Make R.quotient_ring(I) normalize generator in the univariate case (easy to fix!)

[was]

From a Sage Days 14 user (see below).

In short, in the univariate over-a-field case, R.quotient_ring(I) should normalize the generator of the ideal before forming the quotient.

In each case below "I" and "J" are defined by different choices of
generators and are recognized as the same ideal.  In case 1 the
quotients are considered equal and in case 2 they are considered
unequal.

(I checked this with the latest version)

Case 1:
----------

sage: R.<x> = PolynomialRing(QQ)
sage: I = R.ideal([x + x^2, x])
sage: J = R.ideal([2*x + 2*x^2, x])
sage: S = R.quotient_ring(I)
sage: U = R.quotient_ring(J)
sage: I==J
True
sage: S==U
True

Case 2:
----------

sage: R.<x> = PolynomialRing(QQ)
sage: I = R.ideal([x + x^2])
sage: J = R.ideal([2*x + 2*x^2])
sage: S = R.quotient_ring(I)
sage: U = R.quotient_ring(J)
sage: I==J
True
sage: S==U
False

[AlexGhitza]

See attached patch for a fix.

[rishi]

Fixes by taking gcd with itself when only one generator is given, thus assuring a normalized generator.