We are already using Java 7 language features in 1.0 and up: try-with-resources, the diamond operator. I sometimes have to massage this when backporting to 0.98, which is >= 6, not >= 7.
I also don't think we should consider anything to do with Hadoop 3.0 until there's talk of them doing an actual release of it. As far as moving up the minimum Java version for HBase 2.0 to Java 8, this seems reasonable since we've already passed the end of public updates for Java 7. +1 > On May 7, 2015, at 9:06 AM, Sean Busbey <[email protected]> wrote: > > So we want to aim for Hadoop 3 support in hbase 2.0? > > Hard to judge amount of work when it isn't out yet. > > -- > Sean >> On May 7, 2015 8:58 AM, "Jonathan Hsieh" <[email protected]> wrote: >> >>> On Wed, May 6, 2015 at 10:28 PM, lars hofhansl <[email protected]> wrote: >>> >>> With HBASE-11666 in place, can we now start to use JDK 7 language >> features >>> across the board, starting in 1.0.x? >>> I know we had discussion, but it wasn't entirely clear whether that >>> allowed for using JDK 7 language features or not. >>> >>> >>> Or maybe just starting with 2.0? >>> Also - since JDK 7 EOL'ed last month - should we force and allow JDK 8 >>> features in 2.0? >> +1 from me for the jdk8+hbase2.0. I believe hadoop 3.0 will be jdk8 only >> and it makes hbase 2.0 to be as well. >> >> >>> -- Lars >> >> >> >> -- >> // Jonathan Hsieh (shay) >> // HBase Tech Lead, Software Engineer, Cloudera >> // [email protected] // @jmhsieh >>
