On 2 Dec 2016, at 19:09, Reynold Xin <r...@databricks.com<mailto:r...@databricks.com>> wrote:
ThriftHttpCLIService.java code is actually in Spark. That pull request is basically no-op. Overall we are moving away from Hive dependency by implementing almost everything in Spark, so the need to change that repo is getting less and less. that's good, but currently there is still a hive dependency, and as it's a fork of hive, no easy way to get anything needed for Hadoop 3.x in. I suspect Imran Rashid owns that problem... On Fri, Dec 2, 2016 at 10:03 AM, Marcelo Vanzin <van...@cloudera.com<mailto:van...@cloudera.com>> wrote: I believe the latest one is actually in Josh's repository. Which kinda raises a more interesting question: Should we create a repository managed by the Spark project, using the Apache infrastructure, to handle that fork? It seems not very optimal to have this lie in some random person's github account. Given how often it changes, it shouldn't really be too much overhead to maintain it. On Fri, Dec 2, 2016 at 3:57 AM, Steve Loughran <ste...@hortonworks.com<mailto:ste...@hortonworks.com>> wrote: > > What's the process for PR review for the Hive JAR? > > I ask as I've had a PR for fixing a kerberos problem outstanding for a > while, without much response > > https://github.com/pwendell/hive/pull/2 > > I'm now looking at the one line it would take for the JAR to consider Hadoop > 3.x compatible at the API level with Hives 2.x, so that Dataframes work > against Hadoop 3.x; without that ASF Spark is not going to work on Hadoop 3. > > Where should I be submitting those PRs, and what'st the review process? > > thx > > -steve -- Marcelo --------------------------------------------------------------------- To unsubscribe e-mail: dev-unsubscr...@spark.apache.org<mailto:dev-unsubscr...@spark.apache.org>
