very rare circumstances, caused error messages to be sent to space

size (flushes)

Committed Kathryn changes. Added runnable and output subroutine. These should make the coding of Bio::EnsEMBL::Analysis::RunnableDB and Bio::EnsEMBL::Hive::Process more alike.

When -logic_name is given, now only returns/creates the number of workers needed precisely for this analysis.

Added an optional Bio::EnsEMBL::Analysis argument to get_num_needed_workers subroutine, to only get the number of workers needed for this analysis

design element in Worker::batch_size
- old design: setting the batch_size for a worker changed the batch_size for
  all workers by updating the analysis_stats.batch_size of the corresponding
  analysis. Although this code was never triggered in the current system, it
  could have caused potential problems for future developers.
Now setting the batch_size explicitly for a worker only effects the
batch_size for that Worker instance. If it has not been set, the Worker pulls
the batch_size from the analysis_stats record and optionally autocalcs from
avg_msec_per_job if the batch_size<=0.

gets large

added a new debug printing of each job as such
  job_id=9050 blast_1_DEFAULT(17) retry=0 input_id='982'
if the worker has debug turned on

to use a logic_name rather than an analysis_id

  removed option -analysis_stats
  added option -no_analysis_stats

-regfile and -regname by default it looks for the alias 'hive' in the
registry if -regname is not specified

are neither done nor failed.  Changed print_stats to show remaining_job_count
instead of done.  Since this number now counts done to zero, it is much
easier to read how much work is left to do.

added check to test if system is DONE and stop the beekeeper if there is
nothing left to do.

jobs that were created outside the database.  Also added ability to not
write (doesn't execute write_output or dataflow jobs).  Improved logic so
that if a job is specified, but something is wrong that the worker fails
rather than reverting to autonomous behaviour mode.

runWorker.pl now provides much better debugging capability through the
use of the -logic_name, -input_id and -no_write combination.

Fully tested by running a complete homology production test.

jobs that were created outside the database.  Also added ability to not
write (doesn't execute write_output or dataflow jobs).  Improved logic so
that if a job is specified, but something is wrong that the worker fails
rather than reverting to autonomous behaviour mode.
Fully tested by running a complete homology production test.

there is no longer the possibility that a worker might accidentally
claim the job it just failed on so there is no longer a need to check
the hive_id of the job when claiming. Removed check for hive_id

for use within the Hive workflow system.  Includes sections on the design
principles that went into the Hive including listing the emergent behaviour
rules that the Workers were designed around that allows this
controller-less system to function.
  http://en.wikipedia.org/wiki/Emergence
  http://en.wikipedia.org/wiki/Autonomous_Agent
  http://en.wikipedia.org/wiki/Artificial_life

jobs are created outside of the database (ie for debug runs).

level while not killing worker caused catastrophic hammering of analysis_job
table if there was a problem in the code where by all jobs fail.

analysis have failed.

control structure where a process/program has been made aware of the job(s)
they are responsible for controlling.  This is facilited via a job url:
   mysql://ia64e:3306/jessica_compara32b_tree/analysis_job?dbID=6065355
AnalysisJobAdptor::CreateNewJob now returns this url on job creation.
When a job is datflowed, an array of these urls is returned (one for each rule).
Jobs can now be dataflowed from a Process subclass with blocking enabled.
A job can be fetched directly with one of these URLs.
A commandline utility ehive_unblock.pl has been added to unblock a url job.
To unblock a job do:
   Bio::EnsEMBL::Hive::URLFactory->fetch($url)->update_status('READY');
This is primarily useful in asynchronous split process/parsing situations.

I needed to add in a check to prevent the worker from grabbing the same
job back and trying to run it again.  The retry works best when the job is
run on a different machine at a different moment in time (ie a different
hive_id).  This randomizes the run environment.

yet allow the worker to continue working.  Previous design killed the worker
when one of it's jobs fails.  This new design should be more efficient yet
still provide some level of inter-job-mangle-proof-ness.

but on a specific job.  For new system which catches job exceptions and
fails that job, but allows the worker to continue working.

to        eval "require Bio::EnsEMBL::Compara::DBSQL::DBAdaptor";

(taken from %self->input_job->dbID).

have not been run before (< retry_count)