2 ms·
> ranges should be a language feature, as should span and optional That’s an opinion. As you notice, “It's clear that major changes are preferred as library ch
by Someone 3y ago
> ranges should be a language feature, as should span and optional
That’s an opinion. As you notice, “It's clear that major changes are preferred as library changes not language changes”
The spirit of C/C++ is to put just enough in the language itself to allow programmers to implement such things in them.
- maccard 3y ago> That’s an opinion. That's fair, actually. It's an opinion I feel strongly about, though. Honestly, I think c++'s reliance on library features to patch over deficiencies in the language is a cop out. We knew about the unique pointer abi performance issue a decade ago, and decided that string view should be implemented the same way. Ranges are "the kitchen sink", and cause significant compile time issues whether you want to use them or not. Spending time on libraries like fmt (which is an excellent library) is paying lip service to progress by locking in a 5 year old library (at this point). > The spirit of C/C++ is to put just enough in the language itself to allow programmers to implement such things in them To use your own words against you, that's an opinion. Mine is the spirit of C++ (which were talking about here, not C) is to not rock the boat too hard or everyone will have an opinion.