Perhaps you misunderstood. I meant:
u(t) \ne \int_{-\infty}^\infty u(t+4) e^{t(-1-i\omega)-4}\,dt
The integral in the second line is fine to calculate the Fourier transform you're looking for, but it's definitely not equal to u(t) as you wrote.
Try the substitution t'=t+4 to evaluate the integral.