work on addressing some issues with the Jacobian scaling

[zingale]

this applies if we run with scale_system=1
also as part of this change, allow the Circle theorem to use the numerical Jacobian.

[zingale]

this addresses #1242 and #1250

[zingale]

this seems to fix the scaling. For example, on development, running VODE with test_react using aprox13 and inputs_aprox13 with:

./main3d.gnu.ex inputs_aprox13

I get:

Run time = 1.375719445
min number of rhs calls: 6
avg number of rhs calls: 116
max number of rhs calls: 3932
min number of steps: 3
avg number of steps: 94
max number of steps: 3767

If I turn on scaling:

./main3d.gnu.ex inputs_aprox13 integrator.scale_system=1

then the integration is really bad:

Run time = 19.50141599
min number of rhs calls: 6
avg number of rhs calls: 1611
max number of rhs calls: 109891
min number of steps: 3
avg number of steps: 733
max number of steps: 46777

But if I run with scaling on this PR, I get:

Run time = 1.435576058
min number of rhs calls: 6
avg number of rhs calls: 117
max number of rhs calls: 3932
min number of steps: 3
avg number of steps: 94
max number of steps: 3767

almost the same as the "no scaling" result