PolynomialRealDense.quo_rem() returns zero polynomials with wrong degree

[pjbruin]

In Sage 6.7, the following is correct:

sage: R.<x> = QQ[]
sage: z = R.zero()
sage: z.degree()
-1
sage: q, r = z.quo_rem(x)
sage: q.degree()
-1

The following (QQ -> RR) is not:

sage: S.<x> = RR[]
sage: z = S.zero()
sage: z.degree()
-1
sage: q, r = z.quo_rem(x)
sage: q.degree()
-2

The last result should be -1; the given answer implies for example

sage: q == z
False

which is nonsense since q and z are both the zero polynomial.

Component: basic arithmetic

Keywords: polynomial degree quo_rem

Author: Peter Bruin

Branch/Commit: 9dc90cc

Reviewer: Bruno Grenet

Issue created by migration from https://trac.sagemath.org/ticket/18467

[pjbruin]

Branch: u/pbruin/18467-quo_rem

[pjbruin]

Author: Peter Bruin

[pjbruin]

Commit: 9dc90cc

[bgrenet]

comment:2

Looks fine to me! A minor comment: It seems that people prefer not referring to `self` in the documentation, so you could replace the documentation of this function by

Return the quotient with remainder of this polynomial by ``other``.

[bgrenet]

Reviewer: Bruno Grenet

[pjbruin]

comment:3

Replying to @bgrenet:

Looks fine to me! A minor comment: It seems that people prefer not referring to `self` in the documentation

If you really insist I can change it, but I personally prefer self over this polynomial (see #13442 comment:5).

There was some recent discussion on a related question, namely whether self should be documented in INPUT sections of methods. On that question I agree with most people that it should not.

[bgrenet]

comment:4

Replying to @pjbruin:

Replying to @bgrenet:

Looks fine to me! A minor comment: It seems that people prefer not referring to `self` in the documentation

If you really insist I can change it, but I personally prefer self over this polynomial (see #13442 comment:5).

There was some recent discussion on a related question, namely whether self should be documented in INPUT sections of methods. On that question I agree with most people that it should not.

Actually, I was referring to (the beginning of) this discussion. Yet there does not seem to be a consensus. And I do not have a strong opinion on this... so I do not insist!

[vbraun]

Changed branch from u/pbruin/18467-quo_rem to 9dc90cc