Time singleton and resolution changes.

Reapply r1633 (which was reverted by r1635), with some additional modifications.  Consolidate all timer resolution code to time.h, and always depend on TimeTicks::Now for 1ms resolution.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=1893

[Dean McNamee, 2008-09-07 16:33:43]

I had originally reverted this do to bugs.webkit.org being down and making all
the unittests fail.

[Evan Martin, 2008-09-07 16:41:27]

http://codereview.chromium.org/1806/diff/1/4
File base/time.h (right):

http://codereview.chromium.org/1806/diff/1/4#newcode439
Line 439: public:
If only NowSingleton needs access to this, maybe you could just make it a
friend?  (I freely admit I don't fully understand this code.)

http://codereview.chromium.org/1806/diff/1/2
File base/time_win.cc (right):

http://codereview.chromium.org/1806/diff/1/2#newcode133
Line 133: // because it returns the number of millisecond since Windows has
started, which
millisecondS

http://codereview.chromium.org/1806/diff/1/2#newcode165
Line 165: TimeDelta rollover_;  // Accumulation of time lost do to rollover.
s/do/due/

[cpu_(ooo_6.6-7.5), 2008-09-07 20:15:55]

http://codereview.chromium.org/1806/diff/1/2
File base/time_win.cc (left):

http://codereview.chromium.org/1806/diff/1/2#oldcode211
Line 211: return TimeTicks(0);  // Broken, we don't guarantee this function
works.
Now that I see this code I wonder if QueryPerformanceFrequency changes in
laptops with speedstep..

[M-A Ruel, 2008-09-08 01:08:14]

lgtm

http://codereview.chromium.org/1806/diff/1/4
File base/time.h (right):

http://codereview.chromium.org/1806/diff/1/4#newcode442
Line 442: static TickFunction tick_function_;
If you want to make it public, move it to the public: section.
Otherwise, as Evan said, you can make the singleton a friend. But from issues I
had with VS2008 and webkit, I'm not sure it's a good idea.

[darin (slow to review), 2008-09-08 18:15:22]

LGTM

http://codereview.chromium.org/1806/diff/210/20
File base/message_loop.cc (right):

http://codereview.chromium.org/1806/diff/210/20#newcode10
Line 10: #include <mmsystem.h>
nit: please kill this include