In lazy power series rings, 0 + a and 0 - a return 0

[darijgr]

sage: L = LazyPowerSeriesRing(QQ)
sage: x = L.gen()
sage: L.zero().coefficients(4)
[0, 0, 0, 0]
sage: x.coefficients(4)
[0, 1, 0, 0]

Correct so far...

sage: (L.zero() + x).coefficients(4)
[0, 0, 0, 0]
sage: (L.zero() - x).coefficients(4)
[0, 0, 0, 0]

But x - L.zero() and x + L.zero() return correct results.

See also #15248 for another lazy power series bug.

CC: @mwhansen @mantepse @tscrim

Component: algebra

Keywords: LazyPowerSeries, species

Stopgaps: incorrectAnswerMarker

Reviewer: Travis Scrimshaw

Issue created by migration from https://trac.sagemath.org/ticket/15249

[mantepse]

comment:1

One problem is in series_order.py. There, a very simple lattice (in fact, a total order) should be implemented: NonNegativeIntegers together with the two special values unknown and infinity. unknown is the minimal element, infinity the maximal element.

Comparing unknown with another element works, comparing infinity with another element works, but comparing an integer with unknown does not work, because the comparison invoked is from Integer.

How should I create such a lattice really?

[mwhansen]

comment:3

I've done some work on species recently which fixes this as a side effect. This is related to #10084. I'll post the branch asap.

[mantepse]

Changed keywords from lazy-power-series, species to LazyPowerSeries, species

[mantepse]

comment:5

Replying to @mwhansen:

I've done some work on species recently which fixes this as a side effect. This is related to #10084. I'll post the branch asap.

Great! I'm all set to dig into the species code too. For a start, I looked through the tickets that are related to the lazy power series code and changed the keyword to LazyPowerSeries...

Did you see #10931? I guess the patch won't apply anymore, but the philosophy should :-)

Best, Martin

[mwhansen]

comment:6

A fix is in #15673