Add move constructor in ScopedClosureRunner.

base/callback_helpers.h

Index: base/callback_helpers.h
diff --git a/base/callback_helpers.h b/base/callback_helpers.h
index 860803989f4b45cf9c55f4a55d0ebcd785bb047e..a425e47ddcb1b54aaf6b3cf90551d498cdf0c7fa 100644
--- a/base/callback_helpers.h
+++ b/base/callback_helpers.h
@@ -27,11 +27,12 @@ base::Callback<Sig> ResetAndReturn(base::Callback<Sig>* cb) {
   return ret;
 }
 
-// ScopedClosureRunner is akin to scoped_ptr for Closures. It ensures that the
-// Closure is executed and deleted no matter how the current scope exits.
+// ScopedClosureRunner is akin to std::unique_ptr<> for Closures. It ensures
+// that theClosure is executed no matter how the current scope exits.

[danakj, 2016/06/23 20:19:28]

the Closure

[Sergey Ulanov, 2016/06/23 23:30:32]

Done.

 class BASE_EXPORT ScopedClosureRunner {
  public:
   ScopedClosureRunner();
+  ScopedClosureRunner(ScopedClosureRunner&& other);

[danakj, 2016/06/23 20:19:28]

Any reason you're not defining operator= along with this as the style guide
asks?

Can you order this below the "normal" constructors, and separate it with
whitespace?

[Sergey Ulanov, 2016/06/23 23:30:32]

One small concern I have with operator= is that when assigning a=b the closure
in |a| has to be called. That behavior may be confusing and unexpected.  Should
we require that |a| is empty in that case, e.g. with a DCHECK? WDYT?
Done.

[danakj, 2016/06/24 20:06:30]

I think it makes sense that it does what reset does. Can you add a comment on
the operator= and on Reset() in the header saying that it runs the currently
owned closure?

[Sergey Ulanov, 2016/06/24 23:44:01]

Done.

BTW, As far as I can tell there is only one place where Reset() is called and it
looks to be used used incorrectly there:
https://codesearch.chromium.org/chromium/src/content/browser/compositor/surfa...

   explicit ScopedClosureRunner(const Closure& closure);
   ~ScopedClosureRunner();