Protecting those branches would answer the question “If we were to change the 6.6 branch in order to release a 6.6.7, should I put the changes in 6x too” question with “no” ;).
Also, if anyone really wanted to re-open 6x they’d have a known state. Well, not 6x since I’m certain some things have been committed there that haven’t been committed to 6_6… Er… See what I mean? But going forward…. In some weird case where we wanted to release a 6.7 (which I don’t see happening frankly), we could open up the 6x branch again and bump the branch on a case-by-case basis. After “frank and open" discussions about how that’s not our policy.... In general I’m in favor of being unable to screw something up so protecting all the branches we shouldn’t modify from writes is a +1. I’m -1 for removing branches. Erick P.S. And I do note that you carefully did _not_ include 6_6 or 7_7 so we can still do the point releases. > On Apr 5, 2019, at 7:47 AM, Adrien Grand <[email protected]> wrote: > > Removing branches proved a bit controversial in the past, > seehttps://markmail.org/message/6ah3m6zd6v3ik3ie for instance. > > On Fri, Apr 5, 2019 at 3:24 PM Shawn Heisey <[email protected]> wrote: >> >> On 4/5/2019 3:19 AM, Adrien Grand wrote: >>> Apparently it is possible to ask INFRA to mark branches as >>> protected[1]. Should we do it for branches that are not expecting new >>> releases anymore? I think it would make things less trappy. For >>> instance, backporting to branch_7x is almost for sure a mistake since >>> we are not going to release lucene/solr 7.8. >>> >>> [1] https://issues.apache.org/jira/browse/INFRA-18109 >>> >>> What would you think of protecting the following branches: >>> branch_3x >>> branch_4x >>> branch_5_4 >>> branch_5_5 >>> branch_5x >>> branch_6_0 >>> branch_6_1 >>> branch_6_2 >>> branch_6_3 >>> branch_6_4 >>> branch_6_5 >>> branch_6x >>> branch_7_0 >>> branch_7_1 >>> branch_7_2 >>> branch_7_3 >>> branch_7_4 >>> branch_7_5 >>> branch_7_6 >>> branch_7x >> >> In the SVN days, branches were simply deleted when we were through >> cutting releases from them. Deleted branches all came back when we >> converted to git. >> >> Is protecting them the only way to maintain the history? If so, then >> I'm +1. Does deleting them like we did on SVN make it impossible to "go >> back in time" in the repo? >> >> Thanks, >> Shawn >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: [email protected] >> For additional commands, e-mail: [email protected] >> > > > -- > Adrien > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
