Oilpan: tidy up some inter-stack object references.

Oilpan: tidy up some inter-stack object references.

Add uses of STACK_ALLOCATED() so as to allow these objects to safely
keep references to other stack allocated objects.

For other heap allocated objects that keep references to GCed objects,
convert these into appropriate Member<> refs.

R=haraken, philipj@opera.com
BUG=509911
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=200242

[sof, 2015-08-07 07:57:39]

sigbjornf@opera.com changed reviewers:
+ oilpan-reviews@chromium.org

[sof, 2015-08-07 07:57:40]

please take a look.

It would be possible to split these into per-class CLs, if that's preferable...

[haraken, 2015-08-07 08:17:55]

haraken@chromium.org changed reviewers:
+ haraken@chromium.org

[haraken, 2015-08-07 08:17:56]

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
File Source/core/css/parser/CSSParserObserverWrapper.h (right):

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
Source/core/css/parser/CSSParserObserverWrapper.h:42: CSSParserObserver&
m_observer;

Shall we change this to a RawPtrWillBeMember? (I think we have done it in many
places.)

[sof, 2015-08-07 08:20:20]

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
File Source/core/css/parser/CSSParserObserverWrapper.h (right):

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
Source/core/css/parser/CSSParserObserverWrapper.h:42: CSSParserObserver&
m_observer;
On 2015/08/07 08:17:56, haraken wrote:
> 
> Shall we change this to a RawPtrWillBeMember? (I think we have done it in many
> places.)
> 

We can't - it's currently marked as STACK_ALLOCATED().

[haraken, 2015-08-07 08:27:38]

LGTM

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
File Source/core/css/parser/CSSParserObserverWrapper.h (right):

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
Source/core/css/parser/CSSParserObserverWrapper.h:42: CSSParserObserver&
m_observer;
On 2015/08/07 08:20:20, sof wrote:
> On 2015/08/07 08:17:56, haraken wrote:
> > 
> > Shall we change this to a RawPtrWillBeMember? (I think we have done it in
many
> > places.)
> > 
> 
> We can't - it's currently marked as STACK_ALLOCATED().

ah, understand. Maybe it would make more sense to update the plugin so that it
doesn't detect this as a problematic pointer.

[sof, 2015-08-07 09:11:07]

Updated + found another problematic reference in webaudio/

The use STACK_ALLOCATED() in a public/platform/ decl is acceptable by my
reckoning, but please verify.

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
File Source/core/css/parser/CSSParserObserverWrapper.h (right):

https://codereview.chromium.org/1278103002/diff/1/Source/core/css/parser/CSSP...
Source/core/css/parser/CSSParserObserverWrapper.h:42: CSSParserObserver&
m_observer;
On 2015/08/07 08:27:38, haraken wrote:
> On 2015/08/07 08:20:20, sof wrote:
> > On 2015/08/07 08:17:56, haraken wrote:
> > > 
> > > Shall we change this to a RawPtrWillBeMember? (I think we have done it in
> many
> > > places.)
> > > 
> > 
> > We can't - it's currently marked as STACK_ALLOCATED().
> 
> ah, understand. Maybe it would make more sense to update the plugin so that it
> doesn't detect this as a problematic pointer.

Stack references held by stack "hosts" are recognized and supported by the GC
plugin.

This "wrapper" type is indeed always stack allocated, so just annotating it with
STACK_ALLOCATED() is sufficient here. IOW, we can do better :) Updated the CL.

[haraken, 2015-08-07 13:08:38]

LGTM

[sof, 2015-08-07 13:16:46]

sigbjornf@opera.com changed reviewers:
+ jochen@chromium.org, philipj@opera.com

[sof, 2015-08-07 13:16:47]

philipj@, jochen@: could you have a look at public/platform/ change? tia.

[philipj_slow, 2015-08-10 09:56:05]

On 2015/08/07 09:11:07, sof wrote:
> The use STACK_ALLOCATED() in a public/platform/ decl is acceptable by my
> reckoning, but please verify.

I can't find any other cases like this, but LGTM since haraken says it's OK.

[sof, 2015-08-10 10:33:47]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2015-08-10 10:33:49]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1278103002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1278103002/20001

[commit-bot: I haz the power, 2015-08-10 11:48:17]

Committed patchset #2 (id:20001) as
https://src.chromium.org/viewvc/blink?view=rev&revision=200242