gcd is not commutative (on rationals)

[vdelecroix]

sage: gcd([1/1,1/2])
1
sage: gcd([1/2,1/1])
1/2

Though the gcd is perfectly well defined on rationals

sage: (1/1).gcd(1/2)
1/2
sage: (1/2).gcd(1/1)
1/2

[rws]

It's because the loop through the sequence is interrupted as soon as the partial GCD is one, which is only valid for integer GCD:

    for vi in v:
        g = vi.gcd(g, **kwargs)
        if g == one:
            return g

[vklein]

if g == one:
            return g

has been removed.

---

New commits:

​0a599ea

Trac #24880: fix __GCD_sequence

[slabbe]

Documentation won't build correctly as you need double :: before a doctest block of code (after TESTS).

[git]

Branch pushed to git repo; I updated commit sha1. New commits:

​ce541c3

Trac #24880: TESTS: => TESTS::

[vklein]

Replying to slabbe:

Documentation won't build correctly as you need double :: after a doctest block of code (after TESTS).

Fixed