gffstring mishandles transpliced genes

[ian-small]

'locus(gene).strand' in gffstring assumes all positions in the locus are on the same strand, which is not true for trans-spliced genes, e.g. join(11056..11919,complement(94283..95886)) is a valid GenBank locus that can now be faithfully represented using your much-appreciated rewrite of GenomicAnnotations.Locus, but which is rendered as
gene 11056 11919 . +
gene 94283 95886 . +
by GFF.gffstring()

[kdyrhage]

Thanks for bringing up these issues; I only work with prokaryotic data so I rarely encounter more complex locus descriptors myself.