> On 6 May 2015, at 17:50, Vinod Kumar Vavilapalli <[email protected]> 
> wrote:
> 
> +1 for the idea in general.
> 
> We definitely need design discussion and consensus upfront, lest branches 
> turn into dumping grounds and people complaining about why their branches 
> (instead of patches) do not get merged.
> 
> As others noted, for the same reason above, active supervision of branches is 
> desired.
> 
> Echoing Bikas and Zhijie's comments, if two branches are completely 
> orthogonal (may be because it is all new features/code), the order doesn't 
> matter. But if two proposals potentially have a large conflict of code-base 
> or functionality, we should discuss and arrive at a branch-order upfront so 
> as to avoid thrash.
> 
> IAC, I'd definitely like us to avoid "why is this not being done in a branch" 
> conflicts. More often than not, it ends up being a subjective decision. We 
> can actively try to push big enough features into their branches, but we 
> should be open to cases where that may not make all the sense.
> 
> Thanks
> +Vinod


one more thing: anyone who branches needs to set up a jenkins build for that, 
emails sent to them & not the core mailing lists

Reply via email to