Re: [discuss] So what does Gerrit actually get us, anyway?
Please note, the following is a subjective biased view, responding to a subjectively phrased question. It's not intended to be an objective comparison and will conflate a number of concepts together.
I've dealt a lot with the OpenStack development workflow , and very little with Github's workflow. In the OpenStack workflow, Gerrit is primarily used for review, by humans and by CI. In some respects, it's also used for planning (when specs are used), and it's an excellent tool for doing collaborative review on documentation, with good traceability. In contrast, it's hard to find any collaborative review on review.opencontrail.org, by human or by CI, with virtually no comments by humans and frequent 'rechecks' being thrown at the CI to try to get it to pass.
Giving up Gerrit will likely drive TF towards a more monolithic design. Committing to Gerrit properly can help modularise TF into separate, upgradeable components. Personally, I don't find Gerrit a misery when the community is extremely helpful and forthcoming.
I think I might have set the tone a little too aggressive for this message. If it were me, I'd look at the workflows that developers of adjacent and competing projects are using and try to fit in closer with them.
Principal Software Engineer
Netronome | First Floor Suite 1, Block A, Southdowns Ridge Office Park,
Cnr Nellmapius and John Vorster St, Irene, Pretoria, 0157
Phone: +27 (12) 665-4427 | Skype: jangutter | www.netronome.com