Jason et al,

Here's a suggested new flag (with patch, attached) for
/usr/bin/rebaseall. It adds a -w(ait) flag which causes the check for
running Cygwin processes to be done in a loop, breaking out and doing
the rebaseall as soon as it finds a quiescent moment.

We run Cygwin on build servers which are almost always busy. When they
need a rebase (which is rare with modern Cygwin but has happened) it's
hard to find a moment when things are quiet. Of course we can take a
server out of the queue, wait for builds to finish, rebase, and put it
back but that's a slow, manual process. Running "dash -c
'/usr/bin/rebaseall -w 10'" would tell it to try every 10 seconds
until it sees an opening, then go for it and exit.

Arguably, the flag should be -r (retry) but I'll leave that to you if accepted.