Policy for review / approval of PRs

[billsacks]

For extensive PRs, I assume we want at least one reviewer from each of CESM and E3SM.

But what are people's current feelings on small PRs like this recent one from @alperaltuntas - #4944? Do we generally want at least one reviewer from each of CESM and E3SM before merging a PR like this, or do people feel okay with just a review from the CESM side (or E3SM side)? And @mvertens , to what extent do you want a NorESM review on PRs before merging them? This particular PR only impacts MOM, so seems safer to merge without extensive review, but I've been wondering what people want as a general policy.

@jasonb5 @jgfouca @rljacob @jedwards4b @briandobbins @fischer-ncar @mvertens @samsrabin

[mvertens]

@billsacks - thanks so much for reaching out about this. I don't think it's necessary to have a NorESM reviewer on PRs before merging them. I'm include @mvdebolskiy in case he has a different opinion.

[jgfouca]

@billsacks , for a PR like that, I'd say just get it merged as quickly as possible and don't worry too much about getting reviews from people who won't be impacted. However, I will say that PR is at least a little interesting in that it highlights limitations of config_archive that force us to hardcode special cases directly into the python. I've thought for years now that config_archive could use a pretty serious rework.

[jedwards4b]

I have tried to use discretion and not involve the E3SM developers in PR's that do not affect them. I haven't got that right 100% of the time but certainly most of the time.
On the other hand I seek an E3SM review if the change will affect both models.

[billsacks]

Okay, sounds good - thanks, all. As I imagine I'll be taking over for @jedwards4b 's reviewer role more often moving forward, I'll welcome feedback if I'm involving E3SM too much or too little in various reviews.

[mvdebolskiy]

@billsacks you can always ping me on any PR that you think might have significant impact on NorESM. Though, I reviewing might be not necessary 90% of the time.