Add trait std::vec::IntoVec

[lilyball]

std::vec::IntoVec provides the .into_vec() method, similar to
std::slice::CloneableVector. It does not provide .to_vec(), because that
requires Clone, and can instead just be provided on individual types
separately.

[lilyball]

I marked this #[inline] to match WindowsPath. It seems like something that should be trivially inlined anyway, I'm not sure why it wasn't marked as such before.

[alexcrichton]

We've generally been trying to avoid one-off traits, and this does seem a bit like a one-off trait. I think this should be discussed to make sure this is the direction we want to go in.

[lilyball]

Discussion is fine.

The intent of this trait is to be used where current code uses ClonaeableVector. E.g. I have a function

fnfoo<V:CloneableVector<u8>>(v:V){let v = v.into_owned();// do stuff with v}

In order to convert this over to Vec<u8> instead of ~[u8], I need a way to convert both &[T] and ~[T] into a Vec<T>, which is what IntoVec<T> provides. So the equivalent function is

fnfoo<V:IntoVec<u8>>(v:V){let v = v.into_vec();// do stuff with v}

IntoVec<T> is missing to_vec() (the equivalent to CloneableVector<T>::to_owned()) because that method requires Clone, and into_vec() should be usable on ~[T] and Vec<T> without T: Clone (the impl for &[T] still requires Clone, of course). Also I don't think a .to_vec() trait method is particularly useful.

It's worth noting that std::str::Str contains to_strbuf() and into_strbuf() methods.

[lilyball]

@alexcrichton Do you have anything you want to discuss about this pull? Nobody has chimed in so far with any objections beyond your one point about one-off traits.

[alexcrichton]

I don't have much to add over that this still seems like a one-off trait that doesn't need to go in at this time.

[lilyball]

What if I add .to_vec()/.into_vec() into slice::CloneableVector instead? I didn't do that initially because I thought it was unnecessary to require Clone for .into_vec().

[alexcrichton]

Perhaps we can wait for the dust to settle on Vec<T> and ~[T] before reconsidering this patch?

[lilyball]

Ok. I'm going to close it out for now for bors's sake, and I'll revisit later.