Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}

Committed: https://crrev.com/6b2e026b1c11833d558c8fd5c9d427f04ec3757b
Cr-Commit-Position: refs/heads/master@{#384904}

[sof, 2016-04-02 21:34:16]

Description was changed from

==========
Introduce DEFINE_STATIC_PERSISTENT(type, name, arguments)

Provide this macro for defining static Persistent<> references, replacing
uses of DEFINE_STATIC_REF_WILL_BE_PERSISTENT() with it (so as to allow
the latter to be removed in a follow-up.)

R=
BUG=585328
==========

to

==========
Introduce DEFINE_STATIC_PERSISTENT(type, name, arguments)

Provide this macro for defining static Persistent<> references, replacing
uses of DEFINE_STATIC_REF_WILL_BE_PERSISTENT() with it (so as to allow
the latter to be removed in a follow-up.)

R=
BUG=585328
CQ_INCLUDE_TRYBOTS=tryserver.chromium.win:win_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

[sof, 2016-04-03 20:26:00]

Description was changed from

==========
Introduce DEFINE_STATIC_PERSISTENT(type, name, arguments)

Provide this macro for defining static Persistent<> references, replacing
uses of DEFINE_STATIC_REF_WILL_BE_PERSISTENT() with it (so as to allow
the latter to be removed in a follow-up.)

R=
BUG=585328
CQ_INCLUDE_TRYBOTS=tryserver.chromium.win:win_optional_gpu_tests_rel;tryserver.chromium.mac:mac_optional_gpu_tests_rel
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=
BUG=585328
==========

[sof, 2016-04-03 20:41:05]

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

[sof, 2016-04-03 20:41:06]

please take a look.

(Will rebase wrt https://codereview.chromium.org/1851383002/ once it is has
landed.)

[haraken, 2016-04-03 23:46:42]

LGTM. This looks like a nice refactoring!

It would be worth adding a test for DEFINE_STATIC_LOCAL to HeapTest.cpp.

[haraken, 2016-04-03 23:46:51]

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

[haraken, 2016-04-03 23:46:52]

https://codereview.chromium.org/1850413002/diff/80001/third_party/WebKit/Sour...
File third_party/WebKit/Source/wtf/StdLibExtras.h (right):

https://codereview.chromium.org/1850413002/diff/80001/third_party/WebKit/Sour...
third_party/WebKit/Source/wtf/StdLibExtras.h:99: // wrapped up by a strong
off-heap Persistent<T> reference.

Nit: "strong off-heap" might be misleading. "a strong Persistent handle"?

[sof, 2016-04-04 07:09:00]

Added a HeapTest unit test over DEFINE_STATIC_LOCAL()s involving heap objects.

https://codereview.chromium.org/1850413002/diff/80001/third_party/WebKit/Sour...
File third_party/WebKit/Source/wtf/StdLibExtras.h (right):

https://codereview.chromium.org/1850413002/diff/80001/third_party/WebKit/Sour...
third_party/WebKit/Source/wtf/StdLibExtras.h:99: // wrapped up by a strong
off-heap Persistent<T> reference.
On 2016/04/03 23:46:52, haraken wrote:
> 
> Nit: "strong off-heap" might be misleading. "a strong Persistent handle"?
> 

Reworded.

https://codereview.chromium.org/1850413002/diff/100001/third_party/WebKit/Sou...
File third_party/WebKit/Source/wtf/StdLibExtras.h (right):

https://codereview.chromium.org/1850413002/diff/100001/third_party/WebKit/Sou...
third_party/WebKit/Source/wtf/StdLibExtras.h:88: ASSERT(*singleton);
Added this assert to catch out the accidental creation of empty Persistent<>
singletons, i.e., ill-formed uses of DEFINE_STATIC_LOCAL() over persistents
(DEFINE_STATIC_LOCAL(HeapObject, o, ()))

[haraken, 2016-04-04 07:17:31]

LGTM

You can remove the definition of DEFINE_STATIC_REF_WILL_BE_PERSISTENT.

[sof, 2016-04-04 07:19:55]

On 2016/04/04 07:17:31, haraken wrote:
> LGTM
> 
> You can remove the definition of DEFINE_STATIC_REF_WILL_BE_PERSISTENT.

Yes; I won't do that here though.

Once this has landed, all transition types are removable, I think.

