short of more refactoring).

That said, I think that's okay, so here is a refactored version along the lines
that you detailed.

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/con...
File go/src/infra/libs/clock/context.go (right):

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/con...
go/src/infra/libs/clock/context.go:20: func SetClockFactory(ctx context.Context,
f func(context.Context) Clock) context.Context {
On 2015/06/02 06:15:33, iannucci wrote:
>   type ClockFactory func(context.Context) Clock
> 
>   func SetClockFactory(ctx context.Context, f ClockFactory) context.Context
> {...}
> 
> mebbeh?

Done.

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/dur...
File go/src/infra/libs/clock/duration.go (right):

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/dur...
go/src/infra/libs/clock/duration.go:5: package clock
On 2015/06/02 06:15:33, iannucci wrote:
> maybe this could go in
> 
>   clock/clockflag
> 
> ?
> 
> Then `clockflag.Duration`, `clockflag.Time`, etc.

Done.

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/tes...
File go/src/infra/libs/clock/testclock.go (right):

https://codereview.chromium.org/1154213012/diff/1/go/src/infra/libs/clock/tes...
go/src/infra/libs/clock/testclock.go:21: 
On 2015/06/02 06:15:34, iannucci wrote:
> Add a line here
> 
>   var _ Clock = (*TestClock)(nil)
> 
> Then you don't need the `X implements Y` comments.

Done.

[iannucci, 2015-06-02 22:19:03]

talked about some stuff offline. Basically the proposal for timers is to have a
testing callback which gets poked every time a timer is created, which would
allow a test to block until a single 'wave' of timers is created. Then the test
could advance time in a deterministic way, and the various Timers would unblock
as time advanced past them, in order.

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/systemclock/systemclock.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/systemclock/systemclock.go:1: // Copyright (c) 2014 The
Chromium Authors. All rights reserved.
(copyrights)

Also, as discussed, I think it makes sense to have the system clock impl go in
the base package (just so that the default implementation can be the system
clock).

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/systemclock/systemclock.go:21: var _ clock.Clock =
(*systemClock)(nil)
I think this would be even better as

  var _ clock.Clock = systemClock{}

since you don't actually use the pointer receiver for anything.

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/systemclock/timer.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/systemclock/timer.go:14: }
var _ clock.Timer = (*systemTimer)(nil)  // since you DO use the pointer
receiver

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/testclock/testclock.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testclock.go:26: type testClock struct {
mutex for protecting now/timeC/tf

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testclock.go:39: tf:  tf,
timeC?

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/testclock/testtimer.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testtimer.go:47: // Implements Timer.
don't need these comments

[dnj, 2015-06-03 00:19:06]

Patchset #3 (id:60001) has been deleted

[dnj, 2015-06-03 00:25:28]

Updated!

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/systemclock/systemclock.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/systemclock/systemclock.go:1: // Copyright (c) 2014 The
Chromium Authors. All rights reserved.
On 2015/06/02 22:19:02, iannucci wrote:
> (copyrights)
> 
> Also, as discussed, I think it makes sense to have the system clock impl go in
> the base package (just so that the default implementation can be the system
> clock).

Done.

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/systemclock/systemclock.go:21: var _ clock.Clock =
(*systemClock)(nil)
On 2015/06/02 22:19:03, iannucci wrote:
> I think this would be even better as
> 
>   var _ clock.Clock = systemClock{}
> 
> since you don't actually use the pointer receiver for anything.

Done.

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/systemclock/timer.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/systemclock/timer.go:14: }
On 2015/06/02 22:19:03, iannucci wrote:
> var _ clock.Timer = (*systemTimer)(nil)  // since you DO use the pointer
> receiver

Done.

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/testclock/testclock.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testclock.go:26: type testClock struct {
On 2015/06/02 22:19:03, iannucci wrote:
> mutex for protecting now/timeC/tf

Done.

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testclock.go:39: tf:  tf,
On 2015/06/02 22:19:03, iannucci wrote:
> timeC?

Acknowledged.

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
File go/src/infra/libs/clock/testclock/testtimer.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/40001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testtimer.go:47: // Implements Timer.
On 2015/06/02 22:19:03, iannucci wrote:
> don't need these comments

Done.

[dnj, 2015-06-03 03:41:43]

Oh forgot to note: updated gae tests.

