Hi SImon,
On 6 dic, 00:25, Simon King <simon.k...@uni-jena.de> wrote:
> I think that this is the wrong conclusion. It is easily possible to
> call the cached method that is inherited from the base class. Also, it
> is "class", not "category", by the way: The category framework is not
> involved here.
True, I obviously meant class and not category.
> B.foo? Is this a copy-and-past error? It should print A.foo.
Yep, copy-paste error.
>
> It is not clear to me why you would try to do A.foo(E) when you can
> easily do E.foo().
Because in my original problem, I am trying to over-ride the
inefficient base class method with a smarter version. However, the
smarter version can fail (due to some sage <-> gap glitches that I
will try to solve in due time) and if that happens I want to revert to
the original (even if inefficient) method. So my situation is more
like this:
sage: class A:
....: @cached_method
....: def foo(self):
....: print "A.foo"
....:
sage: class B(A):
....: @cached_method
....: def foo(self):
....: try:
....: # some computation that might fail
....: print("B.foo")
....: except: # If the computation fails, fall back to the
original A.foo method
....: A.foo(self)
As the computations involved are really expensive (explicitly listing
all the elements
in a conjugacy class), caching the result is a must.
> I'd say that it is not a bug and not a feature, but a problem - a
> problem that likely is not easy to solve. A cached method is not a
> method but an instance of the class "CachedMethodCaller". It is
> supposed to behave like a BOUND method, and so you call it like
> E.foo().
I might, if it wasn´t overriden by another method in class B.
Note that replacing the call A.foo(self) by A.foo() will also not
produce
the expected result.
>
> Apparently it would be possible to change the implementation of a
> cached method such that A.foo(E) inserts the cached version of foo
> into the dictionary of E - but so far it doesn't work. So, you can not
> call the cached version of the method by A.foo(E). But it is possible
> to call the uncached method, if this is of any help to you:
Not really, my fallback method is really expensive (both speed and
memory-wise) and I absolutely want to avoid doing the same
computations over and over when operating with conjugacy classes.
I guess the only way out is to resort to a hand-knitted cache, as
probably
the overhead of a hand-made definition is less than redoing the whole
computation :-/
Thanks for your help!
Cheers
Javier
--
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org