[sof, 2016-04-04 07:29:00]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2016-04-04 07:29:05]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1850413002/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1850413002/100001

[commit-bot: I haz the power, 2016-04-04 09:27:47]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-04-04 09:27:49]

Failed to apply patch for
third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp:
While running git apply --index -3 -p1;
  error: patch failed:
third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp:512
  Falling back to three-way merge...
  Applied patch to 'third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp'
with conflicts.
  U third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp

Patch:       third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp
Index: third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp
diff --git a/third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp
b/third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp
index
590b80a3b1b71a7dfad4708b1d26412c4597c1d3..cb2825514e678b45aa77a7ba61d7add5f1795612
100644
--- a/third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp
+++ b/third_party/WebKit/Source/core/svg/graphics/SVGImage.cpp
@@ -469,7 +469,7 @@ bool SVGImage::dataChanged(bool allDataReceived)
         // types.
         EventDispatchForbiddenScope::AllowUserAgentEvents allowUserAgentEvents;
 
-        DEFINE_STATIC_LOCAL(Persistent<FrameLoaderClient>,
dummyFrameLoaderClient, (EmptyFrameLoaderClient::create()));
+        DEFINE_STATIC_LOCAL(FrameLoaderClient, dummyFrameLoaderClient,
(EmptyFrameLoaderClient::create()));
 
         if (m_page) {
            
toLocalFrame(m_page->mainFrame())->loader().load(FrameLoadRequest(0, blankURL(),
SubstituteData(data(), AtomicString("image/svg+xml"),
@@ -512,7 +512,7 @@ bool SVGImage::dataChanged(bool allDataReceived)
         RawPtr<LocalFrame> frame = nullptr;
         {
             TRACE_EVENT0("blink", "SVGImage::dataChanged::createFrame");
-            frame = LocalFrame::create(dummyFrameLoaderClient.get(),
&page->frameHost(), 0);
+            frame = LocalFrame::create(&dummyFrameLoaderClient,
&page->frameHost(), 0);
             frame->setView(FrameView::create(frame.get()));
             frame->init();
         }

[sof, 2016-04-04 10:47:33]

The CQ bit was checked by sigbjornf@opera.com

[sof, 2016-04-04 10:47:34]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/1850413002/#ps120001
(title: "rebased")

[commit-bot: I haz the power, 2016-04-04 10:47:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1850413002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1850413002/120001

[sof, 2016-04-04 12:15:19]

The CQ bit was unchecked by sigbjornf@opera.com

[sof, 2016-04-04 12:15:32]

Description was changed from

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=
BUG=585328
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true
==========

[sof, 2016-04-04 12:15:41]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2016-04-04 12:15:56]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1850413002/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1850413002/120001

[commit-bot: I haz the power, 2016-04-04 12:22:15]

Description was changed from

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true
==========

[commit-bot: I haz the power, 2016-04-04 12:22:16]

Committed patchset #7 (id:120001)

[commit-bot: I haz the power, 2016-04-04 12:23:25]

Description was changed from

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

[commit-bot: I haz the power, 2016-04-04 12:23:26]

Patchset 7 (id:??) landed as
https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}

[sof, 2016-04-04 12:43:53]

A revert of this CL (patchset #7 id:120001) has been created in
https://codereview.chromium.org/1855203002/ by sigbjornf@opera.com.

The reason for reverting is: Don't understand what happened here, but
compilation breakage seen
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac%20Builder/....

[sof, 2016-04-04 13:07:19]

Description was changed from

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

[sof, 2016-04-04 13:13:18]

On 2016/04/04 12:43:53, sof wrote:
> A revert of this CL (patchset #7 id:120001) has been created in
> https://codereview.chromium.org/1855203002/ by mailto:sigbjornf@opera.com.
> 
> The reason for reverting is: Don't understand what happened here, but
> compilation breakage seen
>
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Mac%20Builder/....

Can't reproduce locally (nor can CQ), but StdLibExtras.h failed to IWYU
TypeTraits.h. It really should, hence added.

Relanding.

[sof, 2016-04-04 13:13:39]

The CQ bit was checked by sigbjornf@opera.com

[sof, 2016-04-04 13:13:40]

The patchset sent to the CQ was uploaded after l-g-t-m from haraken@chromium.org
Link to the patchset: https://codereview.chromium.org/1850413002/#ps140001
(title: "address compilation failure")

[sof, 2016-04-04 13:13:42]

The CQ bit was unchecked by sigbjornf@opera.com

[sof, 2016-04-04 13:13:54]

Description was changed from

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328
NOTRY=true

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

[sof, 2016-04-04 13:15:28]

The CQ bit was checked by sigbjornf@opera.com

[commit-bot: I haz the power, 2016-04-04 13:15:38]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1850413002/140001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1850413002/140001

[commit-bot: I haz the power, 2016-04-04 14:44:41]

Description was changed from

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

[commit-bot: I haz the power, 2016-04-04 14:44:42]

Committed patchset #8 (id:140001)

[commit-bot: I haz the power, 2016-04-04 14:46:04]

Description was changed from

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}
==========

to

==========
Improve DEFINE_STATIC_LOCAL()'s handling of Blink GCed objects.

