The problem of astrophysical site(s) for the r-process is still shrouded
in mystery despite more than a half century of studies. Although
neutrino driven winds have been considered to be the most likely
success, recent supernova simulations show that it is seriously
difficult to achieve suitable conditions for r-process in the environment.
I will talk about new ideas for the r-process astronomical sites. I will
show the results of r-process nucleosynthesis based on the core-collapse
supernova senario via the Quark-hadron phase transition during the early
post-bounce phase. In addition, I also want to briefly remark the
progress and preliminary results of r-process studies for magnetically
driven core-collapse supernovae.