How it works:
Assumes you have a list of repos in a file called
./all.sh to fork all the repos and generate patches.
Probably a bunch of them will break. Take the output of that
command and pull out the ones that worked (lines ending in "ok")
and put those repos in another file called
and it'll send pull requests for the ones in that file.
Then look at the log in
do.log and try to figure out why
the broken ones failed, and maybe fix it and rerun just those ones.