JB wrote:

> > I bet you
> > can do better by exploiting the specifics of this case!-)
>  
>  Well, I can use the fact that I know the maximal sorting key
>  in advance. So I can append sentinel entries to both lists
>  and this makes things much easier.

Yes.  While this may not be the optimal solution to your
actual problem (as I notice the thread is now exploring
quite different architectures), it's getting better as a
solution to the problem you originally expressed about
"moving list entries", as per subject.  Untested code
below, but getting a bit more readable, I hope...:


au = [None] * 2
def prepare(au, alist, idx):
    alist.append(sentinel)
    it = iter(alist)
    au[idx] = [ it.next(), idx, it.next, [] ]
prepare(au, list1, 0)
prepare(au, list2, 1)

# naming the items of au's items helps readability:
ITEM, INDX, ITER, LIST = range(4)

# writing complicated expressions as functions also helps:
def moredata(au):
    return au[0][ITEM} != sentinel or au[1][ITEM} != sentinel

while moredata(au):
    m = min(au)
    i = m[INDX]
    if i and f(m[ITEM]): i = 0
    au[i][LIST].append(m[ITEM])
    m[ITEM] = m[ITER]()

# now, both lists are exhausted, so just copy the results back
list1[:] = au[0][LIST]
list2[:] = au[1][LIST]


sentinel is any value >  max(max(list1), max(list2)).


Alex

-- 
http://mail.python.org/mailman/listinfo/python-list