From: Marc-Andre Lafortune <redmine@...>
Date: 2010-03-02T13:45:11+09:00
Subject: [ruby-core:28395] [Bug #2830] Some methods raise ArgumentError instead of TypeError

Bug #2830: Some methods raise ArgumentError instead of TypeError
http://redmine.ruby-lang.org/issues/show/2830

Author: Marc-Andre Lafortune
Status: Open, Priority: Normal
Assigned to: Marc-Andre Lafortune, Category: core, Target version: 1.9.2
ruby -v: ruby 1.9.2dev (2010-03-02 trunk 26792) [x86_64-darwin10.2.0] 

Some methods of Ruby 1.9 expect integers/reals and call internally nurat_int_value/nurat_int_check. These functions raise an ArgumentError when the argument is not an Integer, instead of a TypeError.

Thus:
  42.gcd(:foo)     # => ArgumentError, should be TypeError
  42.lcm(:foo)     # => ditto
  42.gcdlcm(:foo)  # => ditto
  Rational(:foo,1) # => ditto

Note that on the other hand:
  Rational(nil, 1) # => TypeError
  Rational(:foo)   # => TypeError

In a similar fashion:
  Complex.rect(nil)  # => ArgumentError, should be TypeError
  Complex.polar(nil) # => ditto


Unless there is objection, I will commit the following patch (and fix RubySpec):

diff --git a/complex.c b/complex.c
index 214d3a2..6742257 100644
--- a/complex.c
+++ b/complex.c
@@ -377,7 +377,7 @@ nucomp_real_check(VALUE num)
        break;
       default:
        if (!k_numeric_p(num) || !f_real_p(num))
-           rb_raise(rb_eArgError, "not a real");
+           rb_raise(rb_eTypeError, "not a real");
     }
 }
 
diff --git a/rational.c b/rational.c
index 65d3cf4..f5a6d26 100644
--- a/rational.c
+++ b/rational.c
@@ -419,7 +419,7 @@ nurat_int_check(VALUE num)
        break;
       default:
        if (!k_numeric_p(num) || !f_integer_p(num))
-           rb_raise(rb_eArgError, "not an integer");
+           rb_raise(rb_eTypeError, "not an integer");
     }
 }


----------------------------------------
http://redmine.ruby-lang.org