extField from RationalPoints2 overtakes symbol "a"

Nobody has claimed this yet.

The documentation warns:
> **N.B. Any data previously stored in this symbol will be lost!** As is the case for GaloisField.

I think this behavior, which is caused by calling `getSymbol`, should be fixed in GF as well and discouraged in other packages. Here is what can happen:

```m2
i1 : needsPackage "RationalPoints2";

i2 : R = QQ[a,b];

i3 : T = ZZ/2[t];

i4 : extField { t^2+t+1 }

o4 = GF 4

o4 : GaloisField

i5 : a

o5 = a

o5 : GF 4
```
The problem is that a call to `extField` could be happening somewhere deep in an algorithm, and consequently "a" will be taken over without the user knowing where it happened.

Start by tracing extField and getSymbol in GaloisField and RationalPoints2, reproducing the provided example to identify where symbol "a" is replaced. Done should prevent an unexpected overwrite in GF and establish how similar behavior is discouraged in other packages.

Written by the indexing model from the issue text.