Hash function objects have different behaviourSection: 23.2.8 [unord.req] Status: Resolved Submitter: Daniel James Opened: 2016-11-24 Last modified: 2020-09-06
Priority: Not Prioritized
View other active issues in [unord.req].
View all other issues in [unord.req].
View all issues with Resolved status.
Discussion:
In 23.2.8 [unord.req] paragraph 12, it says that the behaviour of operator== is undefined unless the
Hash and Pred function objects respectively have the same behaviour. This makes comparing containers
with randomized hashes with different seeds undefined behaviour, but I think that's a valid use case. It's not much
more difficult to support it when the Hash function objects behave differently. I did a little testing and
both libstdc++ and libc++ appear to support this correctly.
operator== or operator!= on unordered containers is undefined unless the Hash and
Pred function refinement"
[2017-01-27 Telecon]
This is a design issue; send to LEWG
[2018-3-17 Resolved by P0809R0, which was adopted in Jacksonville.]
Proposed resolution:
This wording is relative to N4618.
Change 23.2.8 [unord.req] as indicated:
-12- Two unordered containers
aandbcompare equal ifa.size() == b.size()and, for every equivalent-key group[Ea1, Ea2)obtained froma.equal_range(Ea1), there exists an equivalent-key group[Eb1, Eb2)obtained fromb.equal_range(Ea1), such thatis_permutation(Ea1, Ea2, Eb1, Eb2)returnstrue. Forunordered_setandunordered_map, the complexity ofoperator==(i.e., the number of calls to the==operator of thevalue_type, to the predicate returned bykey_eq(), and to the hasher returned byhash_function()) is proportional toNin the average case and toN2in the worst case, whereNisa.size(). Forunordered_multisetandunordered_multimap, the complexity ofoperator==is proportional to ∑Ei2in the average case and toN2in the worst case, whereNis a.size(), andEiis the size of theith equivalent-key group ina. However, if the respective elements of each corresponding pair of equivalent-key groupsEaiandEbiare arranged in the same order (as is commonly the case, e.g., ifaandbare unmodified copies of the same container), then the average-case complexity forunordered_multisetandunordered_multimapbecomes proportional toN(but worst-case complexity remains 𝒪(N2), e.g., for a pathologically bad hash function). The behavior of a program that usesoperator==oroperator!=on unordered containers is undefined unless theHashandPredfunction objects respectively havehas the same behavior for both containers and the equality comparison operator forKeyis a refinement(footnote 258) of the partition into equivalent-key groups produced byPred.