question about xrange performance

[_wolf]

lately i realized a slow running portion of my application, and a
quick profiling nourished the suspicion that, of all things, calls to
`xrange().__contains__` (`x in b` where `b = xrange(L,H)`) is the
culprit. to avoid any other influences, i wrote this test script with
class `xxrange` being a poor man’s `xrange` replacement:


########################################################
class xxrange( object ):
  def __init__( self, start, stop ):
    self.start  = start
    self.stop   = stop
  def __contains__( self, x ):
    return ( x == int( x ) ) and self.start <= x < self.stop

import cProfile
from random import randint
test_integers = [ randint 0, 5000 ) for i in xrange( 8000 ) ]
test_range_a  = xxrange( 10000, 20000 )
test_range_b  = xrange(  10000, 20000 )

def a():
  print test_range_a.__class__.__name__
  for x in test_integers:
    x in test_range_a

def b():
  print test_range_b.__class__.__name__
  for x in test_integers:
    x in test_range_b

cProfile.run('a()')
cProfile.run('b()')
########################################################

now this is the output, surprise:

########################################################
xxrange
         8003 function calls in 0.026 CPU seconds

   Ordered by: standard name

   ncalls  tottime  percall  cumtime  percall filename:lineno
(function)
        1    0.000    0.000    0.026    0.026 <string>:1(<module>)
        1    0.012    0.012    0.026    0.026 xrange-profiler.py:18(a)
     8000    0.014    0.000    0.014    0.000 xrange-profiler.py:9
(__contains__)
        1    0.000    0.000    0.000    0.000 {method 'disable' of
'_lsprof.Profiler' objects}


xrange
         3 function calls in 4.675 CPU seconds

   Ordered by: standard name

   ncalls  tottime  percall  cumtime  percall filename:lineno
(function)
        1    0.000    0.000    4.675    4.675 <string>:1(<module>)
        1    4.675    4.675    4.675    4.675 xrange-profiler.py:23(b)
        1    0.000    0.000    0.000    0.000 {method 'disable' of
'_lsprof.Profiler' objects}
########################################################

can it be that a simple diy-class outperforms a python built-in by a
factor of 180? is there something i have done the wrong way?
omissions, oversights? do other people get similar figures?

cheers