Restore ActivityLogging=SetterForIsolatedWorlds for mouse and keyboard events

Restore ActivityLogging=SetterForIsolatedWorlds for mouse and keyboard events

This was first added in r154552.

In r159511 [PerWorldBindings, ActivityLogging=SetterForIsolatedWorlds]
was replaced with [ActivityLogging=Setter] on the assumption that
PerWorldBindings wasn't there for any particular reason and that
SetterForIsolatedWorlds wouldn't make sense without it.

This was wrong, because activity logging now happens for all worlds,
where it previously only happened for isolated worlds.

Make it so again.

BUG=305112
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=161265

[philipj_slow, 2013-11-03 21:26:16]

A bit of an oops here, better fix this before the branch point in case logging
in all worlds could ruin the ActivityLogging experiment.

The tests for this were added in https://codereview.chromium.org/19540019 but
they apparently didn't start failing when main world access starting logging,
but then I guess you usually don't test all the things that *shouldn't* happen.

[philipj_slow, 2013-11-03 21:33:47]

I should also point out that after this change there isn't any code left using
SetterForAllWorlds, in case that should be purged from the IDL generator as
well.

[abarth-chromium, 2013-11-04 00:48:08]

What impact does this CL have on binary size?

[philipj_slow, 2013-11-04 07:43:35]

On 2013/11/04 00:48:08, abarth wrote:
> What impact does this CL have on binary size?

Size of content_shell binary on 64-bit Linux release.

Before: 94457576 bytes

After: 94505656 bytes

Delta: 48080 bytes

That's more than I would have guessed, looks like PerWorldBindings should be
used sparingly...

[abarth-chromium, 2013-11-04 17:04:30]

Ok, LGTM.  Thanks!

[commit-bot: I haz the power, 2013-11-04 17:05:03]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/philipj@opera.com/57443002/1

[commit-bot: I haz the power, 2013-11-04 18:38:32]

Change committed as 161265