Re-enable PRESUBMIT check for empty unique_ptr<> rvalue

Re-enable PRESUBMIT check for empty unique_ptr<> rvalue

ui/PRESUBMIT.py contained a check for constructing an empty unique_ptr<>
as an rvalue. It's almost always better to use nullptr instead. However, this
check was disabled because it was overly broad, matching things like

  std::vector<std::unique_ptr<T>>()

which cannot be replaced by nullptr.

Make the check stricter so that it will not overmatch, and re-enable it.

Also fix some violations of the rule that were found during testing.

BUG=

Committed: https://crrev.com/e3646b0b4f464ac2a648ec228f04ec2cf86141de
Cr-Commit-Position: refs/heads/master@{#416865}

[Adam Rice, 2016-09-05 08:13:45]

Patchset #1 (id:1) has been deleted

[Adam Rice, 2016-09-05 08:13:54]

Patchset #1 (id:20001) has been deleted

[Adam Rice, 2016-09-05 08:19:14]

Description was changed from

==========
Re-enable check for empty unique_ptr

WRITEME

COMMIT=no



BUG=
==========

to

==========
Re-enable PRESUBMIT check for empty unique_ptr<> rvalue

ui/PRESUBMIT.py contained a check for constructing an empty unique_ptr<>
as an rvalue. It's almost always better to use nullptr instead. However, this
check was disabled because it was overly broad, matching things like

  std::vector<std::unique_ptr<T>>()

which cannot be replaced by nullptr.

Make the check stricter so that it will not overmatch, and re-enable it.

Also fix some violations of the rule that were found during testing.

BUG=
==========

[Adam Rice, 2016-09-05 08:19:15]

ricea@chromium.org changed reviewers:
+ sky@chromium.org

[sky, 2016-09-06 18:19:17]

Did you make sure this handles the case that lead me to revert the change in the
first place?

https://codereview.chromium.org/2311783002/diff/40001/ui/PRESUBMIT.py
File ui/PRESUBMIT.py (right):

https://codereview.chromium.org/2311783002/diff/40001/ui/PRESUBMIT.py#newcode36
ui/PRESUBMIT.py:36: # TODO(sky): this incorrectly catches templates. Fix and
reenable.
Nuke.

https://codereview.chromium.org/2311783002/diff/40001/ui/PRESUBMIT.py#newcode38
ui/PRESUBMIT.py:38: # std::unique_ptr<T>()
Update

[Adam Rice, 2016-09-07 02:33:01]

On 2016/09/06 18:19:17, sky wrote:
> Did you make sure this handles the case that lead me to revert the change in
the
> first place?

It was the following line in ui/views/mus/native_widget_view_manager.cc:
   scoped_ptr<ui::Event> ui_event(event.To<scoped_ptr<ui::Event>>());

The code in question no longer exists, but I have verified that it passes the
new
check.

[Adam Rice, 2016-09-07 02:44:09]

https://codereview.chromium.org/2311783002/diff/40001/ui/PRESUBMIT.py
File ui/PRESUBMIT.py (right):

https://codereview.chromium.org/2311783002/diff/40001/ui/PRESUBMIT.py#newcode36
ui/PRESUBMIT.py:36: # TODO(sky): this incorrectly catches templates. Fix and
reenable.
On 2016/09/06 18:19:17, sky wrote:
> Nuke.

Done.

https://codereview.chromium.org/2311783002/diff/40001/ui/PRESUBMIT.py#newcode38
ui/PRESUBMIT.py:38: # std::unique_ptr<T>()
On 2016/09/06 18:19:17, sky wrote:
> Update

I don't know how to update this. It looks correct to me.

[sky, 2016-09-07 03:43:35]

LGTM

[Adam Rice, 2016-09-07 05:29:06]

The CQ bit was checked by ricea@chromium.org

[commit-bot: I haz the power, 2016-09-07 05:29:30]

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, 2016-09-07 06:32:36]

Description was changed from

==========
Re-enable PRESUBMIT check for empty unique_ptr<> rvalue

ui/PRESUBMIT.py contained a check for constructing an empty unique_ptr<>
as an rvalue. It's almost always better to use nullptr instead. However, this
check was disabled because it was overly broad, matching things like

  std::vector<std::unique_ptr<T>>()

which cannot be replaced by nullptr.

Make the check stricter so that it will not overmatch, and re-enable it.

Also fix some violations of the rule that were found during testing.

BUG=
==========

to

==========
Re-enable PRESUBMIT check for empty unique_ptr<> rvalue

ui/PRESUBMIT.py contained a check for constructing an empty unique_ptr<>
as an rvalue. It's almost always better to use nullptr instead. However, this
check was disabled because it was overly broad, matching things like

  std::vector<std::unique_ptr<T>>()

which cannot be replaced by nullptr.

Make the check stricter so that it will not overmatch, and re-enable it.

Also fix some violations of the rule that were found during testing.

BUG=
==========

[commit-bot: I haz the power, 2016-09-07 06:32:37]

Committed patchset #2 (id:60001)

[commit-bot: I haz the power, 2016-09-07 06:34:22]

Description was changed from

==========
Re-enable PRESUBMIT check for empty unique_ptr<> rvalue

ui/PRESUBMIT.py contained a check for constructing an empty unique_ptr<>
as an rvalue. It's almost always better to use nullptr instead. However, this
check was disabled because it was overly broad, matching things like

  std::vector<std::unique_ptr<T>>()

which cannot be replaced by nullptr.

Make the check stricter so that it will not overmatch, and re-enable it.

Also fix some violations of the rule that were found during testing.

BUG=
==========

to

==========
Re-enable PRESUBMIT check for empty unique_ptr<> rvalue

ui/PRESUBMIT.py contained a check for constructing an empty unique_ptr<>
as an rvalue. It's almost always better to use nullptr instead. However, this
check was disabled because it was overly broad, matching things like

  std::vector<std::unique_ptr<T>>()

which cannot be replaced by nullptr.

Make the check stricter so that it will not overmatch, and re-enable it.

Also fix some violations of the rule that were found during testing.

BUG=

Committed: https://crrev.com/e3646b0b4f464ac2a648ec228f04ec2cf86141de
Cr-Commit-Position: refs/heads/master@{#416865}
==========

[commit-bot: I haz the power, 2016-09-07 06:34:23]

Patchset 2 (id:??) landed as
https://crrev.com/e3646b0b4f464ac2a648ec228f04ec2cf86141de
Cr-Commit-Position: refs/heads/master@{#416865}