Revision history for Perl extension Math::SimpleHisto::XS.

1.22  Wed Sep 28 21:55:00 2011
  - Fix t/80named.t if there's no JSON implementation.

1.21  Wed Sep 28 21:55:00 2011
  - to_soot() method will create a SOOT TH1D object from the
    Math::SimpleHisto::XS histogram IF the optional SOOT library
    is available.

1.20  Tue Sep 27 17:43:00 2011
  - Include Math::SimpleHisto::XS::Named implementing histograms with
    named bins.
  - Implements a "fill_by_bin" method that fills into a given bin
    number instead of locating the target bin by coordinates.
  - Lots of internals refactoring.

1.13  Wed Sep 21 15:47:00 2011
  - Fix dependency on ExtUtils::ParseXS.

1.12  Fri Jun 10 21:29:00 2011
  - Fix segfaulting bug in median/mad calculation.
  - Fix t/50mean.t

1.11  Fri Jun 10 20:17:00 2011
  - Minor test bug fix release, no functional changes.

1.10  Thu Jun  9 13:07:00 2011
  * This release contains incompatible changes to previous releases *
  - INCOMPATBLE CHANGE: The rand() method now works on the
    histogram itself instead of its cumulation. This works by
    lazily constructing and caching a cumulative histogram
    internally and greatly reduces the chance for user errors:
    Before:
      my $ch = $hist->cumulative(1);
      push @rand_like_hist, $ch->rand for 1..100000;
    After:
      push @rand, $hist->rand for 1..100000;
  - Implemented the median() and median_absolute_deviation()
    methods to calculate estimates of those two quantities.
  - highest_bin() returns the index of the highest bin in the
    histogram.

1.03  Wed Jun  1 21:45:00 2011
  - The rand() method, when called on a cumulative distribution,
    returns random numbers sampled from the source distribution.
  - Includes a Mersenne-twister random number generator as
    Math::SimpleHisto::XS::RNG.
  - multiply_constant($k) will scale the histogram contents with $k.
  - The cumulative() method now takes an optional argument that
    is used to normalize the cumulative distribution's last bin.
  - The methods new_from_bin_range($i, $j)
    and new_alike_from_bin_range($i, $j)
    return clones of the histogram that include only the bins $i
    to $j.
  - Several new examples in examples/ and xt/ including an
    example of fitting and plotting histograms.

1.02  Wed May 25 09:00:00 2011
  - Test fixes: On some platforms, is($foo, $bar) is not accurate
    enough. Need an "is_approx" for floats.
  - Using done_testing where test number calculation exceeds the
    tests in complexity.

1.01  Sun May 22 21:48:00 2011
  - Version checks when deserializing
    Note: Future versions of this module will try to be able to
    deserialize the histograms from earlier versions. The other
    way around is not going to work. The indicator for forwards=
    incompatible changes is a major version bump (here: 0 => 1)
  - Variable bin-size histograms.
    (Requiring a total makeover and heavy refactoring)
  - $hist->cumulative to get the cumulation of the histogram

0.05  Thu May 18 18:30:00 2011
  - More documentation on dumping/serializing
  - Accepts any of JSON::XS, JSON::PP, JSON
    as JSON implementations.
  - Fast 'native_pack' serialization format
  - Includes benchmark script for serialization

0.04  Wed May 18  2:00:00 2011
  - Test fix (proper use of eval, DOH)

0.03  Tue May 16 19:00:00 2011
  - Test fix (proper use of SKIP)

0.02  Mon May 16 12:00:00 2011
  - Minor POD/doc fix

0.01  Sun May 15 23:00:00 2011
  - original version