I disagree that snapshot dependencies are showstoppers. Convention would
use release dependencies but we currently run modified versions of these
packages until upstream can include our changes. This is similar to HBase
depending on a forked Hadoop append version until later Hadoop versions
incorporated that functionality.

All the packages are acceptable licenses.
On Aug 4, 2014 12:05 PM, "Ted Dunning" <[email protected]> wrote:

> Happy to give support.  Sorry I was out last week.  But....
>
> I just looked through the release and spotted these snapshot dependencies.
>  The 1.5.0-SNAPSHOT stuff is parquet.  The 1.0.0-mt-incubating-SNAPSHOT
> is drill-contrib-data-parent.  The 0.9-SNAPSHOT is exec-parent.  The bare
> SNAPSHOT is part of the RPM building.
>
> For different reasons, none of these should be there.  The parquet
> dependency is a complete show-stopper.
>
> ./contrib/data/tpch-sample-data/bin/pom.xml:18:
>  <version>1.0.0-m2-incubating-SNAPSHOT</version>
> ./distribution/pom.xml:150:              <release>SNAPSHOT</release>
> ./exec/java-exec/pom.xml:122:      <version>1.5.0-SNAPSHOT</version>
> ./exec/java-exec/pom.xml:137:      <version>1.5.0-SNAPSHOT</version>
> ./exec/jdbc-all/pom.xml:18:    <version>0.4.0-incubating-SNAPSHOT</version>
> ./exec/jdbc-all/pom.xml:226:
>  <baseDrillVersion>1.0.0-m2-incubating-SNAPSHOT</baseDrillVersion>
> ./pom.xml:632:            <version>0.9-SNAPSHOT</version>
>
>
>
> On Mon, Aug 4, 2014 at 8:13 AM, Jacques Nadeau <[email protected]> wrote:
>
> > Okay.  Somehow, I thought that was a requirement only on the first
> release.
> >  I guess that gives us a good reason to move to top-level :)
> >
> > I just pulled back the dist artifacts and started a thread on the general
> > incubator list.  Would love your support there.
> >
> > thanks,
> > Jacques
> >
> >
> > On Sun, Aug 3, 2014 at 11:34 PM, Ted Dunning <[email protected]>
> > wrote:
> >
> > > I just got back on-line and was looking at the release votes.
> > >
> > > That was all good.
> > >
> > > But the next step is that the actual release has to be voted on by the
> > IPMC
> > > as well.  The PPMC (drill PMC) vote is not a binding release vote
> because
> > > drill isn't top-level yet.  Until that happens, the artifacts shouldn't
> > be
> > > put up for distribution.
> > >
> >
>

Reply via email to