7 Jul
2020
7 Jul
'20
9:13 p.m.
On 2020-07-07 18:05, Peter Dimov via Boost wrote:
Niall Douglas wrote:
We discussed whether adding an alternative naming for copy_exception() could be called a major change.
It is both major (API change) and potentially breaking (code that already defines its own make_exception_ptr may break).
Not that I'm arguing with classification of this change as a major one, but I wouldn't consider it a breaking one as I don't think adding any symbols to a library namespace (except where explicitly allowed by contract) is a valid use of the library. This is the same stance as the standard C++ library is taking with regard to user's definitions in namespace std.