We investigate the problem of exact boundary controllability of
semilinear one-dimensional wave equations. We prove that it is possible
to move from any steady-state to any other one by means of a boundary
control, provided that they are in the same connected component of the
set of steady-states. Full paper.