OK, thanks. Tricky. My exposure to classes didn't get as far as the
__call__ method.
That explains how it works.
On Feb 8, 7:17 am, Anthony <abasta...@gmail.com> wrote:
> > Where would I add these?
>
> Add the links wherever you want them in your user interface.
>
> > I am really mystified how the auth() class call receives the method as
> > its argument from these urls assuming that the user action looks like:
>
> > def user()
> > return dict(form=auth())
>
> Auth is a callable class, so when you call auth(), you are calling the
> Auth.__call__ method. That method checks to see if request.args[0] is among
> the set of supported Auth methods, and if so, it calls the associated
> method via:
>
> return getattr(self, args[0])()
>
> So, the __call__ method essentially acts as a router to the other Auth
> methods. Of course, you can also call those methods directly rather than
> relying on request.args and the __call__ router.
>
> > I am really mystified about how one can index the request.args using
> > parens instead of square braces?
>
> > as in:
>
> > <h2>{{=T( request.args(0).replace('_',' ').capitalize() )}}</h2>
>
> request.args is a storage.List object, which acts like a regular Python
> list but is also callable (i.e., it has a __call__ method). You can call it
> with an index, and it will return the same value as if you had indexed via
> the brackets syntax. The only difference is if you call it with an invalid
> index, instead of throwing an exception, it simply returns None. So, let's
> say request.args is an empty list -- in that case, request.args(0) is None,
> but request.args[0] will result in the following error: "IndexError: list
> index out of range". This allows you to do things like passing
> request.args(0) as a value without first having to check whether it exists.
>
> Anthony
