> Still, why avoid changing loop variable? Does python treat looping > over a list differently from looping over an iterator, where it > doesn't know if the iterator future changes while loop running?
Take a look at Objects/listobject.c:listiterobject. It contains an it_index, which is the index into the list of the current iterator position. So if you delete items with indices smaller than the iterator index, the iterator index won't change, causing the iterator to skip over elements, e.g. L=range(10) for i in L: print i del L[0] Appending to lists will cause the new elements to be iterated over later. Whether that's desirable or not depends on the application. E.g. the following loop never terminates L=range(10: for i in L: L.append(i) Notice that the language specification *deliberately* does not distinguish between deletion of earlier and later items, but makes modification of the sequence undefined behavior to allow alternative implementations. E.g. an implementation that would crash, erase your hard disk, or set your house in flames if you confront it with your code still might be a conforming Python implementation. Regards, Martin -- http://mail.python.org/mailman/listinfo/python-list
