On Fri, May 24, 2013 at 10:10:25PM +0000, Animesh Chaturvedi wrote: > [Animesh>] Plan means targeted for 4.2. Community review and technical > quality are of course important and being attended to. The effort to find and > fix issues with this change speaks for the attention to technical quality. I > am simply asking how much time is needed to review.
This discussion is going in circles Animesh. The answer to your question above: as much time as is needed. If the details of a set of changes were created / debated on this list (which these were not really), the review would be much shorter and easier! However, in this case, we are looking at 25 pages of review content on reviewboard, impacting many different features. Let me restate my primary concern (the one that I wish we would focus on): Why would we bring in a significant architectural change late in a release merge cycle? It hurt us to do it with Javelin, and I just don't see *any* valid argument to do it again here. I've said it before in this thread, but I'll say it again: I'm not judging the work. I'm judging the wisdom of the timing of the merge. Why can't we just wait until 4.2 is branched, and then bring this into master? It would go out with 4.3. Is there some sort of unspoken pressure that is causing you to argue against our previous consensus here? I'm really struggling to see what would make us want to do this level of change right before we cut a release branch *again*. I'm especially confused about the need to merge into master when the only delay I'm suggesting is 1 week. It's not like I'm saying "wait for next year". -chip
