Incorrect conversion from ℚ[-i] to SR

[mezzarobba]

This should return -I, not I:

sage: K.<j> = QuadraticField(-1, embedding=CC(0,-1))
sage: SR(j)
I

Works under Sage 9.2, so maybe related to #18036?

CC: @videlec

Component: symbolics

Author: Marc Mezzarobba

Branch/Commit: fddaa2c

Reviewer: Vincent Delecroix

Issue created by migration from https://trac.sagemath.org/ticket/31551

[slel]

comment:2

Possibly related:

• eigenvectors over QQbar are incorrectly conjugated #30518: Eigenvectors over QQbar are incorrectly conjugated

[videlec]

comment:3

Indeed, the following method on gaussian integers looks bad

    def _symbolic_(self, SR):
        r"""
        EXAMPLES::

            sage: SR(1 + 2*i)
            2*I + 1
        """
        from sage.symbolic.constants import I
        return self[1]*I + self[0]

[mezzarobba]

comment:4

Hmm, if I remember right, the GaussianInteger class was intended for ℚ[i] with its standard embedding, so I think there is something else.

[mezzarobba]

comment:5

...but I see no reason not to extend it to support both embeddings.

---

New commits:

fddaa2c

#31151 support both embeddings in NumberFieldElement_gaussian

[mezzarobba]

Branch: u/mmezzarobba/ticket/31151

[mezzarobba]

Commit: fddaa2c

[mezzarobba]

Author: Marc Mezzarobba

[videlec]

comment:6

looks good... waiting for patchbot.

[videlec]

Reviewer: Vincent Delecroix

[mezzarobba]

comment:8

thanks for the quick review!

[mezzarobba]

comment:9

Raising the priority to critical to stress that the fix really should go in Sage 9.3.

[mkoeppe]

comment:10

Setting priority to blocker to bring this ticket to the attention of the release bot.

[vbraun]

Changed branch from u/mmezzarobba/ticket/31151 to fddaa2c