On Thu, Mar 31, 2016 at 11:08 PM, Justin C. Walker <jus...@mac.com> wrote:
> Hi, all,
>
> I have couple of questions regarding "attributes" in Python/Sage:
>
> 1. If hasattr(X, "foo") returns True, does that mean that "X.foo" should not
> blow up?
Yes, this does, because unfortunately hasattr(X, 'foo') actually
access the .foo attribute, which is a pain when dealing with
properties and other descriptors that may take some time to execute.
The inverse, however, is very different, and different between Python
2 and Python 3. On Python 3 any exception in the attribute lookup
will result in hasattr(obj, 'foo') returning False. For example:
$ python2
Python 2.7.6 (default, Jun 22 2015, 17:58:13)
[GCC 4.8.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> class A:
... @property
... def foo(self):
... return 1/0
...
>>> hasattr(A(), 'foo')
False
The point here is that *any* exception from an attribute lookup
results in hasattr False.
On Python 3 they "fixed" that (in my opinion). Only an AttributeError
raised from the attribute lookup results in hasattr False--any other
exception bubbles up:
$ python3
Python 3.4.3 (default, Oct 14 2015, 20:28:29)
[GCC 4.8.4] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> class A:
... @property
... def foo(self):
... return 1 / 0
...
>>> hasattr(A(), 'foo')
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "<stdin>", line 4, in foo
ZeroDivisionError: division by zero
>>> class B:
... @property
... def foo(self):
... raise AttributeError('foo')
...
>>> hasattr(B(), 'foo')
False
(Unfortunately this is still not perfect--an AttributeError for an
unrelated attribute raised by a subroutine when accessing 'foo' can
still result in hasattr('foo') == False :(
One of my top wishlist items for Python 4 is some kind of __hasattr__
special method for classes and/or an equivalent for descriptors that
merely guarantees* that accessing the attribute of that name will not
result in an AttributeError. That would be extremely useful for
checking attribute availability without computing / returning a value.
I've never found a good workaround for that otherwise.
Erik
* This is of course an interface guarantee made by the programmer--it
is still subject to bugs.
> 2. Is there a way to tell, when hasattr(X, "foo") returns True, whether
> "X.foo" can be called?
>
> 3. If "X.foo" is callable, is it expected that "X.foo()" will *not* blow up
> with "Not Implemented"?
>
> I'm hoping that makes sense...
--
You received this message because you are subscribed to the Google Groups
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to sage-devel+unsubscr...@googlegroups.com.
To post to this group, send email to sage-devel@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.