Also, FWIW golang Timer isn't goroutine-safe. Think I should remove mutexes from
testTimer so it doesn't pretend to be?

https://chromiumcodereview.appspot.com/1154213012/diff/80001/go/src/infra/lib...
File go/src/infra/libs/clock/testclock/testtimer.go (right):

https://chromiumcodereview.appspot.com/1154213012/diff/80001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testtimer.go:57: active = t.Stop()
Racy

https://chromiumcodereview.appspot.com/1154213012/diff/80001/go/src/infra/lib...
go/src/infra/libs/clock/testclock/testtimer.go:102: 
Move up

[dnj, 2015-06-03 17:03:36]

Updated again to make testtimer not goroutine-safe. I wouldn't want to give the
false impression that sharing timers across goroutines is suddenly cool.

[iannucci, 2015-06-03 17:37:20]

lg, though I think a lot of the names can improve now that you're all split into
subpackages. Also some of the subpackage 'package' lines are wron.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
File go/src/infra/gae/libs/wrapper/dummy_test.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
go/src/infra/gae/libs/wrapper/dummy_test.go:27: c = clock.SetClock(c, tc)
wdyt about 

  c = testclock.UseTime(c, time.Date(...))  

(which calls clock.SetClock internally)?

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
File go/src/infra/gae/libs/wrapper/memory/memcache.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
go/src/infra/gae/libs/wrapper/memory/memcache.go:34: ctx  context.Context
+1

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
File go/src/infra/gae/libs/wrapper/memory/taskqueue.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
go/src/infra/gae/libs/wrapper/memory/taskqueue.go:67: mathRand *rand.Rand
could we pull out mathRand the same way too? (e.g. delete and just call
wrapper.MathRand on ctx)

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/clockflag/time.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/clockflag/time.go:20: )
or

var _ interface {
 flag.Value
 json.Unmarshaler
 json.Marshaler
} = (*Time)(nil)

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/context.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/context.go:23: func SetClockFactory(ctx context.Context,
f Factory) context.Context {
maybe clock.SetFactory

See 3rd para here: https://golang.org/doc/effective_go.html#package-names

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/context.go:28: func SetClock(ctx context.Context, c
Clock) context.Context {
maybe clock.Set

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/context.go:34: func GetClock(ctx context.Context) (clock
Clock) {
maybe clock.Get

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/systemclock.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemclock.go:5: package clock
package systemclock

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemclock.go:21: func GetSystemClock() Clock {
Get

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/systemtimer.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer.go:5: package clock
systemtimer

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer.go:16: var _ Timer = (*systemTimer)(nil)
where does Timer come from? Should this be clock.Timer?

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer.go:18: func (t *systemTimer) GetC() (c
<-chan time.Time) {
C()

see https://golang.org/doc/effective_go.html#package-names

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/systemtimer_test.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer_test.go:14: func TestSystemTimer(t
*testing.T) {
in general we should be in the habit of starting all test functions with

  t.Parallel()

As both an indicator that the test can be run independently, and so when we run
with GOMAXPROCS>1 (the default in 1.5), we'll get a speed boost.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/testclock/testclock.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/testclock/testclock.go:23: // sychronizing state when
testing.
I think we're still trying to wrap 80 on comments everywhere.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/testclock/testclock_test.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/testclock/testclock_test.go:51: awakeC <- <-afterC
well that's a very odd construction... why not just wait on afterC?

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/testclock/testtimer.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/testclock/testtimer.go:56: t.cancelFunc =
t.clock.whenTimePasses(triggerTime, t.signal)
why not `whenPastTime`, since it's taking an absolute time and not a relative
one?

[dnj, 2015-06-03 18:21:27]

Updated!

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
File go/src/infra/gae/libs/wrapper/dummy_test.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
go/src/infra/gae/libs/wrapper/dummy_test.go:27: c = clock.SetClock(c, tc)
On 2015/06/03 17:37:19, iannucci wrote:
> wdyt about 
> 
>   c = testclock.UseTime(c, time.Date(...))  
> 
> (which calls clock.SetClock internally)?

You need a reference to the test clock in order to poke it. We could do:
ctx, clock := testclock.UseTime(time.Date(...))

