I would rather focus on getting our language manual working: https://cwiki.apache.org/Hive/languagemanual.html https://cwiki.apache.org/Hive/languagemanual-variablesubstitution.html
or that other jira ticket to move our site to that content management system, or another initialize to move our language manual to version control, or work on moving hive to maven, Then focus on something that is not actually a problem like discontinuing support for versions. On Thu, Sep 19, 2013 at 11:41 AM, Owen O'Malley <omal...@apache.org> wrote: > +1 to dropping Hadoop 0.20.2 support in Hive 0.13, which given that Hive > 0.12 has just branched means it isn't likely that Hive 0.13 will come out > in the next 6 months. > > -- Owen > > > On Thu, Sep 19, 2013 at 8:35 AM, Brock Noland <br...@cloudera.com> wrote: > > > First off, I have to apologize, I didn't know there would be such > > passions on both sides of the 0.20.2 argument! > > > > On Thu, Sep 19, 2013 at 10:11 AM, Edward Capriolo <edlinuxg...@gmail.com > > > > wrote: > > > That rant being done, > > > > No worries man, Hadoop versions are something worth ranting about. > > IMHO Hadoop has a history of changing API's and breaking end users. > > However, I feel this is improving. > > > > > we can not and should not support hadoop 0.20.2 > > > forever. Discontinuing hadoop 0.20.2 in say 6 months might be > reasonable, > > > but I think dropping it on the floor due to a one line change for a > > missing > > > convenience constructor is a bit knee-jerk. > > > > Very sorry if I came across with the opinion that we should "drop > > 0.20.2 now" because of the constructor issue. The issue brought up > > 0.20.2's age in my mind and the logical next step is to ask how long > > we plan on supporting it! :) I like the time bounding idea and I feel > > 6 months is reasonable. FWIW, the 1.X series is stable for my needs. > > > > Brock > > >
