Oilpan: convert remaining OwnPtr EventQueue uses.

Oilpan: convert remaining OwnPtr EventQueue uses.

R=haraken@chromium.org
BUG=

Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=176100

[sof, 2014-06-13 06:31:57]

Please take a look.

[haraken, 2014-06-13 06:35:55]

LGTM.

I hope the plugin detects this kind of error. (As we discussed before, compilers
cannot detect OwnPtrs to on-heap objects.)

[sof, 2014-06-13 06:37:40]

The CQ bit was checked by sigbjornf@opera.com

[Mads Ager (chromium), 2014-06-13 06:38:49]

LGTM

This is in an anonymous namespace which is probably why the currently used
plugin does not see it. It probably doesn't run on this code at all. Ian?

[sof, 2014-06-13 06:39:12]

On 2014/06/13 06:35:55, haraken wrote:
> LGTM.
> 
> I hope the plugin detects this kind of error. (As we discussed before,
compilers
> cannot detect OwnPtrs to on-heap objects.)

The oilpan bots were returning red for the CL in Q, so an alternative is for
them to (eventually) be a blocker.

[commit-bot: I haz the power, 2014-06-13 06:39:15]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/sigbjornf@opera.com/336693006/1

[tkent, 2014-06-13 07:01:17]

lgtm.  Thank you!

[commit-bot: I haz the power, 2014-06-13 07:35:16]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/11876)

[commit-bot: I haz the power, 2014-06-13 08:41:01]

Change committed as 176100

[zerny-chromium, 2014-06-13 08:51:07]

On 2014/06/13 06:38:49, Mads Ager (chromium) wrote:
> LGTM
> 
> This is in an anonymous namespace which is probably why the currently used
> plugin does not see it. It probably doesn't run on this code at all. Ian?

I'm not sure what made it into the latest clang roll, but yes, the field decls
of GC managed objects using OwnPtr should be reported. I'll investigate locally.

[zerny-chromium, 2014-06-13 09:24:26]

On 2014/06/13 08:51:07, zerny-chromium wrote:
> On 2014/06/13 06:38:49, Mads Ager (chromium) wrote:
> > LGTM
> > 
> > This is in an anonymous namespace which is probably why the currently used
> > plugin does not see it. It probably doesn't run on this code at all. Ian?
> 
> I'm not sure what made it into the latest clang roll, but yes, the field decls
> of GC managed objects using OwnPtr should be reported. I'll investigate
locally.

Ok. These are not checked because they are not inside a "checked namespace",
which currently includes blink, WebCore and WebKit (and any of their nested
namespaces, also anonymous). These uses are not in such a namespace :-(
Including such in the checking requires a larger change to the plugin in
resolving the various names. I think we should make those changes so that the
plugin can safely run on "non blink" code now that we are sure that the plugin
does not add any significant compile-time overhead.

[zerny-chromium, 2014-06-13 09:29:18]

On 2014/06/13 09:24:26, zerny-chromium wrote:
> On 2014/06/13 08:51:07, zerny-chromium wrote:
> > On 2014/06/13 06:38:49, Mads Ager (chromium) wrote:
> > > LGTM
> > > 
> > > This is in an anonymous namespace which is probably why the currently used
> > > plugin does not see it. It probably doesn't run on this code at all. Ian?
> > 
> > I'm not sure what made it into the latest clang roll, but yes, the field
decls
> > of GC managed objects using OwnPtr should be reported. I'll investigate
> locally.
> 
> Ok. These are not checked because they are not inside a "checked namespace",
> which currently includes blink, WebCore and WebKit (and any of their nested
> namespaces, also anonymous). These uses are not in such a namespace :-(
> Including such in the checking requires a larger change to the plugin in
> resolving the various names. I think we should make those changes so that the
> plugin can safely run on "non blink" code now that we are sure that the plugin
> does not add any significant compile-time overhead.

As Mads suggested off-line, we could at least wrap these tests in an outer
namespace WebCore to allow the current plugin to check them for consistency.