File: newobject1_runme.rb

package info (click to toggle)
swig 4.1.0-0.2
  • links: PTS, VCS
  • area: main
  • in suites: bookworm
  • size: 47,992 kB
  • sloc: cpp: 50,555; ansic: 27,840; java: 15,366; python: 11,221; cs: 8,852; ruby: 6,307; yacc: 6,290; makefile: 5,702; sh: 5,492; perl: 3,818; php: 3,046; ml: 2,094; lisp: 1,756; javascript: 1,751; tcl: 1,499; xml: 115
file content (33 lines) | stat: -rw-r--r-- 940 bytes parent folder | download | duplicates (3)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
#!/usr/bin/env ruby
#
# Simple test of %newobject
#  * The purpose of this test is to confirm that a language module
#  * correctly handles the case when C++ class member functions (of both
#  * the static and non-static persuasion) have been tagged with the
#  * %newobject directive.
#
# 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 invocation, 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
#
# 
# 
# 
#

require 'swig_assert'
require 'swig_gc'
require 'newobject1'

include Newobject1

GC.track_class = Foo
GC.start
100.times { foo1 = Foo.makeFoo; foo2 = foo1.makeMore }
GC.stats if $VERBOSE
swig_assert( 'Foo.fooCount == 200', binding, "but is #{Foo.fooCount}" )
GC.start
swig_assert( 'Foo.fooCount <= 2', binding, "but is #{Foo.fooCount}" )