24:08 interesting how the audience member suggested "a functor" due to the mistaken adoption of the term to mean "function object" only in the context of C++, and Jason either wasn't aware of that historical accident (or was aware but was making a joke) and assumed the audience member was talking about the functor of category theory
For me best book that i learn moor feature is Bjarn C++ Book and Stepanov paper and books. And don’t need more because will speed too much time for some thing and without efficiency >>>.
Im sure there are few who agree with me, but auto is not a good habit in my opinion. Any type/class should never be long enough to use it for readability, and in general it just destroys readability.
Same here too. Going back to the fundamentals of C++ after living comfortably in the luxury of OO languages for more than 12 years, talks like these brings me both revelations and comfort. And also a small does of anxiety from realizing that I have a LOT to catch up on.
If you don't explicitly delete the default move and copy constructors, the copy assignment op, and move assignment op here, you can reassign the Data object to point to something else without explicitly calling the destructor because Data contains a member of type *T (pointer to a template object) in this case: *data So you could write: int main() { Data x(1, 2, 8, 10, 11); Data y(a, b, c); x = y; //Major memory leak, original values from 'x' still exist because '*data' is a pointer }
After finally being able to use C++11 for a couple of years, I have no desire at all to go back to C++98/03. Great talk, I'm now using most of these without even thinking about how I would have had to do it (if I even could) in 98/03.
Solutions to problems created by solutions to problems created by solutions to problems. Just avoid sugar. It's the root of all evil, that's why Fortran survives.