On Wednesday, March 5, 2014 9:38:58 AM UTC+5:30, Ian wrote: > On Tue, Mar 4, 2014 at 8:36 PM, Rustom Mody wrote: > > * ... which summarizes my objection in this thread: Python's 'is' leaks the > > machine abstraction. 'id' does it legitimately (somewhat), > > 'is' does it illegitimately
> Well, since "if x == None" is buggy as a test for sentinel values, > that means the only legitimate non-buggy way to do it is with "if > id(x) == id(None)", which just seems gross to me. Ha Ha! -- Fully agreed! One should not need to leak the machine abstraction for something as basic as None testing As I said earlier my current preference which is least disruptive is to have as builtin def isNone(x): return x is None which should obviate most (basic) uses of is -- https://mail.python.org/mailman/listinfo/python-list
