Mv.__pow__ is not returning Mv instance as return value when raise to power of zero about galgebra

[utensil]

IIUC, this behavior is by design. In such cases, the result should downgrade to scalars naturally, so a Mv is no longer expected.

This is also a recurring theme during maintenance, where you could be dealing with a Mv then suddenly you are dealing with a SymPy expression like a Mul, or suddenly it's just a symbol that the result is simplified to.

This phenomenon is also responsible for the oversight you spot in #513 .

So, this is a non-issue for maintenance, though I agree with you on the UX aspect as a user.

[mammalwong]

@utensil It sounds like a very weird design choice, I have multiple points to support that is weird:

• lets say there is a non-scalar Mv instance $a$, when $a a$ is a scalar, a**2 returns a Mv instance of scalar value under the current implementation, while a**0 exceptionally return sympy.S.One instead of a Mv instance
• in terms of math pow should be a mapping $Cl(p,q,r) \times Cl(p,q,r) \rightarrow Cl(p,q,r)$ even if the result is a scalar, that scalar should still have properties like its grade under the current Ga instance
• in terms of coding it is a very error prone design, there are no documentation or return value type hint to indicate this exceptional case

[utensil]

For __pow__, a fix is pushed, and the existing tests are passing. So it would be fixed for this specific case, unless I find out that it causes other regressions. In general, this remains a wontfix, as you can find S.one usage everywhere in GAlgebra.

[mammalwong]

Thanks for fixing this issue 👍

[utensil]

Thank you, @mammalwong , the fix should be in master now as #510 is now merged.