Actually, it seems that I have already found an instance of the problem you
describe, except that I do not understand it.
In sage/algebras/jordan_algebra.py we have two "def __ne__", one in
JordanAlgebraSymmetricBilinear and one in SpecialJordanAlgebra.
If I remove them (moving the tests to those of __eq__), the following fails:
sage: m = matrix([[0,1],[1,1]])
sage: J.<a,b,c> = JordanAlgebra(m)
sage: x = 4*a - b + 3*c
sage: x != J((4, (-1, 3)))
False
I have no idea why, the element class is defined directly in the parent
JordanAlgebraSymmetricBilinear, and inherits only from AlgebraElement.
Apparently, the problem also arises, if _richcmp_ is implemented in a
superclass. I think I knew that once...
Martin
On Saturday, 2 September 2023 at 13:15:04 UTC+2 Martin R wrote:
> On Saturday, 2 September 2023 at 11:39:12 UTC+2 Oscar Benjamin wrote:
>
> On Sat, 2 Sept 2023 at 08:44, 'Martin R' via sage-devel
> <sage-...@googlegroups.com> wrote:
>
> ...
>
> It is easy for this sort of thing to be overlooked in test code and in
> fact
> messing with __eq__/__ne__ (more so __eq__) can invalidate much of the
> test suite so I would tread carefully.
>
>
> Could you provide an example? I would think that in such a case a doctest
> should catch the problem - what other purpose would a doctest have?
>
> Thank you for your input!
>
--
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 view this discussion on the web visit
https://groups.google.com/d/msgid/sage-devel/6c0cad2a-864c-4635-b4c2-20ec807527b0n%40googlegroups.com.
