Steve Howell a écrit : > --- Bruno Desthuilliers > <[EMAIL PROTECTED]> wrote: > > >>Colin J. Williams a écrit : >> >>>I'm not sure that I like add 3, 5, 7 >>> >>>but it would be nice to be able to drop the >> >>parentheses >> >>>when no argument is required. >>> >>>Thus: close; >>>could replace close(); >> >>This just could not work given Python's object >>model. The parens >>actually *are* the call operator. >> > > > I mostly agree with you, but in the specific use case > of having just a single token on a line, you could > argue that Python could DWIM on calling an object if > the object is callable, since otherwise it's just a > no-op.
It's not. If the name is not defined, it raises an exception (either a NameError or an AttributeError). And if the name resolves to a computed attribute, the getter for this computed attribute will be invoked - with possible side effects. > I think the argument against doing that is > more based on explicit-vs.-implicit principle versus > actual constraints of the object model. I'd say that the object model being the work of a strong proponent of the explicit-vs.-implicit principle, it's probably another chicken-and-egg problem - so we are both right here !-) > Another aspect of Ruby is that the final expression > evaluated in a method actually gets returned as the > result of a method, Unless there's an explict return before... > which has further implications on > whether "close" is simply evaluated or called. I'm sorry but I'm not sure I get the point here. -- http://mail.python.org/mailman/listinfo/python-list
