Support extended precision

[simonp0420]

Each exported function should have one or more unit tests where it is exercised by extended precision types such as BigFloat. Some of the functions will need some recoding to support this.

[stla]

Done for the Carlson functions.

[simonp0420]

Here is a checklist we can use to track completion of this issue:

• [ ] agm
• [ ] am
• [ ] CarlsonRC
• [ ] CarlsonRD
• [x] CarlsonRF
• [ ] CarlsonRG
• [ ] CarlsonRJ
• [ ] EisensteinE2
• [ ] EisensteinE4
• [ ] ellipticE
• [ ] ellipticF
• [ ] ellipticInvariants
• [ ] ellipticK
• [ ] ellipticPI
• [ ] ellipticZ
• [ ] etaDedekind
• [ ] halfPeriods
• [ ] jellip
• [ ] jtheta1
• [ ] jtheta1dash
• [ ] jtheta2
• [ ] jtheta3
• [ ] jtheta4
• [ ] kleinj
• [ ] kleinjinv
• [ ] lambda
• [ ] ljtheta1
• [ ] ljtheta2
• [ ] ljtheta3
• [ ] ljtheta4
• [ ] thetaC
• [ ] thetaD
• [ ] thetaN
• [ ] thetaS
• [ ] wp
• [ ] wsigma
• [ ] wzeta

Edit: Unchecked ellipticF and checked CarlsonRF.