On 12/10/10 10:47 AM, Simon King wrote:
Dear Jason,
On 10 Dez., 14:55, Jason Grout<jason-s...@creativetrax.com> wrote:
In sage/modules/vector_rational_dense.pyx there is a class definition
Vector_rational_dense. It has a cdef'd method _cmp_c_impl, which is
supposed to do the comparison.
...
Hence, the comparison method is consulted *only* when calling the
special method __cmp__ explicitly! So, how else is comparison
achieved??
Perhapshttp://groups.google.com/group/sage-devel/msg/ca37624c4ad655c6
could be helpful? That message deals with comparison between matrices,
but you might find some useful information for tracking down what is
happening.
Thank you, the message is helpful, to some extent. I see that
__richcmp__ and __cmp__ can not be inherited, so that the code from
sage.structure.element should be copied. So, it confirms some of my
suspicions.
I think you misunderstood the post. They *can* be inherited. The point
is that "according to the Python/C API docs [3], either all of __hash__,
__cmp__, and __richcmp__ are inherited, or none are. "
In our case, since none of these three is defined in
vector_rational_dense, they are all inherited from free_module_element.
But now, my confusion has increased. I see that
sage.modules.vector_rational_dense does *not* define its own
__richcmp__ and __cmp__. So, it is even less clear why it works.
I inserted
print "_cmp_c_impl in vector_rational_dense.pyx"
in the vector_rational_dense.pyx _cmp_c_impl function. I inserted
print "_cmp_c_impl in FreeModuleElement (free_module_element.pyx)"
in the free_module_element.pyx _cmp_c_impl function, and
print "__richcmp__ in FreeModuleElement (free_module_element.pyx)"
in the free_module_element.pyx __richcmp__ function. Then I get the
following:
sage: v = vector(QQ, [0,0,0,0])
sage: w = vector(QQ, [-1,3/2,0,0])
sage: cm = sage.structure.element.get_coercion_model(); cm
<sage.structure.coerce.CoercionModel_cache_maps object at 0x101f686b0>
sage: cm.explain(v,w,operator.eq)
Identical parents, arithmetic performed immediately.
Result lives in Vector space of dimension 4 over Rational Field
Vector space of dimension 4 over Rational Field
sage: v==w
__richcmp__ in FreeModuleElement (free_module_element.pyx)
False
sage: cmp(v,w)
__richcmp__ in FreeModuleElement (free_module_element.pyx)
__richcmp__ in FreeModuleElement (free_module_element.pyx)
__richcmp__ in FreeModuleElement (free_module_element.pyx)
1
So I see that it is the free module element class (the superclass of
vector_rational_dense) has the __richcmp__ function defined, and it
calls the special FreeModuleElement._cmp_same_ambient_c function. So we
are not using the coercion system, nor are we using the
vector_rational_dense function, in either of the above calls.
At least, that's the way I understand it.
Thanks,
Jason
--
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
