Le 18/11/15 17:54, Ross Gardler a écrit : > Summarizing: > > In a healthy project I believe that the only significant things that change > between CTR and RTC are: > > 1) speed of commit (CTR is faster) > 2) quality of master, not releases (RTC catches most issues before commit, > CTR shortly after commit) > > I agree with others, nothing in the Apache Way says RTC is bad. Personally I > believe CTR builds communities faster, but there are successful RTC projects. > What really matters is managing the trade off above. > > Justification (mostly a repeat of what has been said already): > > I don't care what the website says. If I have a personal interest in a > project succeeding then I will do everything in my power to ensure it > succeeds. I assume the same is true for everyone else. This means that > "mandatory" reviews are not required because they just get done by the people > who care about project success. > > RTC does not guarantee reviews any more than CTR does, despite what a web > page says. It merely guarantees a way period in which someone will give a > patch a cursory glance. I'm not saying this is the normal RTC behavior, I'm > merely saying this is all that is guaranteed. Fortunately the process doesn't > change the way most people behave in a project, we can still trust them to do > their reviews. > > Furthermore, there seems to be an assumption that CTR means complex or > controversial code will be committed. In my experience this is not true. > People don't like to waste time writing code that may be rejected. What I see > is people discussing such changes, and providing psuedo code and then real > code for review before committing to master. It saves time to get early > review.
+1. Thi is well summarized. We don't care what the project picks (CTR or RTC), because at the end of the day, nobody can ensure the 'review' is done. The only difference between teh two modes is that you expect the +1 to be gathered before the commit in RTC, which is likely to be bypassed at some poing by those who don't have time to review thouroughly the commits. And those who think that with RTC, it will never happen, then they should be aware that it's already a battle to get PMC members to actually *review* a release befoe casting a vote :/ --------------------------------------------------------------------- To unsubscribe, e-mail: general-unsubscr...@incubator.apache.org For additional commands, e-mail: general-h...@incubator.apache.org
