Now it is set in ~/.pinerc by character-set.
The problem is in mapping the system coding names to MIME charset names, e.g., for ru_RU.CP1251 Pine should use \"windows-1251\".
A common library for such mapping would be a good solution. It could re-use a small piece of code from librecode (that deals with aliases and stores additional info about the coding systems).
Emacs has such a map (by means of a property in coding-systems definitoons).
fixed in pine 4.64.
but you should remove charset from .pinerc if you have entry there.