Unless the the repo in question needs to be CI'able by lsstsw, I have generally followed the conventional github model of fork + PR to prevent littering branches in the canonical repo. I don't believe that the published workflow process is really intended for repos which do not contain an EUPS product. I recall Jonathan Sick and I having some discussion about this at the time the guide was written, but have forgotten the details.

Review comments on PR have been addressed. It looks like travis has woken up enough to run the linux matrix cells. OSX still has not triggered yet. I'll give it an our or so for OSX to run and then merge.

John Swinbank Could you re-test in your OSX+GNU env to confirm this issue is resolved?

Thanks Josh — I can confirm this is working for me.

Not a big deal in this case, but I do request that everybody follows our regular workflow. I understand that you have well thought through reasons for not doing so, but that doesn't scale: if everybody started doing their own thing, even with the best of technical justifications, we'd wind up confused and chaotic. (And Jonathan may have written the docs, but he doesn't get to issue exemptions.)

Also — no merge commit? I'm not sure if that's just an oversight or if it was another variation on our regular workflow. I do find them very useful for tracing which ticket a particular piece of work originated on.

I'm scratching my head as to the lack of a merge commit. The only theory I can come up with is that github did a "rebase" PR merge and I mashed the button without noticing a non-standard mode was selected. There doesn't seem to a record post PR merge of the "method" used.