epsilon_0 simplifies incorrectly

[Ryan James]

quantities.epsilon_0 and quantities.constants.epsilon_0 behave differently:

In [17]: epsilon_0.simplified
Out[17]: array(8.8541878176203908e-12) * s**4*A**2/(kg*m**3) # correct
In [18]: constants.epsilon_0.simplified
Out[18]: array(8.5429932106168411e-07) * s*A/m # incorrect

In [24]: constants.epsilon_0
Out[24]: UnitConstant('electric_constant', 8.854187817e-12 * F/m, 'epsilon_0') # correct
In [25]: F.simplified
Out[25]: array(1.0) * s**4*A**2/(kg*m**2) # also correct...
In [26]: x = 1*F/m
In [27]: x
Out[27]: array(1.0) * F/m
In [28]: x.simplified
Out[28]: array(1.0) * s**4*A**2/(kg*m**3) # all good here

so i have no idea why it can simplify farads correctly, and farads per meter, but not when it's constants.epsilon_0

and on a side note, the ascii symbol for quantities.mu_0 reads "epsilon_0" while quantities.constants.mu_0 is correct:

In [20]: mu_0
Out[20]: UnitQuantity('magnetic_constant', 1.25663706144e-06 * N/A**2, 'epsilon_0')
In [21]: constants.mu_0
Out[21]: UnitConstant('magnetic_constant', 1.2566370614e-06 * N/A**2, 'mu_0')

[Darren Dale (dsdale24)]

Thank you for catching this. The symbol F is commonly used for both the Farad and the Faraday constant, and the Faraday constant was registering itself as F and overwriting the farad. This can't be allowed to happen, so I added a check to make sure that previously registered units can not be overwritten. It's fixed in the trunk.

I'll try to get a new release out as soon as possible.