Added command-line options to myisamchk (as suggested by Tim Cutts) to improve performance and only actually run myisamchk if it's required.

Added note to usage about it being best to run it on the actual cluster node that the MySQL instance is running on.

Now can parse a locator with space character in it. Needed as when using Registry with multiple species one needs to specify e.g. -species "Homo sapiens"

1. Use double-fork (not Open3) to communicate with LSF processes. More reliable and no quoting strangeness.

2. Add option to automatically upload external_db table

3. Bug fix: transcript/gene display_xrefs were not incorporating RefSeq xrefs since the split into RefSeq_peptide and RefSeq_dna had not been taken into account in the display xref calculation.

Use double-fork (not Open3) to communicate with LSF processes. More reliable and no quoting strangeness.

Corrected species IDs for (currently commented out) UniGene entries

increments this variable on every call to prepare and do methods