MockClock needs to mock the NowAsDeltaSinceUnixEpoch method.

MockClock needs to mock the NowAsDeltaSinceUnixEpoch method.
Remove the declaration of an undefined AdvanceTime method.

R=rch@chromium.org
BUG=none
TEST=QuicClientSessionTest.CryptoConnectSendsCorrectData does not
fail with different Unix timestamps in ClientHello even when the
machine is slow.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=177297

[wtc, 2013-01-16 22:36:56]

QuicClientSessionTest.CryptoConnectSendsCorrectData failed once
on the "Windows Tests (tsan)" bot. I found that the test
is using a MockClock, but the Unix time stamps in ClientHello
messages did change if I inserted an artificial sleep of
two seconds. It's because we were calling the real
QuicTime::NowAsDeltaSinceUnixEpoch implementation.

Here is the output of a failed run of the
QuicClientSessionTest.CryptoConnectSendsCorrectData test:

QuicClientSessionTest.CryptoConnectSendsCorrectData: 

GMOCK WARNING:
Uninteresting mock function call - returning directly.
Function call: SetConnection(1C3A93C0)
Stack trace:

GMOCK WARNING:
Uninteresting mock function call - returning directly.
Function call: SetTimeoutAlarm(8-byte object <00-46 C3-23 00-00 00-00>)
Stack trace:
quic\test_tools\quic_test_utils.cc(183): error: Failed
Description:
CHLO

Expected:
01  00  00  00   ....
00  00  00  00   ....
01  00  00  00   ....
00  00  00  00   ....
01  00  01  00   ....
00  00  00  00   ....
00  00  00  00   ....
00  00  00  70   ...p
00  43  48  4c   .CHL
4f  07  00  4e   O..N
4f  4e  43  41   ONCA
45  41  44  49   EADI
43  53  4c  4b   CSLK
41  54  4f  56   ATOV
45  52  53  4b   ERSK
45  58  53  43   EXSC
47  53  54  20   GST
00  08  00  04   ....
00  04  00  02   ....
00  08  00  04   ....
00  00  00 *02*  ....
7b  f6  50  72   {.Pr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  41   rrrA
45  53  47  41   ESGA
45  53  48  2c   ESH,
01  00  00  00   ....
00  00  00  00   ....
00  43  32  35   .C25
35  50  32  35   5P25
36  51  42  49   6QBI
43               C

Actual:
01  00  00  00   ....
00  00  00  00   ....
01  00  00  00   ....
00  00  00  00   ....
01  00  01  00   ....
00  00  00  00   ....
00  00  00  00   ....
00  00  00  70   ...p
00  43  48  4c   .CHL
4f  07  00  4e   O..N
4f  4e  43  41   ONCA
45  41  44  49   EADI
43  53  4c  4b   CSLK
41  54  4f  56   ATOV
45  52  53  4b   ERSK
45  58  53  43   EXSC
47  53  54  20   GST
00  08  00  04   ....
00  04  00  02   ....
00  08  00  04   ....
00  00  00 *01*  ....
7b  f6  50  72   {.Pr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  72   rrrr
72  72  72  41   rrrA
45  53  47  41   ESGA
45  53  48  2c   ESH,
01  00  00  00   ....
00  00  00  00   ....
00  43  32  35   .C25
35  50  32  35   5P25
36  51  42  49   6QBI
43               C

[Ryan Hamilton, 2013-01-16 22:42:23]

Ouch!  Nice catch.  Can you do something similar in the internal version too?

https://codereview.chromium.org/11956029/diff/1/net/quic/test_tools/mock_cloc...
File net/quic/test_tools/mock_clock.cc (right):

https://codereview.chromium.org/11956029/diff/1/net/quic/test_tools/mock_cloc...
net/quic/test_tools/mock_clock.cc:18: now_as_delta_since_unix_epoch_ =
now_as_delta_since_unix_epoch_.Add(delta);
It appears that now_as_delta_since_unix_epoch_ is simply:

  now_.Subtract(QuicTime())

Can you return this from NowAsDeltaSinceUnixEpoch instead?

[wtc, 2013-01-16 23:00:59]

Please review patch set 2. I will make the same change to the
internal version.

https://codereview.chromium.org/11956029/diff/1/net/quic/test_tools/mock_cloc...
File net/quic/test_tools/mock_clock.cc (right):

https://codereview.chromium.org/11956029/diff/1/net/quic/test_tools/mock_cloc...
net/quic/test_tools/mock_clock.cc:18: now_as_delta_since_unix_epoch_ =
now_as_delta_since_unix_epoch_.Add(delta);

On 2013/01/16 22:42:23, Ryan Hamilton wrote:
> It appears that now_as_delta_since_unix_epoch_ is simply:
> 
>   now_.Subtract(QuicTime())
> 
> Can you return this from NowAsDeltaSinceUnixEpoch instead?

Good idea! Thanks for the suggestion.

[Ryan Hamilton, 2013-01-16 23:02:27]

lgtm

[commit-bot: I haz the power, 2013-01-16 23:05:00]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/wtc@chromium.org/11956029/4

[commit-bot: I haz the power, 2013-01-17 01:31:56]

Change committed as 177297