On 27 June 2012 01:45, Jonathan Ellis <jbel...@gmail.com> wrote: > I've started commenting on the issues, but to take a higher level > view, I'd say that it looks a lot more tractable than I thought at > first, and props to you and Eric for pushing through with it.
Thanks! > Besides the four (five including 3881) posted so far, I think 4123 > needs to be on the critical path for 1.2; otherwise, you have an > excellent chance of seriously reducing durability by deploying vnodes. Yes, I started on a replication strategy for 4123 which included DF and found that it required a lot more code changes than I had anticipated, so set it aside. This is something we can revisit now that the rest of the patches are there. I intend to write up some of my observations from the first time around. > In the meantime, 3881 can clearly be committed on its own. I'd be > most comfortable committing 4121, 4122, 4125, 4127, 4123 as a group. > From what I've seen, review should proceed fairly quickly, so > hopefully floating the earlier ones a bit longer won't add much pain. Agreed, that makes sense. -- Sam Overton Acunu | http://www.acunu.com | @acunu