2017-01-11 v1.13 pjacklam

 * Reorder entries in CHANGES so newest entries are at the top.

 * Sync test files with Math-BigInt-1.999808.

 * Numerous bug fixes and speed improvements.

2004-01-05 v1.12 Tels

 * tests from BigInt v1.68 - 4572 tests

 * require Math::BigInt v1.68

 * hang in test 1635 due to _fac() not handling $x == 0 properly

 * _lsft() created non-chunked sizes (thus failures in broot())

 * streamlined code (removed return statements etc)

 * _as_hex() and _as_bin() handle the $x == 0 case properly

 * _pow() handles 0 ** Y properly

 * __reduce() check for Y's size in _gcd() was wrongly checking $x->Size()

 * _check() method checks for non-chunked sizes

2002-09-30 v1.11_01 Tels (never released)

 * require Bit::Vector 6.3

 * removed unused binary GCD algorithmn

 * removed unnecc. tests for $x or $y == 0 since Bit::Vector 6.3 handles these

2002-03-24 v1.11 Tels

 * tests from BigInt v1.60 - 4057 tests

2002-03-24 v1.10 Tels

 * tests from BigInt v1.55 - 3877 tests

2002-02-21 v1.09 Tels

 * tests from BigInt v1.51

 * replaced _core_lib() by config()->{lib}

2002-02-19 v1.08 Tels

 * added _fac() for factorial speed improvements

 * streamlined some functions (removed return)

 * _mul() does no longer create a result of max($m,$n)*2, but a $m+$n+2 bits
   Quite some performance gains when $m much is greater/smaller than $n.
   The improvement for small numbers is due to the saved call to max()

        Sat Feb 16 19:12:04 2002 Numbers are relative to v1.50_BitVect_v1.07,
                                 100 denotes 100%.

                            |    v1.50
                            |  BitVect
                            |    v1.08
         -------------------+----------
          1 * 1             |      105
          2 * 2             |      103
          12 * 12           |      103
          123 * 123         |      101
          1234 * 1234       |      104
          12345 * 12345     |      103
          123456*123456     |      100
          1234567*1234567   |      101
          12345678*12345678 |      102
          12345679*12345679 |      102
          12 * 2            |      100
          123 * 2           |      104
          1234 * 2          |      103
          12345 * 2         |      102
          123456 * 2        |      102
          1234567 * 2       |      103
          12345678 * 2      |      103
          123456789 * 2     |      102
          1e10 * 1e10       |      104
          1e100 * 1e100     |      101
          1e1000 * 1e1000   |      100
          1 * -1            |      102
          1e1000 * 1        |      108          <-- !
          1e1000 * 10       |      112          <-- !
          1e1000 * -10      |      110          <-- !
         bmul:              |      103
         ...................|..........

2002-02-10 v1.07 Tels

 * added _lsft() and _rsft() for more speed on shifting

 * tests from MBI v1.50 (3685 vs. 2932 tests)

2001-11-11 v1.06 Tels

 * made _inc/_dec faster and added tests for them

 * fixed bug in __reduce() that made it never  work (oups, forgot debug again)

2001-11-01 v1.05 Tels

 * testsuite from BigInt v1.45

2001-09-03 v1.04 Tels

 * _as_hex(), _as_bin()

 * tests from MBI 1.42

 * removed default exporting, added empty import()

2001-08-03 v1.03 Tels

 * added tests from MBI v1.40

2001-07-22 v1.02 Tels (never released)

 * removed $class/__PACKAGE__ checks entirely

 * fixed 2 ** something to be more efficient

 * made Bit::Vector a requirement

2001-07-13 v1.01 Tels

 * new testsuite from BigInt v1.37, tweaks to it to silence Exporter

 * $class => __PACKAGE__

2001-07-09 v1.00 Tels

 * First release.