Add skstd::unique_ptr and use it.

Add skstd::unique_ptr and use it.

TBR=bsalomon@google.com
The one gpu include change is just to fix swap in implementation.

Committed: https://skia.googlesource.com/skia/+/a3434d83cf988f930a73aaca5ba88f6f6fa8f547

[bungeman-skia, 2015-09-03 18:46:33]

The CQ bit was checked by bungeman@google.com to run a CQ dry run

[commit-bot: I haz the power, 2015-09-03 18:46:44]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1330503006/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1330503006/20001

[commit-bot: I haz the power, 2015-09-03 18:48:08]

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

[commit-bot: I haz the power, 2015-09-03 18:48:09]

Dry run: Try jobs failed on following builders:
  Build-Mac10.8-Clang-x86_64-Release-Trybot on client.skia.compile (JOB_FAILED,
http://build.chromium.org/p/client.skia.compile/builders/Build-Mac10.8-Clang-...)

[bungeman-skia, 2015-09-03 21:11:37]

The CQ bit was checked by bungeman@google.com to run a CQ dry run

[commit-bot: I haz the power, 2015-09-03 21:11:45]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1330503006/60001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1330503006/60001

[commit-bot: I haz the power, 2015-09-03 21:13:46]

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

[commit-bot: I haz the power, 2015-09-03 21:13:46]

Dry run: Try jobs failed on following builders:
  Build-Win-MSVC-x86-Debug-Trybot on client.skia.compile (JOB_FAILED,
http://build.chromium.org/p/client.skia.compile/builders/Build-Win-MSVC-x86-D...)

[bungeman-skia, 2015-09-03 23:14:10]

The CQ bit was checked by bungeman@google.com to run a CQ dry run

[commit-bot: I haz the power, 2015-09-03 23:14:22]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1330503006/100001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1330503006/100001

[bungeman-skia, 2015-09-03 23:15:17]

The following comments explain why the various changes were made along the way
so far (up to patch set 6).

https://codereview.chromium.org/1330503006/diff/20001/include/private/SkTLogic.h
File include/private/SkTLogic.h (right):

https://codereview.chromium.org/1330503006/diff/20001/include/private/SkTLogi...
include/private/SkTLogic.h:123: template <typename R, typename... Args> struct
is_function<R __stdcall (Args...)> : true_type {};
As it turns out the compiler on Windows just ignores __stdcall on x86_64
completely, so this type ends up __cdecl anyway, which causes errors. I'm not
sure why it doesn't go into the type anyway, but then get translated afterword.
There are already plenty of of valid types for which no object can be created
with that type anyway. In any event, x86 and x86_64 need to be considered
different here.

https://codereview.chromium.org/1330503006/diff/20001/include/private/SkTLogi...
include/private/SkTLogic.h:142: using type = conditional_t<is_void<T>::value, T,
T&>;
This (and the existing add_rvalue_reference below) are wrong, one cannot check
for void here to avoid creating a reference to void. The reference to void type
cannot be created at all here, regardless if it's actually going to be assigned
to 'type'. Instead some sort of SFINAE needs to be used to detect void.

https://codereview.chromium.org/1330503006/diff/20001/include/private/SkTLogi...
include/private/SkTLogic.h:147: using type = conditional_t<is_void<T>::value ||
is_reference<T>::value, T, T&&>;
I honestly don't remember why I added this 'is_reference<T> test here. The
result should be the same both ways, considering the reference collapsing rules
T& -> T& (current) or T& && -> T& (without).

https://codereview.chromium.org/1330503006/diff/20001/include/private/SkUniqu...
File include/private/SkUniquePtr.h (right):

https://codereview.chromium.org/1330503006/diff/20001/include/private/SkUniqu...
include/private/SkUniquePtr.h:72: /*constexpr*/ compressed_data() {}
This needs to make sure fPtr starts out nullptr. And possibly initialize the
deleter.

https://codereview.chromium.org/1330503006/diff/20001/tests/CPlusPlusEleven.cpp
File tests/CPlusPlusEleven.cpp (right):

https://codereview.chromium.org/1330503006/diff/20001/tests/CPlusPlusEleven.c...
tests/CPlusPlusEleven.cpp:85: skstd::unique_ptr<Moveable, Deleter<Moveable>>
u(new Moveable(), Deleter<Moveable>());
Need to make sure this 'new' is actually deleted.

[bungeman-skia, 2015-09-03 23:15:52]

bungeman@google.com changed reviewers:
+ mtklein@google.com

[bungeman-skia, 2015-09-03 23:15:52]

I wanted to know how it worked, so I wrote one.

[commit-bot: I haz the power, 2015-09-03 23:19:23]

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

[commit-bot: I haz the power, 2015-09-03 23:19:24]

Dry run: This issue passed the CQ dry run.

[mtklein, 2015-09-04 13:25:43]

https://codereview.chromium.org/1330503006/diff/100001/include/private/SkUniq...
File include/private/SkUniquePtr.h (right):

https://codereview.chromium.org/1330503006/diff/100001/include/private/SkUniq...
include/private/SkUniquePtr.h:20: default_delete(const default_delete<U>&)
/*noexcept*/ {}
funky indent?

https://codereview.chromium.org/1330503006/diff/100001/include/private/SkUniq...
include/private/SkUniquePtr.h:181: //explicit operator bool() const noexcept {
what happened here?

[bungeman-skia, 2015-09-04 19:19:39]

The CQ bit was checked by bungeman@google.com to run a CQ dry run

[commit-bot: I haz the power, 2015-09-04 19:19:52]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1330503006/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1330503006/120001

[bungeman-skia, 2015-09-04 19:20:19]

https://codereview.chromium.org/1330503006/diff/100001/include/private/SkUniq...
File include/private/SkUniquePtr.h (right):

https://codereview.chromium.org/1330503006/diff/100001/include/private/SkUniq...
include/private/SkUniquePtr.h:20: default_delete(const default_delete<U>&)
/*noexcept*/ {}
On 2015/09/04 13:25:42, mtklein wrote:
> funky indent?

Done. Got carried away indenting things.

https://codereview.chromium.org/1330503006/diff/100001/include/private/SkUniq...
include/private/SkUniquePtr.h:181: //explicit operator bool() const noexcept {
On 2015/09/04 13:25:42, mtklein wrote:
> what happened here?

I don't want to use 'explicit operator' yet because in VS2013 vc++ doesn't
actually make them explicit. Maybe if we also built on Windows with VS2015's
vc++ (or clang-cl) to keep things honest in Windows specific code this would
work ok. Until then, I'm hesitant to use it. Chromium currently forbids the use
of 'explicit operator' for this reason. However, I wanted to write it and mark
it so it can be 'fixed' easily in the future, and I'm using it below in
operator==(nullptr).

[commit-bot: I haz the power, 2015-09-04 19:25:16]

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

[commit-bot: I haz the power, 2015-09-04 19:25:17]

Dry run: This issue passed the CQ dry run.

[mtklein, 2015-09-04 19:41:23]

lgtm

[bungeman-skia, 2015-09-07 19:40:36]

The CQ bit was checked by bungeman@google.com

[commit-bot: I haz the power, 2015-09-07 19:40:48]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1330503006/120001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1330503006/120001

[commit-bot: I haz the power, 2015-09-07 19:45:56]

Committed patchset #7 (id:120001) as
https://skia.googlesource.com/skia/+/a3434d83cf988f930a73aaca5ba88f6f6fa8f547