Add isSubsetOf

[rwbarton]

Data.Set has a fuction isSubsetOf :: Ord a => Set a -> Set a -> Bool. I needed the corresponding function for Data.HashSet, and it should be possible to implement it more efficiently than can be done through the public interface of Data.HashSet currently.

[bjin]

I'm also interested in having a isSubmapOf function for HashMap, just like the one for Map and IntMap. I have a few ideas but not sure which one is the best way to implement it. @tibbe

1. Implement isSubmapOf recursively over two HashMap, perform case-by-case analysis like current implementation of unionWith. This is complicated to implement and quite error-prone. I also have concern regarding commit df6c50e. What's the current status of the issue in this commit? Is BitmapIndexed guaranteed to have more than 1 and less than full children now?
2. Implement isSubmapOf like current implementation of (==), transform two HashMap into two lists of Leafs and Collisions with toList' (see df6c50e). The left is linear scanning and case analysis with only Leaf and Collision. The problem with this approach is that in order to run in strictly linear complexity, we need to change the bits in top layer from least-significant to most-significant, this way the keys in the list returned by toList' is sorted in ordinary Word order. This might hurt the performance if the range of hash key is not a full Word.

[kozross]

I would very much like a subset-testing function too.

[treeowl]

Yes, this seems like a good idea. Let's do it the fastest way. I've been wanting to add mergeA for a while; can we do that and use it for this? I don't remember if it has the right shape to do the job efficiently.

[sjakobi]

isSubsetOf was added in #282 and released in v0.2.12.0.