Oilpan: move HTMLParserScheduler to the heap.

Oilpan: move HTMLParserScheduler to the heap.

The benefit being that the back reference it keeps to
HTMLDocumentParser can be traced per normal.

R=kouhei,haraken
BUG=553613
Committed: https://crrev.com/9e418562ced9821f4b8a31650ba4ef721a4845af
Cr-Commit-Position: refs/heads/master@{#359824}

[sof, 2015-11-12 10:53:58]

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

[kouhei (in TOK), 2015-11-12 12:42:53]

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

[kouhei (in TOK), 2015-11-12 12:42:53]

lgtm

[haraken, 2015-11-12 14:20:49]

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

[haraken, 2015-11-12 14:20:49]

https://codereview.chromium.org/1438193002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/html/parser/HTMLParserScheduler.h (right):

https://codereview.chromium.org/1438193002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/html/parser/HTMLParserScheduler.h:107:
RawPtrWillBeUntracedMember<HTMLDocumentParser> m_parser;

A better option would be to move HTMLParserScheduler to the heap.

If that is not possible for some reason, we should use WeakPersistent for this
kind of raw pointer. (I think that that was the conclusion of our previous
discussion about WeakPersistent vs. UntracedMember, if I'm not missing the
point.)

[sof, 2015-11-16 10:56:41]

Description was changed from

==========
Oilpan: account for HTMLParserScheduler::m_parser back ref.

R=
BUG=553613
==========

to

==========
Oilpan: move HTMLParserScheduler to the heap.

The benefit being that the back reference it keeps to
HTMLDocumentParser can be traced per normal.

R=
BUG=553613
==========

[sof, 2015-11-16 10:58:39]

https://codereview.chromium.org/1438193002/diff/1/third_party/WebKit/Source/c...
File third_party/WebKit/Source/core/html/parser/HTMLParserScheduler.h (right):

https://codereview.chromium.org/1438193002/diff/1/third_party/WebKit/Source/c...
third_party/WebKit/Source/core/html/parser/HTMLParserScheduler.h:107:
RawPtrWillBeUntracedMember<HTMLDocumentParser> m_parser;
On 2015/11/12 14:20:49, haraken wrote:
> 
> A better option would be to move HTMLParserScheduler to the heap.
> 
> If that is not possible for some reason, we should use WeakPersistent for this
> kind of raw pointer. (I think that that was the conclusion of our previous
> discussion about WeakPersistent vs. UntracedMember, if I'm not missing the
> point.)

I think that's feasible, the CancellableTaskFactory it keeps have been
problematic in the past, but should now work just fine.

ps#2 does this.

[haraken, 2015-11-16 10:59:49]

LGTM

[sof, 2015-11-16 11:02:42]

Description was changed from

==========
Oilpan: move HTMLParserScheduler to the heap.

The benefit being that the back reference it keeps to
HTMLDocumentParser can be traced per normal.

R=
BUG=553613
==========

to

==========
Oilpan: move HTMLParserScheduler to the heap.

The benefit being that the back reference it keeps to
HTMLDocumentParser can be traced per normal.

R=kouhei,haraken
BUG=553613
==========

[sof, 2015-11-16 11:02:48]

The CQ bit was checked by sigbjornf@opera.com

[sof, 2015-11-16 11:02:48]

The patchset sent to the CQ was uploaded after l-g-t-m from kouhei@chromium.org
Link to the patchset: https://codereview.chromium.org/1438193002/#ps20001
(title: "move the object to the heap instead")

[commit-bot: I haz the power, 2015-11-16 11:03:25]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1438193002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1438193002/20001

[commit-bot: I haz the power, 2015-11-16 12:10:59]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-11-16 12:11:51]

Patchset 2 (id:??) landed as
https://crrev.com/9e418562ced9821f4b8a31650ba4ef721a4845af
Cr-Commit-Position: refs/heads/master@{#359824}