ruby.git/numeric.c, branch v1_8_6_151

merge revision(s) 13902, 13907, 13914:

2008-06-07T18:46:09+00:00

	* numeric.c (flo_divmod): round to the nearest integer.
	  [ ruby-Bugs-14540 ]
	* numeric.c (flodivmod): work around for inifinity.
	* numeric.c (flo_divmod): work around for platforms have no round().
	  [ruby-dev:32247]
	* numeric.c (round): fallback definition.
	* numeric.c (flo_divmod, flo_round): use round() always.
	  [ruby-dev:32269]


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8_6@16931 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

merge revision(s) 13785:

2008-06-07T18:05:03+00:00

	* numeric.c (fix_pow): returns infinity for 0**-1.  [ruby-dev:32084]


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8_6@16919 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

* bignum.c (rb_big_lshift, rb_big_rshift): separated functions

2007-08-22T01:50:21+00:00

  to get rid of infinite recursion.  fixed calculation in edge
  cases.  [ruby-dev:31244]
* numeric.c (rb_fix_lshift, rb_fix_rshift): ditto.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8_6@13188 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

* numeric.c (fix_pow): integer power calculation: 0**n => 0,

2007-08-22T01:26:39+00:00

	  1**n => 1, -1**n => 1 (n: even) / -1 (n: odd).
	* test/ruby/test_fixnum.rb (TestFixnum::test_pow): update test
	  suite.  pow(-3, 2^64) gives NaN when pow(3, 2^64) gives Inf.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8_6@13182 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

* numeric.c (fix_pow): 0**2 should not raise floating point

2007-08-22T01:05:52+00:00

	  exception.  [ruby-dev:31216]


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8_6@13170 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

* numeric.c (fix_cmp, fix_equal): Remove FIX2LONG() to optimize.

2007-02-23T04:24:16+00:00

  suggested in
  http://t-a-w.blogspot.com/2007/02/making-ruby-faster.html.
  [ruby-talk:240223]


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8_6@11823 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

Minor rdoc fix

2006-09-24T20:47:57+00:00

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8@11013 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

* numeric.c (fix_plus): addition in Fixnum will never overflow

2006-09-07T16:35:59+00:00

  long.  a patch from Ondrej Bilka .
  [ruby-core:08794]

* numeric.c (fix_minus): ditto.

* bignum.c (rb_big_pow): eagerly truncate resulting bignum.
  [ruby-core:08794]


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8@10880 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

* numeric.c (flo_hash): improve collision. fixed: [ruby-dev:29352]

2006-09-04T20:57:02+00:00

git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8@10864 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

* common.mk, configure.in, defines.h, eval.c, gc.c, main.c,

2006-09-01T18:47:11+00:00

  numeric.c, ruby.h, ia64.s: backport IA64 HP-UX support.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_8@10827 b2dd03c8-39d4-4d8f-98ff-823fe69b080e