Ok... not exactly what I said but even simpler. This is #17099 which
needs review.

Vincent

2014-10-04 20:31 UTC+02:00, Vincent Delecroix <[email protected]>:
> Ok. The bug is in the canonicalize function of QuaternionAlgebra...
> the function does not care about 0. This is "very optimized" and "very
> wrong" ;-)
>
> I will provide a ticket in a minute.
>
> Vincent
>
> 2014-10-04 20:01 UTC+02:00, Vincent Delecroix <[email protected]>:
>> And more precisely a lack of simplification at some point
>>
>> sage: w[0].denominator()
>> 2
>> sage: z[0].denominator()
>> 4
>>
>> 2014-10-04 19:57 UTC+02:00, Vincent Delecroix
>> <[email protected]>:
>>> Hello Nathann,
>>>
>>> Seems to be the comparison in the number field...
>>>
>>> sage: z[0]
>>> -1/2
>>> sage: w[0]
>>> -1/2
>>> sage: z[0] == w[0]
>>> False
>>> sage: z[0].parent()
>>> Number Field in a with defining polynomial x^3 + x - 1
>>> sage: w[0].parent()
>>> Number Field in a with defining polynomial x^3 + x - 1
>>>
>>> Vincent
>>>
>>> 2014-10-04 19:44 UTC+02:00, Nathan Dunfield <[email protected]>:
>>>> There's something wrong with comparison operators for elements of
>>>> QuaternionAlgebras defined over a number field.  Here's a simple
>>>> example,
>>>> where first it gives the wrong answer and then, after doing some
>>>> arithmetic
>>>>
>>>> with the elements, a mix of right and wrong answers:
>>>>
>>>> sage: K = NumberField(x**3 + x - 1, 'a')
>>>> sage: D.<i,j,k> = QuaternionAlgebra(K, -1, -3)
>>>> sage: x = (i + j)/K(2)
>>>> sage: z =  (-1 + (i + j) + k)/K(2)  - x
>>>> sage: w = i * x
>>>> sage: w, z
>>>> (-1/2 + 1/2*k, -1/2 + 1/2*k)
>>>> sage: w == z
>>>> False
>>>> sage: z == w
>>>> False
>>>> sage: z - w
>>>> 0
>>>> sage: z == w, w == w, z-w == 0
>>>> (True, True, False)
>>>>
>>>> I looked in trac, but I didn't see this.  Is this issue already known,
>>>> and,
>>>>
>>>> if so, is there a simple work-around?
>>>>
>>>> Thanks,
>>>>
>>>> Nathan
>>>>
>>>> --
>>>> 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 [email protected].
>>>> To post to this group, send email to [email protected].
>>>> Visit this group at http://groups.google.com/group/sage-devel.
>>>> For more options, visit https://groups.google.com/d/optout.
>>>>
>>>
>>
>

-- 
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 [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sage-devel.
For more options, visit https://groups.google.com/d/optout.

Reply via email to