bluss

Since we've been so slow to actually push out 1.0, const generics has now landed before 1.0 happened. So now it can be on the checklist. But we have every...

We already have `ArrayVec::from([array literal here])` which serves this purpose quite well, not sure what the benefits of a macro are.

Interesting. Counting expressions -- see the little book of rust macros, or the maplit macros. But is not needed - you can write it to use extend with an exact...

@Alexhuszagh I've thought about this, and 1) To be efficient, that macro needs a method that implements something like this: `ArrayVec::::from_array([1, 2, 3, 4])` which allows starting with a shorter...

Can we do this - don't provide append because it doesn't have a clear error handling story. Users can use extend if they want to do this exact operation already....

Good point, but try-extend can be reserved for (Into-)iterators, and we can't handle them all efficiently in one method without an extra trait. Depending on what trait is needed to...

The policy is that arrayvec 1.0 can not depend on pre-1.0 crates. If I run those rules, that means we could accept this, but the feature will be tossed out...

Ok. I'd implement it like this, if it works to drop direct rand dependency. https://github.com/bluss/arrayvec/compare/quickcheck-0.8?expand=1 Still needs tests. Using a dependency rename with the version should be a good strategy.

Ideally extend shouldn't ptr::write per item, it should optimize better than that, but I suspect the IntoIter implementation is putting limits on how well that can be done. PR welcome,...

This project is in maint mode and there is no feature development (see repo description). Thanks 🙂