This is the mail archive of the libstdc++@gcc.gnu.org mailing list for the libstdc++ project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [Patch]New feature to all kind of stl container.vector::data_move() returns vector::data() that can be moved,just as unique_ptr::release() to unique_ptr::get()


On 8 January 2015 at 17:27, Anthony Shipman <als@iinet.net.au> wrote:
> On Fri, 9 Jan 2015 02:52:42 am Jonathan Wakely wrote:
>> >> extern "C" void some_function(const int*, size_t);
>> >> std::vector<int> get_values(size_t n);
>> >> ...
>> >> some_function(get_values(5).data(), 5);
>> >>
>> >> Today this works correctly. With your suggestion the call to data()
>> >> would release ownership of the memory and it would never be freed.
>> >
>> > Why would this happen? Is there some sort of implied operator=(&&) for
>> > the int* type?
>>
>> I don't understand the question.  Why would what happen?
>>
> That the vector loses ownership of its data. When the vector is destroyed it
> would release the memory with the data. What is the move method that would
> take away ownership of the data if data() returned T*&& to some_function()?

Read the rest of the thread you're replying to. In
https://gcc.gnu.org/ml/libstdc++/2015-01/msg00032.html it was
suggested to make calling data() on an rvalue vector release ownership
of the memory it manages. My example was showing how that causes
currently valid code to start leaking memory.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]