On 17 Oct 2016, at 18:26, Ryan Blue
Are these changes that the Hive community has rejected? I don't see a
compelling reason to have a long-term Spark fork of Hive.
More changes in hive that haven't been picked up
HIVE-11720 is needed to handle very long HTTP headers, which is exactly the
kind of header Active Directory likes to generate
the other one fixes pom dependencies to stop groovy-all getting into to spark.
That's been fixed in sparks own POMs; pushing into the spark hive fork ensures
that nobody else gets it. This matters as you can abuse serialisation and have
SparkContext.objectFile() run arbitrary code in the loader's process. Not ideal.
On Sat, Oct 15, 2016 at 5:27 AM, Steve Loughran
On 15 Oct 2016, at 01:28, Ryan Blue
The Spark 2 branch is based on this one:
Didn't know this had moved.... I had an outstanding PR against patricks which
should really go in, if not already taken up ( HIVE-11720 ;
IMO I think it would make sense if -somehow- that hive fork were in the ASF;
it's got to be in sync with Spark releases, and its not been ideal for me in
terms of getting one or two fixes in, the other one being culling
groovy 2.4.4 as an export (
I don't know if the hive team themselves would be up to having it in their
repo, or if committership logistics would suit it anyway. Otherwise,
approaching infra@ and asking for a forked repo is likely to work with a bit of
On Fri, Oct 14, 2016 at 4:33 PM, Ethan Aubin
In an email thread  from Aug 2015, it was mentioned that the source
to org.spark-project.hive was at
That branch has a 1.2.1.spark version but spark 2.0.1 uses
1.2.1.spark2. Could anyone point me to the repo for 1.2.1.spark2?
To unsubscribe e-mail: