On 23 Oct 2025 05:21, Peter Dimov via Boost wrote:
We already did that for core::string_view, why not do it again, the procedure has been a smashing success and of enormous benefit to everyone.
My proposal here is to not have the alternative macro in Core but instead make StaticAssert satisfy more people. You will only have that problem if you keep the macro in Core. So your sarcasm is misplaced. As for string_view, either way it would have been bad. Specifically, having two string_views (three, if you count std::string_view) is bad since it causes all kinds of ambiguities. If anything, I would have preferred if core::string_view didn't exist or if it *replaced* string_view from Utility. But alas, that was not agreed upon and we have what we have. And for Boost users, I believe, it is better to have just one string_view type that they can use themselves and interface with Boost libraries. If people think I'm wrong, go ahead and do what you think is right.