[bug] pyprima does not adjust linear/nonlinear constraints when removing fixed bounds

[nbelakovski]

The following simple test program raises and error due to a size mismatch with the linear constraints:

from pyprima import minimize, LinearConstraint as LC
import numpy as np

def f(x):
        return np.sum(x**2)

lb = [-1, None, None]
ub = [-1, None, None]
bounds = [(a, b) for a, b in zip(lb, ub)]
lc = LC(np.array([1, 1, 1]), lb=9, ub=15)
res = minimize(f, x0=np.array([1, 1, 1]), constraints=lc, bounds=bounds)

The error arises because while we have code to eliminate variables with fixed bounds, we are not updating the linear constraint matrix (or the nonlinear constraint function) to take into account the new size of x.