fix_mul to return an incorrect result for -2147483648*-2147483648
on 64 bit platforms
* test/ruby/test_integer_comb.rb (class TestIntegerComb): add test case
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@41689 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
(bigor_int): The allocated bignum should have enough size
to store long.
This fixes (bignum fits in a BDIGIT) | (fixnum bigger than BDIGIT)
on platforms which SIZEOF_BDIGITS < SIZEOF_LONG,
such as LP64 with 32bit BDIGIT (no int128).
git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@41639 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
