On Sun, Mar 3, 2013 at 8:50 PM, Harsh J <ha...@cloudera.com> wrote: > Have we agreed (and stated it somewhere proper) that a -1 obtained for > a Windows CI build for a test-patch will not block the ongoing work > (unless it is Windows specific) and patches may still be committed to > trunk despite that? >
This thread is long and possibly hard to follow. Yes, I and several others have stated that for now it is okay to commit even if Windows precommit build posts -1. > > I'm +1 if someone can assert and add the above into the formal > guidelines. I'd still prefer that Windows does its releases separately > as that ensures more quality for its audience and better testing > periods (and wouldn't block anything), but we can come to that iff we > are unable to maintain the currently proposed model. Which do you think is the right place to add this? At this time we are voting for merging into trunk. I prefer having a single release that supports both Linux and windows. Based on working on Windows support I think this is doable and should not hold up releases for Linux.
