Permalink
Commits on May 12, 2011
  1. Bump version number to 1.1.2.0

    tibbe committed May 12, 2011
  2. Update change log

    tibbe committed May 12, 2011
  3. Merge pull request #15 from bos/master

    Add instances for Integer and Ratio
    tibbe committed May 12, 2011
Commits on May 11, 2011
Commits on May 10, 2011
Commits on May 7, 2011
  1. Add instances for Float and Double

    Implementation contributed by Lennart Augustsson.
    tibbe committed May 7, 2011
Commits on Mar 28, 2011
Commits on Mar 27, 2011
Commits on Feb 25, 2011
  1. Add explicit hash instances for tuples, as these are (for small

    component types like Int) slightly more efficient than hashWithSalt.
    jmaessen committed Feb 25, 2011
  2. Tuple and [] instances used to yield the exact same hash collisions

    regardless of the salt---ie, the salt simply permuted the hash value
    without changing its collision properties, effectively making salting
    a no-op for practical purposes when hashing tuples and lists.  Fixed
    by defining tuple instances in terms of hashWithSalt rather than hash
    and combine.
    jmaessen committed Feb 25, 2011
Commits on Feb 22, 2011
  1. Comment out test-suite section until supported by Hackage

    The section was accidentally enabled in a previous commit.
    tibbe committed Feb 22, 2011
  2. Update changelog

    tibbe committed Feb 22, 2011
  3. Bump version number to 1.1.1.0

    tibbe committed Feb 22, 2011
  4. Update maintainer email address

    tibbe committed Feb 22, 2011
  5. Mix in an initial seed when hashing tuples

    Since combine has 0 as a left identity it makes sense to start with a
    non-zero initial value to make sure leading fields that hash to 0
    contribute to the final hash value.
    tibbe committed Feb 22, 2011
  6. -Wall police

    tibbe committed Feb 22, 2011
  7. Implement hashWithSalt for [a]

    tibbe committed Feb 22, 2011
  8. Provide hashWithSalt

    tibbe committed Feb 22, 2011
Commits on Feb 21, 2011
  1. Remove the pUpperMatters test

    This test sometimes fails due to genuine hash collisions.  Such hash
    collisions can always occure as we're hashing a 64-bit value into a
    32-bit value.
    
    We probably need some kind of probabilistic test to test for
    collisions.
    tibbe committed Feb 21, 2011
  2. Hash Int64 and Word64 using XOR on 32-bit machines

    The recipe was adopted from Java, where it seems to work well in
    practice.
    tibbe committed Feb 21, 2011
  3. Clean up tests

    tibbe committed Feb 21, 2011
Commits on Feb 19, 2011
  1. fix 'text' properties in tests

    aslatter committed Feb 19, 2011
Commits on Feb 18, 2011
  1. Tweak Cabal metadata

    tibbe committed Feb 18, 2011
  2. Bumped version number to 1.1.0.0

    tibbe committed Feb 18, 2011
  3. Updated changelog

    tibbe committed Feb 18, 2011
Commits on Feb 14, 2011
Commits on Feb 12, 2011
  1. Bump version

    committed Feb 12, 2011
Commits on Feb 6, 2011