partially this is a RFC, so feel free to drop your thoughts in

upstream, upstart is dead as Ubuntu is mov(ing|ed) to systemd.  just stating facts here, not soliciting opinions.

even if upstart was active, i'm not sure it'd matter too much to us.  we've stayed on v1.2 even though the latest is v1.13.2.  this is because upstream added a lot more (unconditional) features that we simply have no need for (e.g. multiseat support).

however, upstream has added some features we do find interesting, or various bug fixes, and we've backported them as dedicated patches that get applied in the ebuild.

we've also added some unique/custom work in upstart (like selinux & keychain support).  reviewing that work has not been easy as we can't browse the larger context, only the patch file itself.

stats on our current patches:
- 14 patches
- smallest is 320 bytes but largest is 130K (!)
- for all the patches: 61 files changed, 4783 insertions(+), 534 deletions(-)

we're now also talking about doing more custom work.  i think we've passed the point where dedicated cut patches are the best way to work on upstart.

i propose:
- create a chromiumos/third_party/upstart/ git repo
- pull in the existing upstream history and put all their branches under refs/heads/upstream/ (like we've done with some other projects)
- start the master branch at the v1.2 release
- add all our patches on top of that
- convert existing upstart ebuild over to cros-workon
- all future work will be in the git repo

this should be easier to cherry pick back upstream changes and track the metadata.