Change default setting of $Math::GMPq::RETYPE from 0 to 1.

CHANGES

@@ -2,6 +2,7 @@ Revision history for Perl module Math::GMPq
 
 0.57
  - Silence pointless compiler warnings.
+ - Change initial setting of Math::GMPq::RETYPE from 0 to 1.
 
 0.56
  - Skip Math::MPFR checks in t/overload_cross_class.t and t/overload_op_eq.t if

GMPq.pm

@@ -89,12 +89,14 @@ qgmp_urandomb_ui qgmp_urandomm_ui
 
 sub dl_load_flags {0} # Prevent DynaLoader from complaining and croaking
 
-    $Math::GMPq::RETYPE = 0; # set to 1 to enable a Math::GMPq object to be coerced to
+    $Math::GMPq::RETYPE = 1; # This variable used to be initially set to 0, but
+                             # this has changed, beginning with Math::GMPq-0.57.
+                             # This enables a Math::GMPq object to be coerced to
                              # a Math::MPFR object in certain overloaded operations.
                              # (See the 'OPERATOR OVERLOADING' section of the POD
                              # documentation for details.)
-                             # With this variable set to 0, these "certain overloaded
-                             # operations" alluded to will throw a fatal error.
+                             # Setting this variable to 0 will cause these "certain
+                             # overloaded operations" to throw a fatal error.
 
 sub new {
 

GMPq.pod

@@ -548,13 +548,13 @@
     Math::MPFR objects can provided as the second arg  with '+', '-',
     '*', '/' and '**'  operators - in which case the operation will
     return a Math::MPFR object.
-    If $Math::GMPq::RETYPE is set to 1, then Math::MPFR objects can also
-    be provided as the second arg with '+=', '*=', '-=', '/=' and '**='
-    operators. In such instances, the Math::GMPq object becomes
-    (is "retyped" to) a Math::MPFR object that contains the result of
-    the operation, using default Math::MPFR precision and default
-    Math::MPFR rounding mode.
-    $Math::GMPq::RETYPE is initially set to 0 (false).
+    If $Math::GMPq::RETYPE is set to 1 (as is the default setting), then
+    Math::MPFR objects can also be provided as the second arg with
+   '+=', '*=', '-=', '/=' and '**=' operators. In such instances,
+    the Math::GMPq object becomes (is "retyped" to) a Math::MPFR object
+    that contains the result of the operation, using default Math::MPFR
+    precision and default Math::MPFR rounding mode.
+    Set $Math::GMPq::RETYPE to 0, to disable this capability.
 
     To determine the type and value of the second argument the subroutine
     works through the following steps (in order), using the first value

t/overload_op_eq.t

@@ -6,7 +6,8 @@
 # to a true value.
 # A fatal error occurs if $Math::GMPq::RETYPE is
 # set to a false value. The initial value of
-# $Math::GMPq::RETYPE is 0 (false).
+# $Math::GMPq::RETYPE is now 1 (true), but was 0
+# in Math-GMPq-0.56 and earlier.
 
 use strict;
 use warnings;
@@ -22,7 +23,9 @@ my $q  = Math::GMPq->new('1/11');
 my $fr = 0;
 $fr = Math::MPFR->new(17.1) if $have_mpfr;
 
-cmp_ok($Math::GMPq::RETYPE, '==', 0, "retyping not allowed");
+cmp_ok($Math::GMPq::RETYPE, '==', 1, "retyping allowed");
+
+$Math::GMPq::RETYPE = 0; # Disallow retyping
 
 eval {$q *= $fr;};
 if(ref($fr)) {