This patch fixes the polling behaviour of input-xwin.
It currently checks if mansync has requested a flush,
and does so 10 times a second.
This patch opens a pipe(2) between display-x and
input-xwin, so that input-xwin can sit idle in its
select(2) doing nothing, unless mansync requests a
flush (or something happens on the X socket of
This is just a proof of concept thing, it needs more
work. I'm posting here to get feedback on the general