I think it's fine to require that objects in a category support weakrefs. There are certainly categories where you might want integers to be the objects, but then I think that you need to wrap those integers in some class inheriting from CategoryObject.
Simon, where were you planning on adding this new requirement to the documentation? Is there a good place in the Developer guide? David On Mon, Dec 19, 2011 at 16:11, Simon King <[email protected]> wrote: > Hi William > > On 19 Dez., 22:43, William Stein <[email protected]> wrote: >> But I'm personally OK with loosing this feature. The feature we're >> loosing though is this: objects of a category have to support weak >> references. >> >> Is that feature OK to loose? I can't think of a reason we need that >> feature. > > We certainly do not loose that feature - because that feature simply > does not exist. > > Namely, any element is considered to be an object in some category, > such as > sage: 1.category() > Category of elements of Integer Ring > > However, sage.structure.element.Element does not support weak > references, such as > sage: weakref.ref(1) > > --------------------------------------------------------------------------- > TypeError Traceback (most recent > call last) > > /home/simon/SAGE/sage-4.8.alpha3/<ipython console> in <module>() > > TypeError: cannot create weak reference to > 'sage.rings.integer.Integer' object > > > So, there currently *are* objects of some strange category that do not > support weak references - which is why these 3 times 2 doctests fail > when using a weak cache for homsets. > > Cheers, > Simon > > -- > To post to this group, send an email to [email protected] > To unsubscribe from this group, send an email to > [email protected] > For more options, visit this group at > http://groups.google.com/group/sage-devel > URL: http://www.sagemath.org -- To post to this group, send an email to [email protected] To unsubscribe from this group, send an email to [email protected] For more options, visit this group at http://groups.google.com/group/sage-devel URL: http://www.sagemath.org
