GitLab

pipeline_name is now automatically computed from ClassName; simplified workshop's example files and slides

cleaned up the last (optional) slide on pipeline_wide_parameters; removed the exercise about abstracting out the compressor (formerly from CompressFiles_conf)

separated out "long addition" functionality to concentrate on Hive API when writing the Runnable, and not on maths

bugfix: ENSCOMPARASW-132. When all dependent jobs (>1) fail to be created due to unique constraint, they now correctly update status to READY

bugfix: reset the tried jobs to retry_count=1 and untried ones to retry_count=0 when doing a bulk reset

methods dbconn_2_mysql(), dbconn_2_pgsql(), db_connect_command(), db_execute_command() are DEPRECATED - use db_cmd.pl instead

added support for -pipeline_url as an input parameter (no need to supply hive_driver or password in this case)

now performing deep-stack substitution for whatever is dataflown into tables (rather than just dataflowing the output_id)

bugfix: substituting the accu signature on demand from the very depths of emitting job's param_stack

bugfix: make sure longer input_id hashes are correctly fetched from analysis_data table in "param stack" mode

Simplified interface: now db_cmd.pl understands 'CREATE DATABASE' and 'DROP DATABASE' without parameters, given a full URL.