Wrong simplification in symbolic trigonometry involving fractions

[sagetrac-tmonteil]

As reported in this ask question, Sage makes wrong simplifications in symbolic trigonometry involving fractions:

sage: cos(pi+pi*1/42)
cos(1/42*pi)

While it should be -cos(1/42*pi).

CC: @rwst

Component: symbolics

Issue created by migration from https://trac.sagemath.org/ticket/20752

[rwst]

comment:1

Same with sin(pi+pi/42).

[sagetrac-tmonteil]

comment:2

I do not have intermediate versions of Sage with me, but i confirm that it is a regression since on Sage 5.12, we have:

sage: cos(pi+pi*1/42)
cos(43/42*pi)
sage: sin(pi+pi/42)
sin(43/42*pi)

[rwst]

comment:3

No, the changes were just a few months ago, will get to it tomorrow.

[sagetrac-tmonteil]

comment:4

More precisely, on a 7.2.beta4 version of Sage which i can reach by ssh, it works well:

sage: cos(pi+pi*1/42)
cos(43/42*pi)
sage: sin(pi+pi*1/42)
sin(43/42*pi)

So the regression is very recent, could it be due to the last update of pynac/ginac ?

[kcrisman]

comment:7

So the regression is very recent, could it be due to the last update of pynac/ginac ?

Almost certainly so. See e.g. this bug, though I assume it is a different particular one.

[rwst]

comment:8

This is fixed and doctested in #20742.

[embray]

comment:10

Determined to be invalid/duplicate/wontfix (closing as "wontfix" as a catch-all resolution).