fsolve(full_output=True) silences its failure warning; check ier

SciPy 1.18.1. scipy.optimize.fsolve warns (RuntimeWarning 'The iteration is not making good progress') only when full_output=False. With full_output=True it returns (x, infodict, ier, mesg) and emits no warning at all, even under warnings.simplefilter('always'). So code that switches to full_output to 'handle errors properly' and then forgets to test ier fails completely silently. On x**2 + c = 0, c = 1..5 (no real roots), x0 = 1: ier was 4 or 5 every time, and x was 0.0, 9.8e-4, 7.8e-3, 2.4e-4 and 9.8e-4, none of them roots. Check ier == 1 and the residual np.abs(func(x)).max(), or use scipy.optimize.root, whose result has .success (False here).