* ext/bigdecimal/lib/bigdecimal/math.rb (atan), test/bigdecimal/test_bigmath.rb (test_atan): explicitly specify the precision for calculating a reciprocal number of an argument. [Bug #3267]
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/branches/ruby_1_9_2@27735 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
improved precision and performance. based on a patch from Makoto
Yamashita in [ruby-core:25600] and [ruby-core:25602].
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@25013 b2dd03c8-39d4-4d8f-98ff-823fe69b080e