Peter Dimov
Peter Dimov
Hence `to_address`. As I said, having the user type the `&*` is much more error-prone, not to mention wrong in principle because `&*end()`.
Iterators from deque that pass the assert will actually form a valid span. > Requiring the users to type it makes them aware what they are doing. Only in theory....
> This would be runtime-dependent. Meaning the code will pass or fail from one run to another. When it should just always fail to compile. Yes, but we can't reliably...
> Actually, let's add a new `is_contiguous_iterator` trait that will be usable outside `boost::span`. For pre-C++20 it would be true only for pointers. On second thought, this (`boost::core::is_contiguous_iterator`) is a...
It should be public so that people can specialize it for their iterators (if they insist.)
Not specializing for string and vector makes the whole thing completely useless. We either do it and specialize, or do nothing at all, nothing else makes any sense.
No, they can't. It's not possible to specialize the trait for `std::vector::iterator`. More precisely, it would be technically possible if our trait has an `Enable = void` second parameter, but...
See `boost::core::type_name`.
Bind only depends on Config and Core so not much of a win. Lambda currently has 26 dependencies, Phoenix 32, Beast 55. https://pdimov.github.io/boostdep-report/develop/module-weights.html
Support for `BOOST_BEAST_USE_STD_STRING_VIEW` was removed 2 years ago and I don't think that the maintainers of Beast will be willing to reinstate it, so I don't think you can get...