GitLab

Edge case support for hive which allows for non-scheme URIs. We now treat these as if they are DB URIs.

Adding support for SQLite since the URI scheme does not take into account the fact that SQLite is a path and the concept of table does not play well with this

DBNAME is no longer a required attribute as we now allow for direct DB connections. SQLHelper is now an attribute of a DBConnection. Methods added to avoid unintentional DB disconnections, working with the DBH safely and quote DB identifiers if required (useful to avoid SQL injection attacks with auto-generated code). Equals method also improved. Associated test cases written.

Removing URIParser sub-package. It was nice in theory but I'd rather support different schemes using extensions to the URI package

Minor POD corrections

Removing URIParser sub-package. It was nice in theory but I'd rather support different schemes using extensions to the URI package

According to test cases adaptor() should be nullable by passing an undef through to the mutator. Detection of a purposful undef is by checking the length of @_

db_name in external_db. Clean up a few bugs while I'm at it.

Bugs corrected:

  A list (rather than a list reference) was returned if the
  external_db_id for the given db_name could not be found.

  The list returned from list_gene_ids_by_external_db_id() (in
  DBEntryAdaptor) was used as a hash.

For release 64 we've found that features on human HSCHR5_1_CTG1 would
not get fetched since its seq_region_id is an internal_seq_region_id
with a corresponding, but non-existent, external_seq_region_id in the
seq_region_mapping table.

Bringing in assertions for strand as well as for boolean types. Extending the test cases accordingly

"lsf_opt_exonerate" in the configuration file.

if you first check if the transcript exists, you won't get any warnings about undefined values anymore

to difference in seq_region names.