On Saturday, April 24, 2021 at 1:33:36 AM UTC+2 slelievre wrote:
> Could this have to do with one of the following existing tickets?
> 
> - Sage Trac ticket 22008
>  complex_embedding on relative number fields is inconsistent with the 
base field
>   https://trac.sagemath.org/ticket/22008

This is indeed related: it also involves Sage mistaking the extension [L:K] 
for the extension [L:QQ]. But it does not seem to be the exact same 
problem: I am not sure that fixing ticket 17524 would automatically fix the 
problem I am describing here. Well, it *might*; but the only way to be sure 
is to actually pinpoint the problem - and this would involve a complex hunt 
through a lot of different code files that I (being a total newbie to Sage 
development) do not feel like doing right now.

> - Sage Trac ticket 17524
>   polynomial for relative number field elements
>   https://trac.sagemath.org/ticket/17524

This one, OTOH, seems a bit different: in this one, there is no confusion 
between [L:K] and [L:QQ], but simply a problem of consistency between the 
chosen embedding of K when taken by itself and the chosen embedding of K 
when seen as a subfield of L.


On Saturday, April 24, 2021 at 9:51:27 PM UTC+2 vdelecroix wrote:

> At least on 9.3.rc4 it is fixed by using instead 
>
> sage: K.<sqrt2> = QuadraticField(2, embedding=AA(2).sqrt()) 
> sage: L.<s> = K.extension(x^3 + (sqrt2/2 + 1/3)*x^2 + (2*sqrt2/5+3/7)*x - 
> 1) 
> sage: phi0 = hom(L, QQbar, roots[0]) 
> sage: phi1 = hom(L, QQbar, roots[1]) 
> sage: phi2 = hom(L, QQbar, roots[2]) 
> sage: phi0(sqrt2) 
> 1.414213562373095? 
> sage: phi1(sqrt2) 
> 1.414213562373095? + 0.?e-17*I 
> sage: phi2(sqrt2) 
> 1.414213562373095? + 0.?e-17*I 
>

For what it is worth, in my config (version 9.1), this gives a "ValueError: 
relations do not all (canonically) map to 0 under map determined by images 
of generators". So, OK, presumably this has been fixed in the meanwhile. 
But anyway, where can I find documentation for the "hom" method?

Assuming it works, OK, thank you, it is a good workaround. Still, in the 
meanwhile, I think that something needs to be done about 
NumberFieldEmbedding: it should either be fixed, be deprecated, or at least 
the documentation should be updated to signal that it does not work in 
relative fields. So I guess I will now try to open a ticket. 

-- 
You received this message because you are subscribed to the Google Groups 
"sage-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-support/867cb177-9e58-4332-88be-b5cd0c9cf5cfn%40googlegroups.com.

Reply via email to