Fix possible timestamp overflow in the MP4 stream parser.

Fix possible timestamp overflow in the MP4 stream parser.

BUG=345888
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=252849

[damienv1, 2014-02-22 00:36:37]

Need to do a rebase.

[DaleCurtis, 2014-02-22 00:42:31]

chunk mismatch again, reupload :(

[damienv1, 2014-02-22 01:34:03]

Should be fine now.

[DaleCurtis, 2014-02-22 01:41:20]

https://codereview.chromium.org/175693004/diff/140001/media/formats/mp4/track...
File media/formats/mp4/track_run_iterator.cc (right):

https://codereview.chromium.org/175693004/diff/140001/media/formats/mp4/track...
media/formats/mp4/track_run_iterator.cc:60: // To avoid overflow, split the
following calculation:
Why not TimeDelta::FromMicroseconds((static_cast<double>(numer) / denom) *
base::Time::kMicrosecondsPerSecond) ?

[damienv1, 2014-02-22 01:53:01]

https://codereview.chromium.org/175693004/diff/140001/media/formats/mp4/track...
File media/formats/mp4/track_run_iterator.cc (right):

https://codereview.chromium.org/175693004/diff/140001/media/formats/mp4/track...
media/formats/mp4/track_run_iterator.cc:60: // To avoid overflow, split the
following calculation:
On 2014/02/22 01:41:20, DaleCurtis wrote:
> Why not TimeDelta::FromMicroseconds((static_cast<double>(numer) / denom) *
> base::Time::kMicrosecondsPerSecond) ?

I was wondering if in some cases you can loose some precision (mantissa for a
double is coded on 52 bits).
I went for the safe solution (which gives exactly the same result as the
previous implementation except that it supports more cases without overflowing).

[acolwell GONE FROM CHROMIUM, 2014-02-22 02:00:35]

Oh.. overflow avoiding integer multiply divide computation. How I have missed
thee. Sure you're more often 64-bit now instead of 32-bit, but I still find you
comforting. I look forward to seeing your cousin the timestamp rollover
resistant comparison soon. :)

LGTM

https://codereview.chromium.org/175693004/diff/140001/media/formats/mp4/track...
File media/formats/mp4/track_run_iterator.cc (right):

https://codereview.chromium.org/175693004/diff/140001/media/formats/mp4/track...
media/formats/mp4/track_run_iterator.cc:60: // To avoid overflow, split the
following calculation:
On 2014/02/22 01:41:20, DaleCurtis wrote:
> Why not TimeDelta::FromMicroseconds((static_cast<double>(numer) / denom) *
> base::Time::kMicrosecondsPerSecond) ?
The conversion to double can lose precision if |numer| is > 52-bits and |denom|
is a small number like 2.

[damienv1, 2014-02-22 23:27:33]

The CQ bit was checked by damienv@chromium.org

[commit-bot: I haz the power, 2014-02-22 23:27:39]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/damienv@chromium.org/175693004/220001

[commit-bot: I haz the power, 2014-02-23 01:34:32]

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

[commit-bot: I haz the power, 2014-02-23 01:34:33]

Retried try job too often on win_rel for step(s) browser_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[damienv1, 2014-02-23 01:36:55]

The CQ bit was checked by damienv@chromium.org

[commit-bot: I haz the power, 2014-02-23 01:37:02]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/damienv@chromium.org/175693004/220001

[commit-bot: I haz the power, 2014-02-23 08:06:01]

Change committed as 252849