Dedup functionality

[iopq]

I'd like to .sort() and .dedup() the SmallVecs so I can see which values are unique. Failing that, I guess SmallHashSet would work too.

[iopq]

I don't actually know if it would be faster to do .sort(). and .dedup() or to do a .contains() in a loop...

[SimonSapin]

.sort() should Just Work, through DerefMut to [T]. You may also want to try quicksort instead: https://github.com/servo/rust-quicksort it is not a stable sort (which doesn’t matter if you dedup) but it doesn’t allocate.

Vec::dedup can not be used directly, but I think its code can be copied mostly unchanged. It’d just need .truncate(n) to be added. I’d review a PR to do that.

As to which is faster, it depends on how many items you have and probably other things.

[SimonSapin]

It’s often hard to predict what algorithm is gonna be faster on a particular data set, the only way to know is to try and benchmark it.

[McGittyHub]

This is a feature that I would have liked to be able to use in my code.

[mbrubeck]

An implementation of dedup (and also dedup_by and dedup_by_key) would be very welcome. The truncate method was added in #12, so all these methods need to do is swap duplicate elements to the back of the vector and then truncate the vector to drop them. (As mentioned above, you can use the code from std::vec for reference.)

[Thou94]

I would like to work on it, but it's my first time to contribute to an open source project. I would love to get some help, if I stuck somewhere.

[jdm]

@Thou94 A PR to implement those methods has already been submitted in #72.

[mbrubeck]

Fixed by #72, and will be included in version 0.6.0 (#78).