Extend DEFINE_STATIC_LOCAL() to automatically wrap up heap objects
being exposed as static singletons -- the wrapping happening by
way of a Persistent<>.

With that in place, simplify various uses of DEFINE_STATIC_LOCAL()
along with phasing out the use of DEFINE_STATIC_REF_WILL_BE_PERSISTENT()
entirely.

R=haraken
BUG=585328

Committed: https://crrev.com/18dc8ecff5ba68d28fc536f723ae3c57eafa1b4e
Cr-Commit-Position: refs/heads/master@{#384887}

Committed: https://crrev.com/6b2e026b1c11833d558c8fd5c9d427f04ec3757b
Cr-Commit-Position: refs/heads/master@{#384904}
==========

[commit-bot: I haz the power, 2016-04-04 14:46:11]

Patchset 8 (id:??) landed as
https://crrev.com/6b2e026b1c11833d558c8fd5c9d427f04ec3757b
Cr-Commit-Position: refs/heads/master@{#384904}

[Jeffrey Yasskin, 2016-04-05 20:33:14]

A revert of this CL (patchset #8 id:140001) has been created in
https://codereview.chromium.org/1862513004/ by jyasskin@chromium.org.

The reason for reverting is: Caused leaks in
https://build.chromium.org/p/chromium.memory.fyi/builders/Linux%20Tests%20%28...:

Leak_DefinitelyLost
16 bytes in 1 blocks are definitely lost in loss record 346 of 2,713
  malloc (m_replacemalloc/vg_replace_malloc.c:1138)
  WTF::Partitions::fastMalloc(unsigned long, char const*)
(third_party/WebKit/Source/wtf/PartitionAlloc.h:736)
 
blink::PersistentBase<blink::HeapHashSet<blink::WeakMember<blink::FontCacheClient>,
WTF::WeakMemberHash<blink::FontCacheClient>,
WTF::HashTraits<blink::WeakMember<blink::FontCacheClient> > >,
(blink::WeaknessPersistentConfiguration)0,
(blink::CrossThreadnessPersistentConfiguration)0>::operator new(unsigned long)
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)
  blink::fontCacheClients()
(third_party/WebKit/Source/platform/fonts/FontCache.cpp:300)
  blink::FontCache::addClient(blink::FontCacheClient*)
(third_party/WebKit/Source/platform/fonts/FontCache.cpp:308)
  blink::CSSFontSelector::CSSFontSelector(blink::Document*)
(third_party/WebKit/Source/core/css/CSSFontSelector.cpp:56)
  blink::CSSFontSelector::create(blink::Document*)
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)


Leak_DefinitelyLost
16 bytes in 1 blocks are definitely lost in loss record 352 of 2,713
  malloc (m_replacemalloc/vg_replace_malloc.c:1138)
  WTF::Partitions::fastMalloc(unsigned long, char const*)
(third_party/WebKit/Source/wtf/PartitionAlloc.h:736)
  blink::PersistentBase<blink::HeapHashSet<blink::Member<blink::Widget>,
WTF::MemberHash<blink::Widget>, WTF::HashTraits<blink::Member<blink::Widget> >
>, (blink::WeaknessPersistentConfiguration)0,
(blink::CrossThreadnessPersistentConfiguration)0>::operator new(unsigned long)
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)
  blink::widgetsPendingTemporaryRemovalFromParent()
(third_party/WebKit/Source/core/html/HTMLFrameOwnerElement.cpp:53)
 
blink::HTMLFrameOwnerElement::UpdateSuspendScope::performDeferredWidgetTreeOperations()
(third_party/WebKit/Source/core/html/HTMLFrameOwnerElement.cpp:91)
  blink::HTMLFrameOwnerElement::UpdateSuspendScope::~UpdateSuspendScope()
(third_party/WebKit/Source/core/html/HTMLFrameOwnerElement.cpp:103)


Leak_DefinitelyLost
16 bytes in 1 blocks are definitely lost in loss record 347 of 2,713
  malloc (m_replacemalloc/vg_replace_malloc.c:1138)
  WTF::Partitions::fastMalloc(unsigned long, char const*)
(third_party/WebKit/Source/wtf/PartitionAlloc.h:736)
  blink::PersistentBase<blink::HeapHashSet<blink::WeakMember<blink::Document>,
WTF::WeakMemberHash<blink::Document>,
WTF::HashTraits<blink::WeakMember<blink::Document> > >,
(blink::WeaknessPersistentConfiguration)0,
(blink::CrossThreadnessPersistentConfiguration)0>::operator new(unsigned long)
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)
  blink::Document::liveDocumentSet()
(third_party/WebKit/Source/core/dom/Document.cpp:366)
  blink::Document::Document(blink::DocumentInit const&, unsigned char)
(third_party/WebKit/Source/core/dom/Document.cpp:509)
  blink::HTMLDocument::HTMLDocument(blink::DocumentInit const&, unsigned char)
(third_party/WebKit/Source/core/html/HTMLDocument.cpp:71)
.

