Euler numbers revamp

[rwst]

• move euler_number from combinatorics to functions
• replace pexpect-maxima with maxima-lib implementation
• make GP/Pari the main version with euler(n)=subst(bernpol(2*n+1), 'x, 1/4)*4<sup>(2*n+1)*(-1)</sup>(n+1)/(2*n+1) (Charles R Greathouse IV)
• mpmath has fp Euler numbers

Component: symbolics

Issue created by migration from https://trac.sagemath.org/ticket/17770

[jdemeyer]

comment:1

What you're proposing seems like 4 tickets, not 1 ticket...

[rwst]

comment:2

Replying to @jdemeyer:

What you're proposing seems like 4 tickets, not 1 ticket...

It's not substantially more than any other symbolic function ticket. Not that there is heavy use that would justify several tickets.

[jdemeyer]

comment:3

Replying to @rwst:

It's not substantially more than any other symbolic function ticket. Not that there is heavy use that would justify several tickets.

Neither of this matters. The question is: can the issue be easily and naturally be split up? If the answer to that is yes, it should be done.

I have just seen too often that a ticket does A, B, C and there are issues with C. These issues with C turn out to be much harder than originally thought. Therefore, A and B are never merged or merged with a big delay, even though they could have been merged quickly.

Note: personally I don't care much about this ticket, so if a reviewer doesn't mind, then it's obviously not a problem.

[rwst]

comment:5

See #20763