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
File go/src/infra/gae/libs/wrapper/memory/taskqueue.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/gae/libs/...
go/src/infra/gae/libs/wrapper/memory/taskqueue.go:67: mathRand *rand.Rand
On 2015/06/03 17:37:19, iannucci wrote:
> could we pull out mathRand the same way too? (e.g. delete and just call
> wrapper.MathRand on ctx)

Done.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/clockflag/time.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/clockflag/time.go:20: )
On 2015/06/03 17:37:19, iannucci wrote:
> or
> 
> var _ interface {
>  flag.Value
>  json.Unmarshaler
>  json.Marshaler
> } = (*Time)(nil)

Oh cool, didn't think of this notation. Done.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/context.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/context.go:23: func SetClockFactory(ctx context.Context,
f Factory) context.Context {
On 2015/06/03 17:37:20, iannucci wrote:
> maybe clock.SetFactory
> 
> See 3rd para here: https://golang.org/doc/effective_go.html#package-names

ClockSetTheClockFactoryForContext

(okay)

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/context.go:28: func SetClock(ctx context.Context, c
Clock) context.Context {
On 2015/06/03 17:37:20, iannucci wrote:
> maybe clock.Set

Done.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/context.go:34: func GetClock(ctx context.Context) (clock
Clock) {
On 2015/06/03 17:37:20, iannucci wrote:
> maybe clock.Get

Done.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/systemclock.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemclock.go:5: package clock
On 2015/06/03 17:37:20, iannucci wrote:
> package systemclock

Not anymore, is in main "clock" package.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemclock.go:21: func GetSystemClock() Clock {
On 2015/06/03 17:37:20, iannucci wrote:
> Get

(In "clock" package now).

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/systemtimer.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer.go:5: package clock
On 2015/06/03 17:37:20, iannucci wrote:
> systemtimer

(In "clock" package now).

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer.go:16: var _ Timer = (*systemTimer)(nil)
On 2015/06/03 17:37:20, iannucci wrote:
> where does Timer come from? Should this be clock.Timer?

(In "clock" package now).

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer.go:18: func (t *systemTimer) GetC() (c
<-chan time.Time) {
On 2015/06/03 17:37:20, iannucci wrote:
> C()
> 
> see https://golang.org/doc/effective_go.html#package-names

(In "clock" package now).

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/systemtimer_test.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/systemtimer_test.go:14: func TestSystemTimer(t
*testing.T) {
On 2015/06/03 17:37:20, iannucci wrote:
> in general we should be in the habit of starting all test functions with
> 
>   t.Parallel()
> 
> As both an indicator that the test can be run independently, and so when we
run
> with GOMAXPROCS>1 (the default in 1.5), we'll get a speed boost.

Done.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/testclock/testclock.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/testclock/testclock.go:23: // sychronizing state when
testing.
On 2015/06/03 17:37:20, iannucci wrote:
> I think we're still trying to wrap 80 on comments everywhere.

I did a best-effort run. I wonder if this is enforceable via golint settings...

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/testclock/testclock_test.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/testclock/testclock_test.go:51: awakeC <- <-afterC
On 2015/06/03 17:37:20, iannucci wrote:
> well that's a very odd construction... why not just wait on afterC?

Heh no idea.

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
File go/src/infra/libs/clock/testclock/testtimer.go (right):

https://codereview.chromium.org/1154213012/diff/140001/go/src/infra/libs/cloc...
go/src/infra/libs/clock/testclock/testtimer.go:56: t.cancelFunc =
t.clock.whenTimePasses(triggerTime, t.signal)
On 2015/06/03 17:37:20, iannucci wrote:
> why not `whenPastTime`, since it's taking an absolute time and not a relative
> one?

"Past" implies ">", but I want to enforce ">=".

I went with "invokeAt".

[iannucci, 2015-06-04 00:20:46]

The CQ bit was checked by iannucci@chromium.org

[iannucci, 2015-06-04 00:20:47]

lgtm

[commit-bot: I haz the power, 2015-06-04 00:21:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1154213012/180001

[commit-bot: I haz the power, 2015-06-04 00:26:51]

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

[commit-bot: I haz the power, 2015-06-04 00:26:52]

Try jobs failed on following builders:
  infra_tester on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/infra_tester/bu...)

[dnj, 2015-06-04 00:31:21]

The CQ bit was checked by dnj@chromium.org

[dnj, 2015-06-04 00:31:22]

The patchset sent to the CQ was uploaded after l-g-t-m from
iannucci@chromium.org
Link to the patchset: https://codereview.chromium.org/1154213012/#ps200001
(title: "Added coverage files.")

[commit-bot: I haz the power, 2015-06-04 00:31:33]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1154213012/200001

[commit-bot: I haz the power, 2015-06-04 00:34:33]

Committed patchset #9 (id:200001) as
https://chromium.googlesource.com/infra/infra/+/a6bb752f1c7a93854d3bb17c31d60...