Avoid heap allocations with GestureEventDataPacket

Avoid heap allocations with GestureEventDataPacket

There are typically only a few events in a given gesture packet.  Avoid heap
allocating in the common case by switching from a raw vector to a StackVector.

BUG=395168
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=284728

[jdduke (slow), 2014-07-19 14:48:31]

tdresser@: PTAL (didn't realize we had StackVector until today...).  As a
follow-up I'm planning to add some basic perf tests that exercise the GR,
running it through a sequence of gestures. I imagine std::vector was just fine
here, but I've become paranoid about heap allocations, especially after seeing
wins from switching ui::LatencyInfo to use SmallMap.

[tdresser, 2014-07-21 12:50:57]

On 2014/07/19 14:48:31, jdduke wrote:
> tdresser@: PTAL (didn't realize we had StackVector until today...).  As a
> follow-up I'm planning to add some basic perf tests that exercise the GR,
> running it through a sequence of gestures. I imagine std::vector was just fine
> here, but I've become paranoid about heap allocations, especially after seeing
> wins from switching ui::LatencyInfo to use SmallMap.

LGTM.

Probably not worth merging back to M37 though - we can manage with an
std::vector for one release.

[jdduke (slow), 2014-07-21 15:56:33]

sadrul@: Owner review for ui/events/BUILD.gn? Thanks.

[tdresser, 2014-07-21 16:03:00]

On 2014/07/21 15:56:33, jdduke wrote:
> sadrul@: Owner review for ui/events/BUILD.gn? Thanks.

Test LGTM.

[sadrul, 2014-07-22 16:51:33]

lgtm

[jdduke (slow), 2014-07-22 17:35:41]

The CQ bit was checked by jdduke@chromium.org

[commit-bot: I haz the power, 2014-07-22 17:36:53]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/jdduke@chromium.org/403173004/60001

[commit-bot: I haz the power, 2014-07-22 18:20:09]

Change committed as 284728