Alphare added a comment.
In D7116#104685 <https://phab.mercurial-scm.org/D7116#104685>, @kevincox wrote: > I've seen very good results with https://github.com/servo/rust-fnv in the past so it is probably worth including that in the comparison and possibly using it. It is especially good for small keys which seems like a common case in hg. The following comparison shows that for input > 20 bytes, fnv has worse overall performance than xx: https://cglab.ca/~abeinges/blah/hash-rs/. I think that keys bigger than 20 bytes are pretty common: `mercurial/dirstate.py` is nested once and is already over 20 bytes. Maybe we could use fnv for cases other than the dirstate where we have mostly small keys, but that seems pretty cumbersome to me. REPOSITORY rHG Mercurial CHANGES SINCE LAST ACTION https://phab.mercurial-scm.org/D7116/new/ REVISION DETAIL https://phab.mercurial-scm.org/D7116 To: Alphare, #hg-reviewers, kevincox Cc: durin42, kevincox, mercurial-devel _______________________________________________ Mercurial-devel mailing list Mercurial-devel@mercurial-scm.org https://www.mercurial-scm.org/mailman/listinfo/mercurial-devel