Raymond Hettinger added the comment: Alok, overall the patch looks pretty good and you've done great work on it.
However, in working through its details, I find myself having major misgivings about doubling the size and complexity of the code for something that may not be ever benefit any real code. Terry noted that range() supports values bigger than the word size but the needs there are much different. Programs can reasonably use ranges with large start points, but an islice() call would have to iterate over *start* values before it begins returning any usable values: list(range(sys.maxsize+10, sys.maxsize+20)) # maybe a good idea list(islice(count(), sys.maxsize + 10, sys.maxsize + 20)) # probably not a good idea When we finally get 64-bit everywhere (not there yet), I think the code in this patch would never get exercised. Even in the 32-bit world, islicing over 2**32 inputs doesn't seem like a great idea. ---------- _______________________________________ Python tracker <rep...@bugs.python.org> <http://bugs.python.org/issue6305> _______________________________________ _______________________________________________ Python-bugs-list mailing list Unsubscribe: https://mail.python.org/mailman/options/python-bugs-list/archive%40mail-archive.com
