Give all files to the parser, not just the first one.  The parser will
have to know what order the files are in (e.g. the Interpro parser knows
that the second file is a HTML file that contains release information).

NB: This means that the ordering of the list of URLs in the source_url
meta data table is important!

than calling the run() method as a class method.

Fixed bug that made the xref_parser.pl fall over on missing "LOCAL:"
files.

of BaseParser so that the file name comes last.  This will soon allow
some parsers to accept more than one file at a time.

characters from the MD5 hex digest...

use index() to do a string comparison against this to figure out whether
to parse the files or not.

Cleanup.

fails, the feature is reconstructed from transforms of the unagpped
components of the feature
- Added some checks to Transcript/Exon transform that warns and skips
supporting features that do not transform successfully

Regex cleanup and general parsing loop cleanup.

To force decompression, use 'xref_parser.pl' with the new '-unzip'
command line switch.

Removed '-compressed' command line switch.

pipe.  Use the command line switch -compressed with xref_parser.pl to do
this.  The default behaviour (without the -compressed switch) is still
to uncompress gzipped files.