Issue1784950

>>> filter(lambda x :x==1,[1,2])
[1]
>>> filter(lambda (x) :x==1,[1,2])
Traceback (innermost last):
  File "<console>", line 1, in ?
  File "<console>", line 1, in <lambda>
TypeError: iteration over non-sequence


(the latter works in python)

A little more exploratory work:

Jython 2.2:

>>>  filter( lambda (x):x==1, [(1,),(2,)] )
[(1,)]

While in Python 2.4:

>>> filter( lambda (x):x==1, [(1,),(2,)] )
[]

In fact, in a simpler example,

In Jython 2.2:

>>> f = lambda (x):x
>>> f(1)
Traceback (innermost last):
  File "<console>", line 1, in ?
  File "<console>", line 1, in <lambda>
TypeError: iteration over non-sequence
>>> f((1,))
1
>>> f((1,2))
Traceback (innermost last):
  File "<console>", line 1, in ?
  File "<console>", line 1, in <lambda>
ValueError: unpack tuple of wrong size
>>> 

While in Python 2.4: 

>>> f = lambda (x):x
>>> f(1)
1
>>> f((1,))
(1,)
>>> f((1,2))
(1, 2)
>>> 

It seems that Jython takes lambda (x) a little too literally, assuming that the argument passed will be only a single element tuple. While Python interprets it like any other function definition.

I will look into the matter further, time permitting.

HTH,
~musically_ut

I've uploaded a naive fix for this here:

http://sourceforge.net/tracker/index.php?func=detail&aid=1793995&group_id=12867&atid=312867

Fixed in 2.5
Need a unit test here

test_complex_args provides ample test coverage of this now fixed bug.