stored truncated batch_size (when job_limit < batch_size).  Fixed with a
'smart' worker->batch_size method which returns the lesser of either the
analysis->stats->batch_size or worker->job_limit

failed job to the failed worker.

added 'SYNCHING' and 'LOADING' to analysis_stat.status

added disconnect_if_idle before the sleep

The synchronization of the analysis_stat summary statistics was done by
the beekeeper at the top of it's loop.  For graphs with 40,000+ analyses
this centralized syncing became a bottle neck.  This new system allows
the Queen attached to each worker process to synchronize it's analysis.
Syncing happens when a worker 'checks in' and when it dies.  The sync on
'check in' only updates if the stats are >60secs out of date to prevent
over syncing.
The beekeeper still needs to do whole system syncs when a subsection has
finished and the next section needs to be 'unblocked'.  For homology this
will happen 2 times in a 16 hour run.

to turn this on like this anymore

which takes a hash ref as a parameter and returns a string which can
be evaled back to the hash.  $hash_ref = eval(encode_hash($hash_ref));

allows simple $stats->update_status('DONE');

so that branch_code is set explicitly rather than replying on the return
value of the write_output method.  Switched Worker.pm to use this value.

changed to varchar(255) (but dropped joining to analysis_data table)
If modules need more than 255 characters of input_id
they can pass the anaysis_data_id via the varchar(255) : example {adid=>365902}

workers can change batch_size as they run.

1) input_id is the command
2) input_id is formated like '{did=>123}'
  where did is short hand for analysis_data_id and the real command
  is stored in the analysis_data table

All STDOUT and STDERR from the command are autoamtically captured and redirected
to files (locations stored in analysis_job_file table).
Very simple idea, but might prove useful.

all parts of the Hive system.  Allows one to have a single use/include
use Bio::EnsEMBL::Hive;

Use instead DataflowRule and DataflowRuleAdaptor

on all anlyses, not just the ones with entries in the analysis_job table.
New logic is also faster.

logic as -loop option, but returns right away). Also modified checkk_for_dead
to take into account jobs with 'EXIT' status.

table.  Doing join on analysis_job.input_analysis_data_id=analysis_data.analysis_data_id
gives same performance as having analysis_job.input_id in table rather than second query

Removed select before store (made new method store_if_needed if that functionality is required by users)
and added option in AnalysisJobAdaptor::CreateNewJob to pass input_analysis_data_id
so if already know the CreateNewJob will be as fast as before.  Plus there are no limits on the
size of the input_id string.

if one isn't in the database

input_analysis_data_id int(10) which joins to analysis_data table.
added output_analysis_data_id int(10) for storing output_id
External analysis_data.data is LongText which will allow much longer
parameter sets to be passed around than was previously possible.
AnalysisData will also allow processes to manually store 'other' data and
pass it around via ID reference now.

isn't a corresponding Analysis entry in the analysis table

Estentially a mini filesystem so that data that would normally be stored in
NFS files and referenced via a path, can now be stored in the database and
referenced via a dbID. Data is a LONGTEXT.
Can be used to store configuration data, paramater strings,
BLOSSUM matrix data, uuencode of binary data .....

Added get_available_adaptors to get the pairs of name / adaptor modules. This should make it easier to read etc