DB group, I am exploring the possibility of making sqlline an Apache DB sub-project.
Sqlline is a JDBC shell along the lines of Oracle's SQL*Plus and Derby's ij, but it is generic: it works with any JDBC driver. It was founded about 10 years ago, and was forked (without malice) into Hive as beeline and into my github project https://github.com/julianhyde/sqlline. Both forks have seen reasonable use (Hive's by the Hive community, and mine by Drill, Phoenix, Cascading among others), ongoing development activity, and regular releases. I am currently undertaking the painful task of "un-forking" -- producing a development line that contains the union of all patches, without unnecessary dependencies such as Hive. The Hive-dependent pieces of beeline would remain in sub-classes in beeline. See https://issues.apache.org/jira/browse/HIVE-6361 for details. For pragmatic reasons, I will be un-forking into a new Hive module, org.apache.hive.sqlline. But there is a concern that the project will over time become too dependent on Hive. Even if we commit to not introduce dependencies, perception is reality. Also, project development would have to occur on Hive's (very busy) dev list, and, pragmatically, its releases would have to be synchronized with Hive's releases. As a sub-project of DB, sqlline could continue to be developed in the Apache way but with more room to breathe. Do you think sqlline would be a good fit as a DB sub-project? Julian
