Add nonstd::unique_ptr move assigment operator.

Add nonstd::unique_ptr move assigment operator.

std::unique_ptr supports move assignment as in:
  ptr2 = std::move(ptr1);

R=jyasskin@chromium.org

Committed: https://pdfium.googlesource.com/pdfium/+/dd7a7f012424ec8505830710ac0dd0183203c189

[Tom Sepez, 2015-09-22 18:28:19]

tsepez@chromium.org changed reviewers:
+ jyasskin@chromium.org, thestig@chromium.org

[Tom Sepez, 2015-09-22 18:28:19]

Jeffrey, for review.  Small omission from our stub impl.

[Jeffrey Yasskin, 2015-09-22 20:12:55]

LGTM given the changes below.

"ptr1.reset(std::move(ptr2))" doesn't work because reset() takes a T*, not a
unique_ptr<T>. It'd be possible to overload it, but the committee doesn't
generally like having two ways to spell a single operation, and
operator=(unique_ptr) is the way to spell this one.

https://codereview.chromium.org/1358163002/diff/1/third_party/base/nonstd_uni...
File third_party/base/nonstd_unique_ptr.h (right):

https://codereview.chromium.org/1358163002/diff/1/third_party/base/nonstd_uni...
third_party/base/nonstd_unique_ptr.h:181: if (that.ptr_ != ptr_) {
Implement this as "reset(that.release());"

https://codereview.chromium.org/1358163002/diff/1/third_party/base/nonstd_uni...
third_party/base/nonstd_unique_ptr.h:206: class unique_ptr<C[]> : public
unique_ptr_base<C> {
Probably keep unique_ptr<C[]> consistent with unique_ptr<C> by adding the move
assignment here too.

[Tom Sepez, 2015-09-22 22:05:13]

https://codereview.chromium.org/1358163002/diff/1/third_party/base/nonstd_uni...
File third_party/base/nonstd_unique_ptr.h (right):

https://codereview.chromium.org/1358163002/diff/1/third_party/base/nonstd_uni...
third_party/base/nonstd_unique_ptr.h:181: if (that.ptr_ != ptr_) {
On 2015/09/22 20:12:55, Jeffrey Yasskin wrote:
> Implement this as "reset(that.release());"

Done. clever.

https://codereview.chromium.org/1358163002/diff/1/third_party/base/nonstd_uni...
third_party/base/nonstd_unique_ptr.h:206: class unique_ptr<C[]> : public
unique_ptr_base<C> {
On 2015/09/22 20:12:55, Jeffrey Yasskin wrote:
> Probably keep unique_ptr<C[]> consistent with unique_ptr<C> by adding the move
> assignment here too.

Done.

[Tom Sepez, 2015-09-22 22:07:03]

Committed patchset #2 (id:20001) manually as
dd7a7f012424ec8505830710ac0dd0183203c189 (presubmit successful).

[Lei Zhang, 2015-09-23 04:50:35]

https://codereview.chromium.org/1358163002/diff/20001/pdfium.gyp
File pdfium.gyp (right):

https://codereview.chromium.org/1358163002/diff/20001/pdfium.gyp#newcode726
pdfium.gyp:726: 'third_party/base/nonstd_unique_ptr_unittest.cpp',
You forgot BUILD.gn, but I got it in my CL.

[Tom Sepez, 2015-09-23 19:12:32]

On 2015/09/23 04:50:35, Lei Zhang wrote:
> https://codereview.chromium.org/1358163002/diff/20001/pdfium.gyp
> File pdfium.gyp (right):
> 
> https://codereview.chromium.org/1358163002/diff/20001/pdfium.gyp#newcode726
> pdfium.gyp:726: 'third_party/base/nonstd_unique_ptr_unittest.cpp',
> You forgot BUILD.gn, but I got it in my CL.

Thanks.