regression in simplify_full

[zimmermann6]

In Sage 5.12:

sage: e=log(1/2*sqrt(2)*(sqrt(2) + 1)) + log(1/2*sqrt(2)*(sqrt(2) - 1))
sage: e.simplify_full()
-log(2)
sage: (2*e).simplify_full()
2*log(1/2*sqrt(2)*(sqrt(2) + 1)) + 2*log(1/2*sqrt(2)*(sqrt(2) - 1))

In Sage 5.9 the last command gave instead:

sage: (2*e).simplify_full().simplify_full()
2*log(sqrt(2) - 1) + 2*log(sqrt(2) + 1) - 2*log(2)

This is related to #10983.

CC: @orlitzky

Component: calculus

Reviewer: Paul Zimmermann, Karl-Dieter Crisman

Issue created by migration from https://trac.sagemath.org/ticket/15362

[kcrisman]

comment:1

I assume this stems from #12737? Just guessing.

[kcrisman]

comment:2

Indeed, in 5.12:

sage: (2*e).simplify_radical()
-2*log(2) + 2*log(sqrt(2) + 1) + 2*log(sqrt(2) - 1)

(Note the slight Pynac-related, unimportant, reordering.)

If I recall correctly, Paul was one of the bigger voices behind removing simplify_radical from full simplification, so hopefully #10983 can just change their doctest here.

[zimmermann6]

comment:3

replacing simplify_full by simplify_radical indeed solves the two issues we have in #10983, and it is clearer.

I thus propose to close that ticket.

Paul

[kcrisman]

Reviewer: Paul Zimmermann, Karl-Dieter Crisman

[kcrisman]

comment:4

Okay, let's close this.