On 29/03/2017 12:30 , Vitalii Koshura wrote: > I will not do any any cherry-picks because I have write access to the repo.
You mean "no write access"? > If you will look at the current branch tree you will see that every new > release branch bases on previous release branch. And master branch is a > sandbox. It contains some commits which are too risky to be included > into release. Both of that needs to be changed anyway to ensure proper release management. Since master is used for integration is should always be as stable as possible (don't break master!). That usually means development has to happen in dedicated feature branches. This also ensures that independent developments don't affect each other which will make the whole process more efficient. > Also if you will take a look at any release banch you will > see that all commits were integrated there separately one by one. This doesn't mean that this approach is sensible or that it should be continued in the future. Again, cherry-picking like that is error-prone (as you said, it's hard) and cumbersome. Our goal needs to be to make the release process much safer/simpler and even more efficient/faster. > I think it is the best way to continue this work. I strongly disagree for the reasons above :-) > If you think this is a wrong way to do this please suggest your own. It's basically what I outlined in my previous mail and above. Cheers, Oliver _______________________________________________ boinc_dev mailing list boinc_dev@ssl.berkeley.edu https://lists.ssl.berkeley.edu/mailman/listinfo/boinc_dev To unsubscribe, visit the above URL and (near bottom of page) enter your email address.
