default behaviour is to use registry and staging, but can also use alternative database on other server

Added 'coord_system' as an option.

Minor formatting.

there should always be a common_name, throw if not

option added to specify connection details in the xref_config.ini file if your database is not on staging

both can be loaded via the registry or via the parameters set up in the xref_config.ini file

if data_uri is defined, all connection details can be read for vega and core databases
source_prefix is read from data_uri as well, so no need to hard-code it in the module
if there is a source_prefix, it is a merged species, else, skip the module alltogether

source_name and host are read from the xref_config.ini file, as they are already there, rather than hard-code them in the module
If no host is found, this should not be a merged species

Exit but not failed if no external references found, as some species (pig) can be merged without official naming

typo ignoreed replaced by ignored

added method transcript_names_from gene, which assigns uses whatever name has been assigned to the gene for all of its transcripts, with an extension number (-201, -202...)
for merged species, this method will be skipped as the official naming module does that already

removed ignore for uniprot_genename with low evidence
uniprot_genename assignments are trusted

a blank line was removed in the process

pig (sus_scrofa) is now a merged species, needs to be added to the list of existing three, human, mouse and zebrafish

Added a no_cache override feature plus testcase. Useful in times of memory crisis coupled with annoyingly slow lookups of small data.

[ENSCORESW-271]. Fixed the test being too chatty during cleanup by setting the RUNTESTS_HARNESS_NORESTORE before firing off a thread

[ENSCORESW-270]. Fixed issue with can() call which was returning a boolean flag rather than the coderef can() actually returns. Switched to local variables and the entire thing looks fine now

former script nonsense_transcript_attribs.pl moved to module format to be run as part of the production pipeline
list of populations has been updated after discussion with variation

if there is only one row, result can be retrieved by fetching the first element of the de-referenced array, which is itself an array where each element corresponds to a column

new (resurrected source ncRNA_EG::EG, which replaces RNAMMER, TRNASCAN and RFAM::EG sources); passes as an argument the logic_name instead of a biotype

Introduced a new argument, logic_name, so we can fetch genes by logic_names rather than by biotype, easier for ensemblgenomes to pick up all ncRNA genes, as there are so many biotypes behind them. Goes with new source, nCRNA_EG::EG (resurrected actually). Because this source does not correspond with any external_db ( e.g. RFAM ) , as it was, so I had to add some logic to associate the xref source with an external_db on the fly.

Warn is assembly_exception table is empty.

Revised fetch_all_by_Slice to allow for a proper feature count method that includes overlaps and clones. Response to lack of performance in generic_count for Variation (request from Laurent Gil)

Added musa acuminata, solanum tuberosum species; added two new sources, phytozome_gmax and pgsc; added UniParc source to plants genomes