On Wed, May 28, 2014 at 9:01 AM, Bill Havanki <[email protected]>wrote:
> Is it possible to break the review into smaller, easier-to-review chunks? > I like that. Could construct a dependency graph of the chunks to be reviewed and post that as a map for us. > > > On Tue, May 27, 2014 at 9:16 PM, Josh Elser <[email protected]> wrote: > > > Well, organizing something at the summit is a possibility, but I'm > already > > signed up to give a talk there on the subject, so I'd rather just > leverage > > that instead. > > > > I don't want to just wave the "CTR flag" and merge it into master, but, > at > > the same time, I want to make sure that those who *are* interested in > > getting a preview have the opportunity to do so. I also recognize that > the > > code is rather intimidating to just sit down and look at via reviewboard. > > > > Overall, if people are genuinely interested in getting an overview from > an > > implemented stance, I'm happy to provide that. But, at the same time, I > > don't want people to feel obligated to review if they really aren't > > interested. Does that make sense? > > > > > > On 5/27/14, 8:46 PM, David Medinets wrote: > > > >> Perhaps snag a few people during the summit and lock them in a room for > >> the > >> review? > >> > >> > >> On Tue, May 27, 2014 at 8:39 PM, Josh Elser <[email protected]> > wrote: > >> > >> Everyone, > >>> > >>> It's been about a week and I haven't gotten any feedback on the review > >>> [1] > >>> I put up for the replication work yet, so I'm wondering if it would be > >>> better to take a different approach. > >>> > >>> The tests I've written (which are included in the review), give me > quite > >>> a > >>> bit of confidence in the implementation as designed, not to mention the > >>> "higher-level" tests I've run. > >>> > >>> So, I'm wondering if it would be better to find a good time to do some > >>> sort of dev-focused video presentation (google hangout, likely), and > >>> give a > >>> verbal overview. Would this be something that people would prefer > instead > >>> of looking through ~20k lines of code? Do people care either way? > >>> > >>> I'm happy to try to make this as easy as possible. > >>> > >>> - Josh > >>> > >>> [1] https://reviews.apache.org/r/21752/ > >>> > >>> > >> > > > -- > // Bill Havanki > // Solutions Architect, Cloudera Govt Solutions > // 443.686.9283 >
