Retrial of the first step to port file_util::CountFilesCreatedAfter()....

Retrial of the first step to port file_util::CountFilesCreatedAfter().
CountFileCreatedAfter() takes FilePath and Time instead of wstring and FILETIME.

We still need to implement the function for POSIX.
I'll work on the change after this change.

In the previous attempt, it seemed that the test for CountFilesCreatedAfter() failed due to discrepancy between system clock and Time::Now() on the build bot.
Time::Now() doesn't always use system clock for better performance and the difference can be large on virtual machines.
I added Time::NowFromSystemTime() to solve this issue.

BUG=9833

[hamaji, 2009-04-16 14:02:53]

Automated tests succeeded 4 times with this change.

[Erik does not do reviews, 2009-04-16 16:25:36]

LGTM with a few nits

http://codereview.chromium.org/75033/diff/1/7
File base/time_mac.cc (right):

http://codereview.chromium.org/75033/diff/1/7#newcode45
Line 45: return Now();
add a comment explaining why this is just pointing at the same implementation

http://codereview.chromium.org/75033/diff/1/5
File base/time_posix.cc (right):

http://codereview.chromium.org/75033/diff/1/5#newcode45
Line 45: return Now();
add a comment explaining why this is the same implementation

http://codereview.chromium.org/75033/diff/1/2
File base/time_win.cc (right):

http://codereview.chromium.org/75033/diff/1/2#newcode137
Line 137: return Now();
Rather than returning Now(), you could just return Time based on initial_time
since you don't need the drift correction.  I'm OK with this as-is if you'd
prefer though since it's effectively equivalent.

[hamaji, 2009-04-16 17:45:11]

Thanks for the review!

http://codereview.chromium.org/75033/diff/1/7
File base/time_mac.cc (right):

http://codereview.chromium.org/75033/diff/1/7#newcode45
Line 45: return Now();
On 2009/04/16 16:25:36, Erik Kay wrote:
> add a comment explaining why this is just pointing at the same implementation

Done.

http://codereview.chromium.org/75033/diff/1/5
File base/time_posix.cc (right):

http://codereview.chromium.org/75033/diff/1/5#newcode45
Line 45: return Now();
On 2009/04/16 16:25:36, Erik Kay wrote:
> add a comment explaining why this is the same implementation

Done.

http://codereview.chromium.org/75033/diff/1/2
File base/time_win.cc (right):

http://codereview.chromium.org/75033/diff/1/2#newcode137
Line 137: return Now();
On 2009/04/16 16:25:36, Erik Kay wrote:
> Rather than returning Now(), you could just return Time based on initial_time
> since you don't need the drift correction.  I'm OK with this as-is if you'd
> prefer though since it's effectively equivalent.

Done.

[Erik does not do reviews, 2009-04-16 18:05:08]

http://codereview.chromium.org/75033/diff/2002/1019
File base/time_win.cc (right):

http://codereview.chromium.org/75033/diff/2002/1019#newcode137
Line 137: return initial_time;
We shouldn't be relying on implicit conversion.  Change to return
Time(initial_time);

[hamaji, 2009-04-16 18:12:14]

Thanks!

http://codereview.chromium.org/75033/diff/2002/1019
File base/time_win.cc (right):

http://codereview.chromium.org/75033/diff/2002/1019#newcode137
Line 137: return initial_time;
On 2009/04/16 18:05:08, Erik Kay wrote:
> We shouldn't be relying on implicit conversion.  Change to return
> Time(initial_time);

Ah, I see. Fixed. I also fixed the return statement in Time::Now().

[Erik does not do reviews, 2009-04-16 19:52:53]

LGTM

I'll try to get this uploaded the bots later today.

[hamaji, 2009-04-17 22:27:43]

On 2009/04/16 19:52:53, Erik Kay wrote:
> LGTM
> 
> I'll try to get this uploaded the bots later today.

Thanks, could you submit this if the test is OK?

Thanks!