Workaround for failing complex_double test on Cygwin

[embray]

I fixed this a while ago, but realized I don't think I ever made a ticket for it.

This test fails due to minor numerical differences (again, probably originating in the log() function) so that the polynomial returned by z.algdep(5) ends up being slightly different (it gets an extra factor of x^2). This makes the test a little more flexible to such differences.

Component: porting: Cygwin

Branch/Commit: u/embray/bug/cygwin/complex_double_test @ 77feef0

Reviewer: Erik Bray

Issue created by migration from https://trac.sagemath.org/ticket/21605

[embray]

comment:1

#21292 remains stalled, but this simple test fix doesn't require it; better to just fix this.

[embray]

Changed dependencies from #21292 to none

[jdemeyer]

comment:2

#22759 solves the same problem in a different (arguably better) way.

[embray]

comment:3

Oh, that's new--I'll have a look.

[embray]

comment:4

I see now--this should fix the underlying problem I think, but I'll have to test it.

[tscrim]

comment:5

Since #22759, I think we can close this as a dup.

[jdemeyer]

Changed author from Erik Bray to none

[jdemeyer]

Reviewer: Erik Bray