Closed N-Dekker closed 1 week ago
It appears more complicated for Array
than it was for Array2D
, to implement a fast noexcept move-assignment operator. Because unlike Array2D
, Array
does have a m_LetArrayManageMemory
flag. While its base class vnl_vector
also has a m_LetArrayManageMemory
flag. These two flags may have different (Boolean) values, for one and the same object. To be continued...!
Update, I find it too complicated to add move semantics on top of optional internal memory management, for an object that has two possibly contradicting (!) m_LetArrayManageMemory
flags. (One of Array
, and one of its base class vnl_vector
.) So I give up for now 🤷
noexcept
"move semantics" toArray2D
"