[AstroPy] Wrong earth mass in astropy.constants

[Andrej Prsa]

> The IAU passed a resolution relatively recently addressing exactly
> this problem, and recommending standard values, while acknowledging
> that best estimates might continue to diverge slightly from these.
> So it seems like these IAU standard values would be a good option to
> consider for use in astropy: 
> 
> https://arxiv.org/abs/1510.07674
> 
> Resolution B3 here:
> 
> https://www.iau.org/static/resolutions/IAU2015_English.pdf

Note that these are /nominal/ values rather than current best
estimates. They are by definition exact and are used as rulers, not as
CBEs. See this for an extended discussion:

	http://adsabs.harvard.edu/abs/2016AJ....152...41P

We override the constants from astropy in phoebe to conform to nominal
units (which is arguably the only suitable set for our work as per the
reference above). Thus, package subsectioning sounds like a good
solution for these issues.

My 2c worth,
Andrej