GitLab

use -pipeline_url instead of defining EHIVE_USER and EHIVE_PASS for all tests, and drop the database afterwards

This can solve issues about quoting the arguments

Join the arguments if they are given as an array-ref, and take care of the
quotes !

schema_change: removed analysis_stats[_monitor].num_required_workers, substituted with AS::estimate_num_required_workers()

It now takes a DBConnection object instead of a dbc_hash

It allows to capture stderr, and output a more useful message when the command fails.

The snapshot is taken whilst the dumping job is still running. It means that
on a restored database, beekeeper will first garbage-collect the job, and
then re-run it.

Previously, the job would re-dump everything and overwrite the snapshot.
This is not the case any more.

Conflicts:
	modules/Bio/EnsEMBL/Hive/DBSQL/CoreDBConnection.pm

renamed shortened DBC parameters to their longer versions; made sure they are all passed to the constructor

avoiding RELOCATED workers by applying incemental backoff-and-retry approach from Ethernet CSMA/CD protocol in dbc::connect()

bugfix: stop registering some NO_WORK workers as UNKNOWN, by changing the order of Valley/DB reconciliation

bugfix: avoid getting negative analysis_stats.ready_job_count by separating the "grab" and "release" phases of batching

bugfix: avoid getting negative analysis_stats.ready_job_count by separating the "grab" and "release" phases of batching

bugfix: avoid getting negative analysis_stats.ready_job_count by separating the "grab" and "release" phases of batching