FindGrp() is somewhat slow as we do not store the untruncated hash value in the groups, so we have to compare all strings (well, we order them on insertion, but

FindGrp() is somewhat slow as we do not store the untruncated hash value in the groups, so we have to compare all strings (well, we order them on insertion, but it's still 2 to 3 on the average - an average 3 integer comparisons would be nicer). But fixing that requires an ABI break, so it has to wait for a chance. Like in a year or something.