[iOS/ARC] Fix a violation of ODR in base::BindBlock under ARC.

[iOS/ARC] Fix a violation of ODR in base::BindBlock under ARC.

The template function base::BindBlock invoked the constructor of
base::mac::ScopedBlock that has two different signature whether
ARC (automatic reference counting) is enabled or not.

If two different files are compiled with a different ARC setting
(one has ARC enabled, the other has it disabled) and both invoke
base::BindBlock with the same type of block (i.e. same signature)
then the function has two different implementation (which is a
violation of the ODR) and one of them is selected at link-time.

As the base::BindBlock implementation does retain the block, this
can only lead to over-retaining an object (as the implementation
leaks if the ARC version of base::mac::ScopedBlock is invoked) not
to over-releasing.

BUG=666027
Committed: https://crrev.com/3baab423ea642cc0851a482841091fd8ac9f8b3c
Cr-Commit-Position: refs/heads/master@{#434182}

[sdefresne, 2016-11-23 13:37:24]

The CQ bit was checked by sdefresne@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-23 13:37:37]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[stkhapugin, 2016-11-23 13:40:17]

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

[stkhapugin, 2016-11-23 13:40:18]

lgtm

[commit-bot: I haz the power, 2016-11-23 14:06:59]

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

[commit-bot: I haz the power, 2016-11-23 14:07:00]

Dry run: This issue passed the CQ dry run.

[sdefresne, 2016-11-23 15:05:19]

The CQ bit was checked by sdefresne@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-23 15:05:38]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[sdefresne, 2016-11-23 15:07:58]

sdefresne@chromium.org changed reviewers:
+ mark@chromium.org, rohitrao@chromium.org

[sdefresne, 2016-11-23 15:07:58]

Please take a look.

The added unit tests fails without the patch as expected (at least one of them
will fail, the second will only fail if the ARC version of the base::BindBlock
template is selected).

[ RUN      ] BindObjcBlockTestARC.TestBlockReleased
../../base/mac/bind_objc_block_unittest_arc.mm:119: Failure
Value of: weak_nsobject == nil
  Actual: false
Expected: true
[  FAILED  ] BindObjcBlockTestARC.TestBlockReleased (0 ms)

[ RUN      ] BindObjcBlockTest.TestBlockReleased
../../base/mac/bind_objc_block_unittest.mm:116: Failure
Value of: nullptr
  Actual: 8-byte object <00-00 00-00 00-00 00-00>
Expected: weak_nsobject
Which is: 96-byte object <30-BB 0D-00 80-60 00-00 00-00 00-00 00-00 00-00 58-54
55-4D 00-00 00-00 00-00 00-00 65-20 00-00 00-00 00-00 58-54 55-4D 00-00 00-00
00-00 00-00 00-04 00-00 00-04 00-00 FF-FF FF-FF FF-FF FF-FF D7-C6 B9-AF 00-80
FF-FF 58-54 55-4D 58-54 55-4D C0-33 30-1A 01-00 00-00 FF-FF FF-FF FF-FF FF-FF>
[  FAILED  ] BindObjcBlockTest.TestBlockReleased (1 ms)

[commit-bot: I haz the power, 2016-11-23 15:36:27]

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

[commit-bot: I haz the power, 2016-11-23 15:36:28]

Dry run: This issue passed the CQ dry run.

[sdefresne, 2016-11-23 15:55:14]

The CQ bit was checked by sdefresne@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-23 15:55:32]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[sdefresne, 2016-11-23 16:08:45]

The CQ bit was checked by sdefresne@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-23 16:09:05]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Mark Mentovai, 2016-11-23 16:22:52]

LGTM

[commit-bot: I haz the power, 2016-11-23 16:32:32]

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

[commit-bot: I haz the power, 2016-11-23 16:32:33]

Dry run: This issue passed the CQ dry run.

[rohitrao (ping after 24h), 2016-11-23 16:32:42]

lgtm

[sdefresne, 2016-11-23 16:35:43]

The CQ bit was checked by sdefresne@chromium.org

[sdefresne, 2016-11-23 16:35:44]

The patchset sent to the CQ was uploaded after l-g-t-m from
stkhapugin@chromium.org
Link to the patchset: https://codereview.chromium.org/2522943003/#ps60001
(title: "Use EXPECT_NSEQ instead of EXPECT_EQ/EXPECT_TRUE.")

[commit-bot: I haz the power, 2016-11-23 16:36:13]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-23 16:41:25]

CQ is committing da patch.

Bot data: {"patchset_id": 60001, "attempt_start_ts": 1479918943877290,
"parent_rev": "a3507ab89b491d40bb04afcd72d3a0302f34b5b2", "commit_rev":
"4c487dcdf5304895f8847e4864b20f5d0ad584fc"}

[commit-bot: I haz the power, 2016-11-23 16:41:51]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2016-11-23 16:45:03]

Description was changed from

==========
[iOS/ARC] Fix a violation of ODR in base::BindBlock under ARC.

The template function base::BindBlock invoked the constructor of
base::mac::ScopedBlock that has two different signature whether
ARC (automatic reference counting) is enabled or not.

If two different files are compiled with a different ARC setting
(one has ARC enabled, the other has it disabled) and both invoke
base::BindBlock with the same type of block (i.e. same signature)
then the function has two different implementation (which is a
violation of the ODR) and one of them is selected at link-time.

As the base::BindBlock implementation does retain the block, this
can only lead to over-retaining an object (as the implementation
leaks if the ARC version of base::mac::ScopedBlock is invoked) not
to over-releasing.

BUG=666027
==========

to

==========
[iOS/ARC] Fix a violation of ODR in base::BindBlock under ARC.

The template function base::BindBlock invoked the constructor of
base::mac::ScopedBlock that has two different signature whether
ARC (automatic reference counting) is enabled or not.

If two different files are compiled with a different ARC setting
(one has ARC enabled, the other has it disabled) and both invoke
base::BindBlock with the same type of block (i.e. same signature)
then the function has two different implementation (which is a
violation of the ODR) and one of them is selected at link-time.

As the base::BindBlock implementation does retain the block, this
can only lead to over-retaining an object (as the implementation
leaks if the ARC version of base::mac::ScopedBlock is invoked) not
to over-releasing.

BUG=666027

Committed: https://crrev.com/3baab423ea642cc0851a482841091fd8ac9f8b3c
Cr-Commit-Position: refs/heads/master@{#434182}
==========

[commit-bot: I haz the power, 2016-11-23 16:45:04]

Patchset 4 (id:??) landed as
https://crrev.com/3baab423ea642cc0851a482841091fd8ac9f8b3c
Cr-Commit-Position: refs/heads/master@{#434182}