[GROOVY-5785] HashSet removeAll() operations are slow

• Type: Bug
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 2.0.5
• Component/s: groovy-jdk
• Assignee: Unassigned
• Reporter: Uri Moszkowicz
• Created: 05/Nov/12 14:47
• Updated: 08/Nov/12 16:36

Description

Consider the following (simple but not very Groovy) code:

def hashSet1 = (0..<100000) as HashSet
def hashSet2 = (50000..<150000) as HashSet
def hashSet3 = hashSet2 - hashSet1
println hashSet3.size()

This script takes a very long time to complete due to the implementation of the minus() operation used in computing hashSet3 since it is a non-linear operation. By comparison, using Java methods it completes in less than a second:

def hashSet1 = (0..<100000) as HashSet
def hashSet2 = (50000..<150000) as HashSet
def hashSet3 = hashSet2.clone()
hashSet3.removeAll(hashSet1)
println hashSet3.size()

It seems to me that a simple fix is just to call removeAll() in Collection or worst case just for HashSets to solve this problem but it seems too simple (maybe implications for equals() that I'm missing?). If I'm not missing anything, then I'd be happy to create a patch which fixes this performance bug but I just want to make sure that I'm not missing anything.