Fixed ruby's newobject1 and newobject2 demos.

Added swig_assert.rb and started using it in some tests.
Added my patches to ruby.cxx, rubyrun.swg, rubystrings.swg
and rubytracking.swg [see: sourceforge]



git-svn-id: https://swig.svn.sourceforge.net/svnroot/swig/trunk@9691 626c5289-ae23-0410-ae9c-e8d60b6d4f22

Examples/test-suite/ruby/newobject2_runme.rb

@@ -1,15 +1,29 @@
+#!/usr/bin/env ruby
+#
+# Ruby's GC is somewhat broken in that it will mark some more stack space
+# leading to the collection of local objects to be delayed.
+# Thus, upon invokation, it sometimes you can wait up to several
+# instructions to kick in.
+# See: http://blade.nagaokaut.ac.jp/cgi-bin/scat.rb/ruby/ruby-core/7449
+#
+# Author::    gga
+# Copyright:: 2007
+# License::   SWIG
+#
+
+require 'swig_assert'
 require 'newobject2'
 
 include Newobject2
 
-foo1 = makeFoo
-raise RuntimeError if fooCount != 1
 
-foo2 = makeFoo
-raise RuntimeError if fooCount != 2
+100.times { foo1 = makeFoo }
+GC.start
+swig_assert( 'fooCount == 1', "but is #{fooCount}" )
 
-foo1 = nil ; GC.start
-raise RuntimeError if fooCount != 1
 
-foo2 = nil ; GC.start
-raise RuntimeError if fooCount != 0
+@foos = []
+100.times { @foos << makeFoo }
+swig_assert( 'fooCount == 101', "but is #{fooCount}" )
+
+GC.start