fix parent of 0th Bernoulli polynomial

[fchapoton]

CC: @videlec @orlitzky

Component: number theory

Author: Frédéric Chapoton

Branch/Commit: 078c078

Reviewer: Michael Orlitzky

Issue created by migration from https://trac.sagemath.org/ticket/32970

[fchapoton]

Commit: 756f605

[fchapoton]

New commits:

756f605

fix parent of 0-th Bernoulli polynomial

[fchapoton]

Branch: u/chapoton/32970

[orlitzky]

comment:2

I think you can avoid the coercion entirely by returning x**0. It looks 3-4x faster on my machine.

[slel]

comment:3

Would parent(x).one() work?

[orlitzky]

comment:4

Replying to @slel:

Would parent(x).one() work?

Yeah, x.parent().one() does the same thing should skip the intermediate exponentiation method.

[fchapoton]

comment:5

the only alternative is to use parent(x), as x can be an int, with no parent method.

and the the "parent" may not have any one method..

[orlitzky]

comment:6

Replying to @fchapoton:

the only alternative is to use parent(x), as x can be an int, with no parent method.

and the the "parent" may not have any one method..

Hmm, I would just stick with x**0 then. It's the most pedagogically correct option, and avoids the parent() issues. Anything that you can feed into a polynomial, by definition, should support a zeroth power that gives the corresponding "one."

[sagetrac-git]

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

078c078

using x**0

[sagetrac-git]

Changed commit from 756f605 to 078c078

[orlitzky]

Reviewer: Michael Orlitzky

[orlitzky]

comment:8

LGTM, thanks

[slel]

comment:9

Setting milestone to 9.6 now that 9.5 is out.

[vbraun]

Changed branch from u/chapoton/32970 to 078c078