Conversion of PowerSeries -> SR bug

[vbraun]

If I convert a powerseries into the symbolic ring, it does not have any variables.

sage: R.<x> = PowerSeriesRing(QQ)
sage: s = R([1,2,3,4,5],prec=10)
sage: s
1 + 2*x + 3*x^2 + 4*x^3 + 5*x^4 + O(x^10)
sage: SR(s)
1 + 2*x + 3*x^2 + 4*x^3 + 5*x^4 + O(x^10)
sage: SR(s).variables()
()

Running simplify() fixes it:

sage: SR(s).simplify().variables()
(x,)

Component: symbolics

Author: Ralf Stephan

Branch/Commit: 26df71b

Reviewer: Volker Braun

Issue created by migration from https://trac.sagemath.org/ticket/10846

[vbraun]

comment:1

A possibly related bug is Trac #9769: Coercon problems to symbolic ring

[rwst]

Branch: u/rws/conversion_of_powerseries____sr_bug

[rwst]

Author: Ralf Stephan

[rwst]

Commit: 99916f9

[rwst]

New commits:

99916f9

10846: fix conversion of power series to symbolic series

[rwst]

comment:8

Deleted.

[rwst]

comment:9

Note that GiNaC will not add an order term to series from polynomials even if requested. I would consider this a bug. The reason is that pseries are formal, anyway, so they should be fully manipulatable.

This restriction can be circumvented with a simple trick.

[sagetrac-git]

Changed commit from 99916f9 to 26df71b

[sagetrac-git]

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

26df71b

10846: handle order term correctly

[rwst]

comment:11

Passes all tests in a 6.5 patchbot run.

[vbraun]

Reviewer: Volker Braun

[rwst]

comment:13

Followup in #18094 which uncovers a bug with bigoh handling.

[vbraun]

Changed branch from u/rws/conversion_of_powerseries____sr_bug to 26df71b