Wrong indefinite integral of some odd functions between opposite endpoints

[sagetrac-tmonteil]

As reported on this ask question, Maxima seems to know that integrating an odd function between two opposite endpoints leads to zero, however it applies this rule even when the integral is not defined:

sage: integrate(sin(x), x, -oo, oo)
0
sage: integrate(tan(x), x, -oo, oo)
0

However,

sage: integrate(arctan(x), x, -oo, oo)
ValueError: Integral is divergent.

Upstream: Reported upstream. No feedback yet.

CC: @sagetrac-jakobkroeker

Component: symbolics

Keywords: maxima, integration

Stopgaps: wrongAnswerBug

Issue created by migration from https://trac.sagemath.org/ticket/17109

[kcrisman]

Upstream: Reported upstream. No feedback yet.

[kcrisman]

comment:1

I've reported this upstream here. You are right that this is not the usual definition of such an improper integral - though of course sometimes it's fun to mess with students this way!

[sagetrac-tmonteil]

comment:2

Thanks !

[sagetrac-jakobkroeker]

Stopgaps: wrongAnswerBug

[fchapoton]

Changed keywords from maxima to maxima, integration