[sof, 2016-04-05 20:43:59]

On 2016/04/05 20:33:14, Jeffrey Yasskin wrote:
> A revert of this CL (patchset #8 id:140001) has been created in
> https://codereview.chromium.org/1862513004/ by mailto:jyasskin@chromium.org.
> 
> The reason for reverting is: Caused leaks in
>
https://build.chromium.org/p/chromium.memory.fyi/builders/Linux%20Tests%20%28...:
> 
> Leak_DefinitelyLost
> 16 bytes in 1 blocks are definitely lost in loss record 346 of 2,713
>   malloc (m_replacemalloc/vg_replace_malloc.c:1138)
>   WTF::Partitions::fastMalloc(unsigned long, char const*)
> (third_party/WebKit/Source/wtf/PartitionAlloc.h:736)
>  
>
blink::PersistentBase<blink::HeapHashSet<blink::WeakMember<blink::FontCacheClient>,
> WTF::WeakMemberHash<blink::FontCacheClient>,
> WTF::HashTraits<blink::WeakMember<blink::FontCacheClient> > >,
> (blink::WeaknessPersistentConfiguration)0,
> (blink::CrossThreadnessPersistentConfiguration)0>::operator new(unsigned long)
>
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)
>   blink::fontCacheClients()
> (third_party/WebKit/Source/platform/fonts/FontCache.cpp:300)
>   blink::FontCache::addClient(blink::FontCacheClient*)
> (third_party/WebKit/Source/platform/fonts/FontCache.cpp:308)
>   blink::CSSFontSelector::CSSFontSelector(blink::Document*)
> (third_party/WebKit/Source/core/css/CSSFontSelector.cpp:56)
>   blink::CSSFontSelector::create(blink::Document*)
>
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)
> 
> 
> Leak_DefinitelyLost
> 16 bytes in 1 blocks are definitely lost in loss record 352 of 2,713
>   malloc (m_replacemalloc/vg_replace_malloc.c:1138)
>   WTF::Partitions::fastMalloc(unsigned long, char const*)
> (third_party/WebKit/Source/wtf/PartitionAlloc.h:736)
>   blink::PersistentBase<blink::HeapHashSet<blink::Member<blink::Widget>,
> WTF::MemberHash<blink::Widget>, WTF::HashTraits<blink::Member<blink::Widget> >
> >, (blink::WeaknessPersistentConfiguration)0,
> (blink::CrossThreadnessPersistentConfiguration)0>::operator new(unsigned long)
>
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)
>   blink::widgetsPendingTemporaryRemovalFromParent()
> (third_party/WebKit/Source/core/html/HTMLFrameOwnerElement.cpp:53)
>  
>
blink::HTMLFrameOwnerElement::UpdateSuspendScope::performDeferredWidgetTreeOperations()
> (third_party/WebKit/Source/core/html/HTMLFrameOwnerElement.cpp:91)
>   blink::HTMLFrameOwnerElement::UpdateSuspendScope::~UpdateSuspendScope()
> (third_party/WebKit/Source/core/html/HTMLFrameOwnerElement.cpp:103)
> 
> 
> Leak_DefinitelyLost
> 16 bytes in 1 blocks are definitely lost in loss record 347 of 2,713
>   malloc (m_replacemalloc/vg_replace_malloc.c:1138)
>   WTF::Partitions::fastMalloc(unsigned long, char const*)
> (third_party/WebKit/Source/wtf/PartitionAlloc.h:736)
>   blink::PersistentBase<blink::HeapHashSet<blink::WeakMember<blink::Document>,
> WTF::WeakMemberHash<blink::Document>,
> WTF::HashTraits<blink::WeakMember<blink::Document> > >,
> (blink::WeaknessPersistentConfiguration)0,
> (blink::CrossThreadnessPersistentConfiguration)0>::operator new(unsigned long)
>
(/mnt/data/b/build/slave/chromium-rel-linux-valgrind-tests-2/build/src/out/Release/extensions_unittests)
>   blink::Document::liveDocumentSet()
> (third_party/WebKit/Source/core/dom/Document.cpp:366)
>   blink::Document::Document(blink::DocumentInit const&, unsigned char)
> (third_party/WebKit/Source/core/dom/Document.cpp:509)
>   blink::HTMLDocument::HTMLDocument(blink::DocumentInit const&, unsigned char)
> (third_party/WebKit/Source/core/html/HTMLDocument.cpp:71)
> .

Hmm; the Persistent<> wrappers to these static singletons aren't leaks, really.
And that CL didn't fundamentally change the situation as Persistent<>s were used
before, so I don't get what memcheck is picking up on now.