ensemblgenomes-17-70'.

Sprout from branch-e70-db-portability 2013-01-24 00:49:18 UTC cvs2git <cvs2git> 'This commit was manufactured by cvs2svn to create branch 'branch-e70-db-'
Cherrypick from branched-annot-ext-devel 2013-02-04 09:54:19 UTC Mark McDowall <mcdowall@ebi.ac.uk> 'Modifications for associated_xrefs':
    modules/Bio/EnsEMBL/DBSQL/DBEntryAdaptor.pm
    modules/Bio/EnsEMBL/OntologyXref.pm

Filter terms out which span more than 1 ontology. This is not the right way to do it but is the best for the moment

Only use the coordinates with the smaller seq_region_id (usually 1-22,X,Y) to avoid mapping the LRG to the patch region

Adding a no_error flag to allow for simple single value querying where the number of rows returned could be ambigious

Go through an object and make sure everything is populated once we go through a call of $gene->load().

given two databases of the same species, compares the xrefs used for gene display to describe the statistics of what has changed between databases

Applied wait constraints to all three calls. Could have refactored, but the system calls are fiddly and I'll probably make more mistakes that way.

Added a queue waiting command to prevent e.g. Human and Mouse projecting simultaneously when Mouse relies upon Human to be complete and correct.

If PASS is not pulled out into a local variable then passed into DBI we will attempt to connect to a database (DBI->trace() confirms the passing of a password). Perl is very odd

Counting features over HAPs/Patches/PARs was not working. Fixed by adding a SQL based filtering stage. The normal code will not use this as it is quite nasty

data should rely on protein evidence levels provided by Uniprot

[ENSCORESW-315]. Fixing a bug where the bind params need to be reused on subsequent queries. MySQL/DBI/DBD::mysql is really bad at warning us when we submit a query with a single ? and no bind parameters (it compensates). Wish it didn't